Bluegrass Messengers

The Texts of King Horn- James R. Hurt 1970


The Texts of King Horn
by James R. Hurt 
Journal of the Folklore Institute, Vol. 7, No. 1 (Jun., 1970), pp. 47-59

The Texts of King Horn
JAMES R. HURT

The medieval romance is in many ways a surprisingly mysterious genre, at least in its English manifestations. It was the dominant literary form in Europe for about four hundred years, from the twelfth to the sixteenth centuries, and a very large body of texts survive in a number of languages. The English metrical romances have a respectable place among these
texts, and the best of them are among the masterpieces of medieval literature. Yet we know very little for sure about these romances, who wrote them, how they were disseminated, and what kind of audience read them or listened to them sung. Elementary textbooks often tell us that they were composed by minstrels, but there is no convincing evidence that this was often the case. In a few instances there seem to be indications that clerics composed romances, but general statements about
clerical origins are as hard to support as assertions of minstrel authorship.

And were romances intended to be read aloud publicly, sung, chanted, read silently and privately? Were they intended, by and large, for aristocratic audiences, bourgeois audiences, peasants? What was the place of a minstrel in medieval English society? How did he learn and practice his art?

These are all hard questions. Though some of them may be answered tentatively for individual romances, there is still much we need to learn about the romances, their form and their social setting, before we can make many generalizations about them.[1]
Fortunately, however, we can ask these old questions in a somewhat different and perhaps more fruitful way as a result of the vigorous study of oral poetry that has been going on in a number of fields for the last few years. The so-called "oral-formulaic theory" of oral poetry originated in some studies of Homer by Milman Parry in the 1930's. Seeking a contemporary poetic tradition analogous to Homer's, Parry and Albert B. Lord went to Yugoslavia, where they found a flourishing tradition of folk epic, carried on by "guslars," who chanted lengthy verse narratives to the accompaniment of a simple stringed instrument. Parry and Lord recorded a number of these poems and on the basis of their study of them formulated a general theory of oral poetry, described most fully in Lord's The Singer of Tales.[2]

Students of other literatures were quick to follow the lead of Parry and Lord and in the past few years oral-formulaic studies have been published of Old English narrative poetry, Middle English romance, the French chanson de geste, the Spanish romancero, Gaelic poetry, the English and Scottish popular ballad, the Middle High German epic, and others.[3] In the course of these studies, the Parry-Lord version of the theory has been criticized, modified, and in some cases, perhaps, even distorted. The result, however, has been a much clearer idea of the nature of oral poetry than was possible before.

I propose to suggest some of the implications this work has for the study of Middle English metrical romance by taking up one of the most interesting of these romances, King Horn, summarizing some of the critical and historical problems it presents, and suggesting how they may be reconsidered in the light of the oral-formulaic theory.

The story of Horn survives in four distinct forms. We are concerned with the metrical romance version known as King Horn, but it also forms the subject of an Anglo-French romance, Horn et Rimenhild, by a certain Master Thomas. The three manuscripts which survive date from the late thirteenth century. Another version is Horn Childe and Maiden Rimnild, a fourteenth century poem from Northern England which survives in only one manuscript. Last is the ballad version of the story, Hind Horn, of which Child printed eight versions (Child # 17). None of these versions of the story seems to be derived directly from another. Horn Childe and Maiden Rimnild and the ballad seem to be distinctively Northern versions of the story, with a number of Irish touches. King Horn is a Southern version of the same story, while Horn et Rimenhild is a comparatively late and sophisticated version of the story by a man familiar with both the Northern and Southern traditions.[4]

