EPIC AND ROMANCE
ESSAYS ON MEDIEVAL LITERATURE
BY
W. P. KER
FELLOW OF ALL SOULS COLLEGE, OXFORD
PROFESSOR OF ENGLISH LITERATURE IN UNIVERSITY COLLEGE
LONDON
Contents
MACMILLAN AND CO., LIMITED
ST. MARTIN'S STREET, LONDON
1931
COPYRIGHT
First Edition (8vo) 1896
Second Edition (Eversley Series) 1908
Reprinted (Crown 8vo) 1922, 1926, 1931
PRINTED IN GREAT BRITAIN
BY R. & R. CLARK, LIMITED, EDINBURGH
PREFACE
These essays are intended as a general description of some of the principal forms of narrative literature in the Middle Ages, and as a review of some of the more interesting works in each period. It is hardly necessary to say that the conclusion is one "in which nothing is concluded," and that whole tracts of literature have been barely touched on—the English metrical romances, the Middle High German poems, the ballads, Northern and Southern—which would require to be considered in any systematic treatment of this part of history.
Many serious difficulties have been evaded (in Finnesburh, more particularly), and many things have been taken for granted, too easily. My apology must be that there seemed to be certain results available for criticism, apart from the more strict and scientific procedure which is required to solve the more difficult problems of Beowulf, or of the old Northern or the old French poetry. It is hoped that something may be gained by a less minute and exacting consideration of the whole field, and by an attempt to bring the more distant and dissociated parts of the subject into relation with one another, in one view.
Some of these notes have been already used, in a course of three lectures at the Royal Institution, in March 1892, on "the Progress of Romance in the Middle Ages," and in lectures given at University College and elsewhere. The plot of the Dutch romance of Walewein was discussed in a paper submitted to the Folk-Lore Society two years ago, and published in the journal of the Society (Folk-Lore, vol. v. p. 121).
I am greatly indebted to my friend Mr. Paget Toynbee for his help in reading the proofs.
I cannot put out on this venture without acknowledgment of my obligation to two scholars, who have had nothing to do with my employment of all that I have borrowed from them, the Oxford editors of the Old Northern Poetry, Dr. Gudbrand Vigfusson and Mr. York Powell. I have still to learn what Mr. York Powell thinks of these discourses. What Gudbrand Vigfusson would have thought I cannot guess, but I am glad to remember the wise goodwill which he was always ready to give, with so much else from the resources of his learning and his judgment, to those who applied to him for advice.
W. P. KER.
London, 4th November 1896.
POSTSCRIPT
This book is now reprinted without addition or change, except in a few small details. If it had to be written over again, many things, no doubt, would be expressed in a different way. For example, after some time happily spent in reading the Danish and other ballads, I am inclined to make rather less of the interval between the ballads and the earlier heroic poems, and I have learned (especially from Dr. Axel Olrik) that the Danish ballads do not belong originally to simple rustic people, but to the Danish gentry in the Middle Ages. Also the comparison of Sturla's Icelandic and Norwegian histories, though it still seems to me right in the main, is driven a little too far; it hardly does enough justice to the beauty of the Life of Hacon (Hákonar Saga), especially in the part dealing with the rivalry of the King and his father-in-law Duke Skule. The critical problems with regard to the writings of Sturla are more difficult than I imagined, and I am glad to have this opportunity of referring, with admiration, to the work of my friend Dr. Björn Magnússon Olsen on the Sturlunga Saga (in Safn til Sögu Islands, iii. pp. 193-510, Copenhagen, 1897). Though I am unable to go further into that debatable ground, I must not pass over Dr. Olsen's argument showing that the life of the original Sturla of Hvamm (v. inf. pp. 253-256) was written by Snorri himself; the story of the alarm and pursuit (p. 255) came from the recollections of Gudny, Snorri's mother.