The story of King Horn is so thoroughly conventional that it might almost be told in a series of Thompson's motif headings. Horn is the son of King Murry and Queen Godhild of Suddenne. (I use the versions of the names in the Cambridge manuscript.) Saracens invade the kingdom, kill the king, and set Horn adrift in a small boat with twelve comrades. The boat drifts to Westernesse, where Horn quickly wins the favor of the king, Aylmar, who has Horn learn to sing and play the harp and makes him an honored servant. The king's daughter, Rymenhild, falls so deeply in love with Horn that she almost dies of lovesickness. Horn, when he learns of her love, objects that he is only a servant and Rymenhild succeeds in having him and his twelve comrades knighted. Horn immediately distinguishes himself by killing Saracens and becomes an honored member of the court. But he is betrayed by his comrade Fykenhild, who covets Rymenhild himself and turns Aylmar against Horn. Horn flees to Ireland, after asking Rymenhild to wait seven years for him and receiving a magic ring from her. In Ireland, under the name ot
Cutberd, he rises to fame as quickly as he had in Westernesse, but at the end of seven years he learns that Rymenhild is being forced to marry King Modi, one of his enemies. He quickly returns to Westernesse, enters the court disguised as a palmer, reveals himself to Rymenhild by dropping the ring in her cup, and rescues her from the unwanted marriage. Restored to favor with King Aylmar, he parts with Rymenhild again to return to Suddenne to avenge his father.

The remainder of the story is a rather stranger eduplication o f the earlier episodes. Again, Rymenhild is threatened, this time by the ubiquitous Fykenhild, and again Horn rescues her, this time disguised as a harper. He rewards his friends, marries Rymenhild (at last!) and returns as king to his native Suddenne.

The roots of this tale in the mass of Western European folk narrative material will be readily recognized. (Gerald Bordman has catalogued forty-two T hompson m otifs i n the romance.[5]) M ore r elevant t o my present purpose,h owever,a re the peculiaritiesin the threet exts we have of the romance. The oldest of these (the "C" text) has been dated at about 1260
and is in the University Library, C ambridge. The other two texts (the "L" and "O" texts), a pparently w ritten about fifty years after the C text, are in the British Museum and the Bodleian Library, Oxford.[6]

These texts have been a great problem to editors. Although all three adhere closely to the same sequence of events, they differ so much in detail that it is hard to find a single line that is exactly the same in all three.N or aret he differencessu cht hat they can be resolved by traditional methods of textuals tudy. To illustrate, the C text contains7 65 couplets; if we take the appearanceo f the same rhymea s a very loose criteriono f identity, 107 of these couplets are different from the corresponding
passagesi n the other manuscriptsa nd seventy-threme ore occur in only one of the other manuscripts. This means t hat the threet exts have only 1170 lines in common, out of a total of 1530 lines in the C text. And if we impose a stricter test of identity than the presence of the same rhyme word, the figure is far lower.
Editors have attempted to solve this puzzle in various ways. Joseph

Ritson, the first editor, merely printed the L text in his Ancient Engleish
MetricalR omancees( 1802)w ithoutt ryingt o establishi ts relationshipto
the other two texts. The C text was published in 1845 by Francisque
Michel, and the O text in 1872 by Carl Horstmann. But the first real
50 JAMES R. HURT
THE TEXTS OF King Horn
effort to relate the three texts was Theodor Wissmann's introduction to
his critical edition of the poem in 1881. After comparing the three
manuscripts, he concluded that the only explanation for their radical
divergences was that they had been transmitted orally: "das Lied von
King Horn in miindliche Ueberlieferung von einem Sanger dem andern
iibertragen wurde."7
This view gets short shrift from Joseph Hall in the introduction to his
edition of King Horn in 1901. In this excellent edition, one of the best we
have of a medieval romance, Hall prints all three texts side by side. He
nevertheless believes that the disparities can be attributed wholly to
scribal corruption or rewriting. The three texts derive from a common,
written original, through several intermediary texts, he thinks, but "all
have rewritten, generally with poor effect, passages which have been
corrupted in process of transmission." Wissmann's theory of oral
transmission, he says, "gets no support from what we know of the history
of all other Middle-English romance texts."8
A third attempt to explain the relationship of the manuscripts is by
Walter H. French.9 In a lengthy study of the poem, he attacks the
views of both Wissmann and Hall and constructs a completely different
interpretation of the evidence. L and 0, he thinks, derive from a common
source which was very similar to the parent text. C stems from a different
source and is full of corruptions. The scribes of all three freely rewrote
the lines because they did not understand the iambic trimeter in which
French surmises the original must have been written.
The radically different interpretations of the same evidence by these
three scholars suggest that traditional methods of textual study are not
wholly adequate to deal with texts like King Horn. Wissmann constructs
a genetic chart, only to reject it, Hall provides another chart, and French
opposes Hall's chart by constructing another of his own. Recent studies
of medieval romance as oral poetry suggest that the whole problem should
be reconsidered. It is no longer possible to say, as Hall did, that a theory
of oral transmission is incompatible with "the history of other Middle-
English romance texts."
The fullest study of medieval romances as oral-formulaic poetry has
7 "Das Lied von King Horn," Quellen und Forschungenz ur Sprach- und Culturgeschichte
(Strassburg, 1881), Vol. XLV, p. v.
8 Hall, p. xiv.
9 Walter H. French, Essays on King Horn (Ithaca, New York, 1940).
51
JAMESR . HURT
been carried out by Albert C. Baugh.10 Baugh extends to Middle English
metrical romance the methods used by Parry and Lord in their studies of
Homeric epic and the songs of the Yugoslavian guslars. According to the
Parry-Lord theory of oral-formulaic composition, a singer may improvise
a long narrative poem by filling out a general memorized outline with
conventional elements. These elements are of two kinds: formulas (word
groups "regularly employed under the same metrical conditions to express
a given essential idea") and themes ("recurrent elements of narration or
of description in traditional oral poetry"). A formula is thus a stock
verbal and metrical unit; a theme is a stock nonverbal narrative element.
The phrase "Beowulf makelode" in Beowulf is a formula; the stock
sequence of details in the description of a knight being armed is a theme.
The Matter of England romances, Baugh finds, are full of such phrases
and such sequences, and they function as similar stock sequences have
been shown to function in other oral poetry. He supports this thesis by
detailed analyses of passages from Beves of Hampton, Guy of Warwick,
and Richard the Lion-Hearted. He concludes:
Those who have developed the theory of oral composition have argued that ...
the Iliad or the Chanson de Guillaume ... were composed by those who recited
them and in the process of recitation, just as the long narrative poems of
Jugoslavia are composed by the guslars today.... The Middle English romances
cannot in general have arisen in this way, since in so many cases they are
translationso f Frenchr omances,o ften so faithfult o their sourcea s to make it
most likely that the English translator was working with his French original
open beforeh im. However ... I have tried to show that improvisationh as had
a share in the production of the different versions of the same romance that
have come down to us.1l
There are several problems in this conclusion, of which Baugh is well
aware and to which we shall return in a moment, but first, in what sense
(if in any) is King Horn an oral-formulaic poem? The usual procedure
in oral-formulaic studies is to examine samples of the text for verbal
units which meet the requirements of formulas, that is, which are closely
similar verbally and interchangeable metrically with other units in the
same or similar poems. If a sizable portion of the poem is found to be
made up of such units, it is regarded as "oral-formulaic." Here are the
first eighteen lines of King Horn. (The C text is quoted.)
10 "Improvisationi n the Middle English Romances,"P roceedingso f the American
PhilosophicaSl ociety, CIII (1959), 418-454.
1 Ibid., pp. 439-440.
52
THE TEXTS OF King Horn
Alle beon he blibe
bat to my song lybe:
a sang ihc schal sou singe
Of Murry pe kinge 4
King he was biweste
So longe so hit laste.
Godhildh et his quen
Faire ne mi3te non ben. 8
He hadde a sone bat het horn,
Fairer ne miste non beo born.
Ne no rein vpon birine,
Ne sunne vpon bischine: 12
Fairer nis non pane he was,
He was bri3t so pe glas,
He was whit so pe flur,
Rose red was his colur. 16
In none kinge riche
Nas non his iliche.
I have italicized, in the customary fashion, the lines that appear to be
formulaic, judging from the parallels to them that Hall cites in his