In the Chansons de Geste a great discovery has been made since my essay was written; the Chançun de Willame, an earlier and ruder version of the epic of Aliscans, has been printed by the unknown possessor of the manuscript, and generously given to a number of students who have good reason to be grateful to him for his liberality. There are some notes on the poem in Romania (vols. xxxii. and xxxiv.) by M. Paul Meyer and Mr. Raymond Weeks, and it has been used by Mr. Andrew Lang in illustration of Homer and his age. It is the sort of thing that the Greeks willingly let die; a rough draught of an epic poem, in many ways more barbarous than the other extant chansons de geste, but full of vigour, and notable (like le Roi Gormond, another of the older epics) for its refrain and other lyrical passages, very like the manner of the ballads. The Chançun de Willame, it may be observed, is not very different from Aliscans with regard to Rainouart, the humorous gigantic helper of William of Orange. One would not have been surprised if it had been otherwise, if Rainouart had been first introduced by the later composer, with a view to "comic relief" or some such additional variety for his tale. But it is not so; Rainouart, it appears, has a good right to his place by the side of William. The grotesque element in French epic is found very early, e.g. in the Pilgrimage of Charlemagne, and is not to be reckoned among the signs of decadence.
There ought to be a reference, on p. 298 below, to M. Joseph Bédier's papers in the Revue Historique (xcv. and xcvii.) on Raoul de Cambrai. M. Bédier's Légendes épiques, not yet published at this time of writing, will soon be in the hands of his expectant readers.
I am deeply indebted to many friends—first of all to York Powell—for innumerable good things spoken and written about these studies. My reviewers, in spite of all differences of opinion, have put me under strong obligations to them for their fairness and consideration. Particularly, I have to offer my most sincere acknowledgments to Dr. Andreas Heusler of Berlin for the honour he has done my book in his Lied und Epos (1905), and not less for the help that he has given, in
W. P. K.
Oxford, 25th Jan. 1908.
CONTENTS
CHAPTER I
INTRODUCTION
I
The Heroic Age
PAGE Epic and Romance: the two great orders of medieval narrative 3 Epic, of the "heroic age," preceding Romance of the "age of chivalry" 4 The heroic age represented in three kinds of literature—Teutonic Epic, French Epic, and the Icelandic Sagas 6 Conditions of Life in an "heroic age" 7 Homer and the Northern poets 9 Homeric passages in Beowulfand in the Song of Maldon 10
11 Progress of poetry in the heroic age 13 Growth of Epic, distinct in character, but generally incomplete, among the Teutonic nations 14
II
Epic and Romance
The complex nature of Epic 16 No kind or aspect of life that may not be included 16 This freedom due to the dramatic quality of true (e.g. Homeric) Epicas explained by Aristotle 17
17 Epic does not require a magnificent ideal subject
such as those of the artificial epic (Aeneid, Gerusalemme Liberata, Paradise Lost) 18
18 The Iliad unlike these poems in its treatment of "ideal" motives (patriotism, etc.) 19 True Epic begins with a dramatic plot and characters 20 The Epic of the Northern heroic age is sound in its dramatic conception
and does not depend on impersonal ideals (with exceptions, in the Chansons de geste) 20
21 The German heroes in history and epic (Ermanaric, Attila, Theodoric) 21 Relations of Epic to historical fact 22 The epic poet is free in the conduct of his story
but his story and personages must belong to his own people 23
26 Nature of Epic brought out by contrast with secondary narrative poems, where the subject is not national 27 This secondary kind of poem may be excellent, but is always different in character from native Epic 28 Disputes of academic critics about the "Epic Poem" 30 Tasso's defence of Romance. Pedantic attempts to restrict the compass of Epic
but needs a strong dramatic imagination to keep Romance under control 32
33
III
Romantic Mythology
Mythology not required in the greatest scenes in Homer 35 Myths and popular fancies may be a hindrance to the epic poet, but he is compelled to make some use of them 36 He criticises and selects, and allows the characters of the gods to be modified in relation to the human characters 37 Early humanism and reflexion on myth—two processes: (1) rejection of the grosser myths; (2) refinement of myththrough poetry
40 Two ways of refining myth in poetry—(1) by turning it into mere fancy, and the more ludicrous things into comedy;
(2) by finding an imaginative or an ethical meaning in it