voluminous notes. Hall notes that although the idea of the first couplet
is common, "there is nothing quite parallel to this." But for the other
lines (except for line 9, in which Horn's name appears), he cites as many
as thirty-eight parallels of varying degrees of similarity. Furthermore, he
identifies several couplets (lines 11-12, 15-16, 17-18) as stock combinations.
This extremely high proportion of formulas may be partially due to
the fact that these lines constitute a conventional opening, but a glance
at Hall's notes on the rest of the poem will show that the difference is
not great. A somewhat different sort of evidence is provided by the
following passage, in which the divergences among the three texts are
apparent. The scene is the knighting of Horn by King Aylmar (lines
495-502 in the C text).
L O
be day bigon to springe be day by gan to springe
horn cor byfore be kynge Horn cam bi forn be kinge
wib his tuelf fere Wit swerde horn he girde
alle ber ywere Rit honder hys herte
horn knyht made he He sette him on stede
wib ful gret solempnite Red so any glede
sette him on a stede And set him on his fotes
red so eny glede Bobe spores and botes
smot him a lute wiht And smot alitel with
& bed him buen a god knyht And bed him ben god knict
53
C
ke day bigan to springe,
Horn com biuore le kinge
Mid his twelf yfere:
Sume hi were lukere.
Horn he dubbede to kni3te
Wit swerd & spures bri3te.
He sette him on a stede whit:
ter nas no kni3t hym ilik.
He smot him a litel wi3t
& bed him beon a god kni3t.
The extent of the variations among the texts is demonstrated pretty
clearly here. Only the first and last couplets are even roughly identical
in the three texts; one more occurs in both L and O. The couplets,
moreover, are all end-stopped and constitute interchangeable units. The
L text's "horn knyht made he/wit ful gret solempnite" could be substituted
for the O text's "Wit swerde horn he girde/Rit honder hys herte"
without much affecting the sense or the style. Furthermore, all these
lines are widely paralleled in other romance literature, as Hall's notes
show.
Let me cite one more short passage in which the differences among
the texts are even greater. This is the scene in which King Aylmar exiles
Horn (lines 705-714 in the C text).
L
he fond horn vnder arme
in rymenyldes barme
go out quot aylmer te kyng
Horn tou foule fundlyng
forb out of boures flore
for rymenild bin hore
wend out of londe sone
her nast kou nout to done
wel sone bote kou flette
myd suert yshal ke sette
0
He fond horn wit arme
in rimenyldes barme
Henne out qwad aylmer king
Henne tou foule wendling
Out of boure flore
Fram reymyld ti hore
Sone bote ke flette
Wit swerd hy wole be hette
Hout of londe sone
Here hauest bou nowt to done
C
He fond horn in arme.
On Rymenhilde barme
'Awei vt,' he sede, 'fule keof!
Ne wurstu me neuremore leof.
Wend vt of my bure
Wit muchel messauenture.
54 JAMES R. HURT
THET EXTOS FK ing Horn
Wel sone bute tu flitte,
Wib swerde ihc be anhitte.
Wend vt of my londe
Ober bu schalt haue schonde.'
Again, only two couplets, the first and the last in the L text, appear in
all three texts. L and 0, in this passage, are comparatively close in their
correspondence, but C uses different formulas to express the same
sequence of ideas.
King Horn also yields many examples of oral-formulaic themes. We
have already touched on two such themes obliquely: the stock sequence
of ideas in a minstrel opening tag - a call for attention, a hope that the
listeners will be entertained, and a statement of the nature of the story
to follow - and the stock sequence in the knighting ceremony -
dubbing, arming, and mounting. Other themes appear in the arming
scenes, the battle scenes, the love scenes (brief as they are), scenes in
which strangers meet - indeed the ehtire narrative is a tissue of interwoven
themes. But for the sake of brevity, let us take three representative
"journey" passages and examine the sequence of ideas in them. Only the
C text will be cited.
117ff. be se bigan to flowe
& horn child to rowe;
be se bat schup so fasste drof
be children dradde berof
Al be day & al be ni3t,
Til hit sprang dai li3t
Til Horn sa3 on be stronde
Men gon in be londe.
1007ff. Horn dude him in be weie
On a god Galeie
be wind him gan to blowe
in alitel bro3e.
be se bigan to posse
Ri3t in to Westernesse.
He let his schup stonde
& 3ede to londe.
1499ff. Horn tok Rymenhild bi be honde
& ladde hure to be stronde,
be se bigan to flowe
55
& horn gan to Rowe.
Horn gan for to ride,
be wind him bleu wel wide.
He ariuede in yrlonde.
Several formulas occur in these passages: for example the couplet in
lines 117-118 and in lines 1503-1504 (which occurs a third time in Horn,
in lines 1095-1096) and the "strand/land" rhyme. In addition, however,
we can see a stock sequence of ideas that constitutes a theme. In each
of these journey passages, the character embarks, one or two couplets
describe the wind or the sea, and the character arrives. There is no
attempt at realistic description, merely a stock sequence of "journey"
elements, which may be expressed in formulas.
A full formulaic analysis of King Horn would, of course, be the work
of a thick volume, not an article, and I believe that the sample evidence
which has been presented may be taken as an indication, if not a proof,
that oral-formulaic methods have had some part in the composition or
the transmission of the poem.
The question to which we now may pass is, what use can we make of
this information? Does it throw any light on the question cf the textual
history of the poem, with which the editors have had so many difficulties?
A simple and attractive possibility would be to say that the three texts
derive from three separate oral performances. I am afraid that this
possibility must be immediately rejected, however. For one thing, the
fifty year difference between the dates of C and of L and O would mean
that either (1) the same singer had dictated the L and O texts fifty years
after the C text, (2) L and O are copies of an earlier text derived from the
C singer's dictation, or (3) two or more singers dictated the three texts.
The first possibility is blatantly improbable, the second is open to the
same objections as the traditional accounts of the textual history, and
the third raises serious questions about the possible degree of similarity
between performances of the same romance by different singers. These
questions are difficult to answer - transcriptions of more than one
performance of an oral poem are rare, except for the Balkan epics
collected by Parry and Lord - but the oral-formulaic theory itself would
lead us to expect greater differences among the texts than in fact exist,
even if they had been dictated by the same singer, much less by two or
more.
56 JAMES R. HURT
THE TEXTS OF King Horn
A second possibility is more promising, though it raises a serious issue
that has been much debated in oral-formulaic studies. This is the possibility
that the scribes themselves functioned as oral-formulaic poets and
reshaped their source texts along the lines we have described. The issue
that such a hypothesis raises is that of the relationship between oralformulaic
poetry and writing. The original theory is clearly and exclusively
a theory of improvisation. The oral poet uses formulas and
themes in order to be able to "improvise" long passages of verse on the
spot, before an audience. The literate poet, with leisure to invent and
shape, has no need to rely exclusively on formulas and is indeed psychologically
unable to do so. This is so important a point that it is worthwhile
to quote Lord himself on the subject:
It is necessaryf or us to face squarelyt he problem of "transitional"te xts. Is
there in realitys uch a phenomenona s a text which is transitionalb etweeno ral
and written literary tradition? ... I believe that the answer must be in the negative,
because the two techniques are, I submit, contradictory and mutually
exclusive. Once the oral technique is lost, it is never regained. The written
technique, on the other hand, is not compatible with the oral technique, and
the two could not possibly combine, to form another, a third, a "transitional"
technique.12
The question of the possibility of oral-formulaic techniques within a
written tradition is particularly troublesome in the study of Anglo-Saxon
and Middle English poetry, where certain texts which are demonstrably
direct translations from written sources have been shown to be as heavily
formulaic as other texts which may have been orally composed, such as
Beowulf As Baugh notes, the medieval romances themselves present this
problem, since many of them are close translations of written French
originals. Faced with these contradictions, a number of English medievalists
have fallen back upon a hypothesis that all Anglo-Saxon and Middle
English "oral poetry" is transitional. Even the earliest surviving English
poetry may be the product of literate poets recalling a still earlier oral
tradition.
Lord himself has pointed out the difficulty in this way out of the
dilemma. He acknowledges that historically there must be a period of
transition between oral and written, but insists that even within such a
period any particular text must be either oral-formulaic or literary, never
12 Lord, pp. 128-129.
57
JAMES R. HURT
transitional. The literary texts may be reminiscent of the oral tradition,
but they will be different in kind, not merely in degree.
Formula analysis, providing of course, that one has sufficient material for
significantr esults,i s, therefore,a ble to indicatew hethera ny given text is oral
or "literary."A n oral text will yield a predominanceo f clearly demonstrable
formulas,w ith the bulk of the remainder" formulaic,"a nd a small numbero f
nonformulaic expressions. A literary text will show a predominance of nonformulaic
expressions, with some formulaic expressions, and very few clear
formulas.13
If we use the term "oral-formulaic" in reference to a literary work, no
matter how traditional in style, therefore, we must make it very clear
that we are using the term in a derived sense, and we must thoroughly
rethink our definitions of such terms as "formula" and "theme" to fit an
altogether different situation.
Regardless of the history of other Middle English metrical romances,
King Horn seems likely to have been, at some point in its history before
it was written down, an oral-formulaic poem in the strict sense of the
term. The density of its formulas and its thoroughgoing structuring by
themes seem strong evidence that the story material which elsewhere
became a ballad and a literary romance here became an oral-formulaic
romance, continually recreated by an oral singer. But the written versions
which survive, although they must derive from a version dictated by a
singer, have undergone revision, probably by scribes also working oralformulaically,
but in the second, derived sense of the term. The relatively
close correspondences among the three texts indicate that the poem has
not been recreated each time; the divergences in formulaic detail indicate
that the transcription has been altered by scribes with some understanding
of the oral-formulaic method.
Such an account of the poem does not solve the difficult problems of
its textual history; it may even imply new and even more difficult ones.
But it is worthwhile if it states the problems in a way that reflects the
nature of the poem better than traditional methods of textual study do.
The development of nineteenth century textual study into modern scientific
bibliography has been spectacularly successful, but perhaps we have
not fully realized how thoroughly these methods reflect a print-centered
culture, in which, as Lord says, "writing has fixed the norm of a stable
first creation in art."14 When we cross the boundary into the study of the
'1
Ibid., p. 130.
58
THE TEXTS OF King Horn
products of an oral culture, the concept of a fixed original has little
meaninga ny more.P erhapsw e now havet he materialsa vailablef or more
suitable methods for studying the texts of oral literature.
Universityo f Illinois
Urbana, Illinois
I Ibid., p. 101.