40 Instances in Icelandic literature—Lokasenna 41 Snorri Sturluson, his ironical method in the Edda 42 The old gods rescued from clerical persecution 43 Imaginative treatment of the graver myths—the death of Balder; the Doom of the Gods 43 Difficulties in the attainment of poetical self-command 44 Medieval confusion and distraction 45 Premature "culture" 46 Depreciation of native work in comparison with ancient literature and with theology 47 An Icelandic gentleman's library 47 The whalebone casket 48 Epic not wholly stifled by "useful knowledge" 49
IV
The Three Schools—Teutonic Epic—French Epic—The Icelandic Histories
Early failure of Epic among the Continental Germans 50 Old English Epic invaded by Romance (Lives of Saints, etc.) 50 Old Northern (Icelandic) poetry full of romantic mythology 51 French Epic and Romance contrasted 51 Feudalism in the old French Epic (Chansons de Geste) not unlike the prefeudal "heroic age" 52 But the Chansons de Geste are in many ways "romantic" 53 Comparison of the English Song of Byrhtnoth (Maldon, a.d. 991) with the Chanson de Roland 54 Severity and restraint of Byrhtnoth 55 Mystery and pathos of Roland 56 Iceland and the German heroic
CHAPTER II
THE TEUTONIC EPIC
I
The Tragic Conception
Early German poetry 65 One of the first things certain about it is that it knew the meaning of tragic situations 66 The Death of Ermanaric in Jordanes 66 The story of Alboin in Paulus Diaconus 66 Tragic plots in the extant poems 69 The Death of Ermanaric in the "Poetic Edda" (Hamðismál) 70 Some of the Northern poems show the tragic conception modified by romantic motives, yet without loss of the tragicpurport—Helgi and Sigrun
72 Similar harmony of motives in the Waking of Angantyr 73 Whatever may be wanting, the heroic poetry had no want of tragic plots—the "fables" are sound 74 Value of the abstract plot (Aristotle) 74
II
Scale of the Poems
List of extant poems and fragments in one or other of the older Teutonic languages (German, English, and Northern) inunrhymed alliterative verse
76 Small amount of the extant poetry 78 Supplemented in various ways 79 1. The Western Group (German and English) 79 Amount of story contained in the several poems, and scale of treatment 79 Hildebrand, a short story 80 Finnesburh, (1) the Lambeth fragment (Hickes); and (2) the abstract of the story in Beowulf 81 Finnesburh, a story of (1) wrong and (2) vengeance, like the story of the death of Attila, or of the betrayal of Roland 82 Uncertainty as to the compass of the Finnesburh poem (Lambeth) in its original complete form 84 Waldere, two fragments: the story of Walter of Aquitaine preserved in the Latin Waltharius 84 Plot of Waltharius 84 Place of the Waldere fragments in the story, and probable compass of the whole poem 86 Scale of Maldon
and of Beowulf
89 General resemblance in the themes of these poems—unity of action 89 Development of style, and not neglect of unity nor multiplication of contents, accounts for the difference of length between
earlier and later poems
91 Progress of Epic in England—unlike the history of Icelandic poetry 92 2. The Northern Group 93 The contents of the so-called "Elder Edda" (i.e. Codex Regius 2365, 4to Havn.)
to what extent Epic 93
93 Notes on the contents of the poems, to show their scale; the Lay of Weland 94 Different plan in the Lays of Thor, Þrymskviða and Hymiskviða 95 The Helgi Poems—complications of the text 95 Three separate stories—Helgi Hundingsbane and Sigrun 95 Helgi Hiorvardsson and Swava 98 Helgi and Kara (lost) 99 The story of the Volsungs—the long Lay of Brynhild
contains the whole story in abstract
giving the chief place to the character of Brynhild 100
100
101 The Hell-ride of Brynhild 102 The fragmentary Lay of Brynhild (Brot af Sigurðarkviðu) 103 Poems on the death of Attila—the Lay of Attila (Atlakviða), and the Greenland Poem of Attila (Atlamál) 105 Proportions of the story 105 A third version of the story in the Lament of Oddrun (Oddrúnargrátr) 107 The Death of Ermanaric (Hamðismál) 109 The Northern idylls of the heroines (Oddrun, Gudrun)—the Old Lay of Gudrun, or Gudrun's story to Theodoric 109 The Lay of Gudrun (Guðrúnarkviða)—Gudrun's sorrow for Sigurd 111 The refrain 111 Gudrun's Chain of Woe (Tregrof Guðrúnar) 111 The Ordeal of Gudrun, an episodic lay 111 Poems in dialogue, without narrative—
(1) Dialogues in the common epic measure—Balder's Doom, Dialogues of Sigurd, Angantyr—explanations
in prose, between the dialogues
(2) Dialogues in the gnomic or elegiac measure:
(a) vituperative debates—Lokasenna, Harbarzlióð (in irregular verse), Atli and Rimgerd
(b) Dialogues implying action—The Wooing of Frey (Skírnismál)
112
112
114 Svipdag and Menglad (Grógaldr, Fiölsvinnsmál)
(2) summary, i.e. giving the whole of a long story in abstract, with details of one part of it (Weland, etc.)