_______________
Footnotes:

1. Two excellent accounts of some of these problems are Albert C. Baugh, "The Authorship of the Middle English Romances," Modern Humanities Research Association Bulletin, No. 22 (1950), pp. 13-28; and "The Middle English Romance: Some Questions o f Creation, Presentation, and Preservation," Speculum X, LII (1967),1 -31.

2. Albert B. Lord, The Singer of Tales (Cambridge, Mass., 1960).

3. Francis P . Magoun, Jr., "The Oral-Formulaic C haractero f Anglo-Saxon N arrative Poetry," Speculum, XXVIII (1953), 446-467; Ronald A. Waldron, "Oral-Formulaic Techniquea nd Middle-EnglishA lliterativeP oetry," Speculum,X XXII (1957),7 92-804;
Jean Rychner, La chanson de geste: essai sur l'art epique des jongleurs (Geneva, 1955); Bruce A. Beatie, "Oral Traditional Composition in the Spanish Romancero of the Sixteenth Century," Journal of the Folklore Institute, I (1964), 92-113; James Ross, "FormulaicC ompositioni n Gaelic Oral Literature,"M odernP hilology,L VII (1959), 1-12; JamesH . Jones, "Commonplacesa nd Memorizationi n the Oral Traditiono f the English and Scottish Popular Ballads," Journal of American Folklore, LXXIV (1961),
97-112, and Albert B. Friedman, "The Formulaic Improvisation Theory of Ballad Tradition - A Counterstatement,"J ournal of American Folklore, LXXIV (1961), 113-115; Robert L. Kellogg, "The South Germanic Oral Tradition," in Franciplegius: Medieval and Linguistic Studies. Studies in Honor of Francis P. Magoun, Jr., ed. Jess B. Bessinger, Jr. and Robert P. Creed (New York, 1965), pp. 66-74. For a good criticism of the theory, see Robert D. Stevick, "The Oral-Formulaic Analyses of Old English
Verse," Speculum, XXXVII (1962), 382-389, and for a good survey and further bibliography, see Michael Curschmann, "Oral Poetry in Mediaeval English, French, and German Literature: Some Notes on Recent Research," Speculum, XLII (1967), 36-52.

4 These versions and their relationships are described more fully by Joseph Hall (ed.), King Horn: A Middle-EnglishR omance( Oxford, 1901), pp. li-lvi.

5. Gerald Bordman, Motif-Index of the English Metrical Romances (= FF Communications, No. 190) (Helsinki, 1963).
6. The manuscripts are thoroughly described by Hall, pp. vii-xv.