118 The second class is unfit for agglutination 119 Also the first, when it is looked into 121 The Teutonic Lays are too individual to be conveniently fused into larger masses of narrative 122
III
Epic and Ballad Poetry
Many of the old epic lays are on the scale of popular ballads 123 Their style is different 124 As may be proved where later ballads have taken up the epic subjects 125 The Danish ballads of Ungen Sveidal (Svipdag and Menglad)and of Sivard (Sigurd and Brynhild) 126
127 The early epic poetry, unlike the ballads, was ambitious and capable of progress 129
IV
The Style of the Poems
Rhetorical art of the alliterative verse 133 English and Norse 134 Different besetting temptations in England and the North 136 English tameness; Norse emphasis and false wit (the Scaldic poetry) 137 Narrative poetry undeveloped in the North; unable to compete with the lyrical forms 137 Lyrical element in Norse narrative 138 Volospá, the greatest of all the Northern poems 139 False heroics; Krákumál (Death-Song of Ragnar Lodbrok) 140 A fresh start, in prose, with no rhetorical encumbrances 141V
The Progress of Epic
Various renderings of the same story due (1) to accidents of tradition and impersonal causes; (2) to calculation andselection of motives by poets, and intentional modification of traditional matter
144 The three versions of the death of Gunnar and Hogni compared—Atlakviða, Atlamál, Oddrúnargrátr 147 Agreement of the three poems in ignoring the German theory
the largest epic work in Northern poetry, and the last of its school 153
155 The "Poetic Edda," a collection of deliberate experiments in poetry and not of casual popular variants 156
VI
Beowulf
Beowulf claims to be a single complete work 158 Want of unity: a story and a sequel 159 More unity in Beowulf than in some Greek epics. The first 2200 lines form a complete story, not ill composed 160 Homeric method of episodes and allusions in Beowulfand Waldere 162
163 Triviality of the main plot in both parts of Beowulf—tragic significance in some of the allusions 165 The characters in Beowulf abstract types 165 The adventures and sentiments commonplace, especially in the fight with the dragon 168 Adventure of Grendel not pure fantasy 169 Grendel's mother more romantic 172 Beowulf is able to give epic dignity to a commonplace set of romantic adventures 173
CHAPTER III
THE ICELANDIC SAGAS
I
Iceland and the Heroic Age
The close of Teutonic Epic—in Germany the old forms were lost, but not the old stories, in the later Middle Ages 179 England kept the alliterative verse through the Middle Ages 180 Heroic themes in Danish ballads, and elsewhere 181 Place of Iceland in the heroic tradition—a new heroic literature in prose 182II
Matter and Form
The Sagas are not pure fiction 184 Difficulty of giving form to genealogical details 185 Miscellaneous incidents 186 Literary value of the historical basis—the characters well known and recognisable 187 The coherent Sagas—the tragic motive 189of Laxdæla
of Egils Saga 190
191
192 Vápnfirðinga Saga, a story of two generations 193 Víga-Glúms Saga, a biography without tragedy 193 Reykdæla Saga 194 Grettis Saga and Gísla Saga clearly worked out 195 Passages of romance in these histories 196 Hrafnkels Saga Freysgoða, a tragic idyll, well proportioned 198 Great differences of scale among the Sagas—analogies with the heroic poems 198
III
The Heroic Ideal
Unheroic matters of fact in the Sagas 200 Heroic characters 201 Heroic rhetoric 203 Danger of exaggeration—Kjartan in Laxdæla 204 The heroic ideal not made too explicit or formal 206IV
Tragic Imagination
Tragic contradictions in the Sagas—Gisli, Njal 207 Fantasy 208 Laxdæla, a reduction of the story of Sigurd and Brynhild to the terms of common life 209 Compare Ibsen's Warriors in Helgeland 209 The Sagas are a late stage in the progress of heroic literature 210 The Northern rationalism 212 Self-restraint and irony 213 The elegiac mood infrequent 215 The story of Howard of Icefirth—ironical pathos 216 The conventional Viking 218 The harmonies of Njálaand of Laxdæla 219
222 The two speeches of Gudrun 223