The Legends of Ermanaric

Caroline Brady (1905–1980) was an American philologist whose scholarship focused on Old English and Old Norse works. Br

857 71 17MB

English Pages X+342 [362] Year 1943

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

The Legends of Ermanaric

  • Commentary
  • A lot of pencil underlining

Table of contents :
I. The Gothic Ermanaric Legend Known to Jordanes 1
Notes to chapter i 18
II. The Norse Jörmunrekkr Legend 23
The texts 26
The "Hamðismal" 26
The "Ragnarsdrapa" 37
"Snorra Edda" 44
The "Völsungasaga" 48
The new relationships, characters, and motifs 51
The story of the slaying of Svanhildr 51
The nucleus: relationship between Jörmunrekkr and Svanhildr; motive for slaying; figures of son and counselor 51
The names Randvér and Bikki 54
The hanging of Randvér 62
The lineage of Svanhildr 65
The story of the vengeance for Svanhildr 68
Change in protagonist: Guðrun becomes the avenger 68
Jónakr 76
Erpr 76
The stoning of Hamðir and Sörli 85
The intervention of Óðinn 91
Development of the Norse Version 99
Notes to chapter ii 109
III. The Version of the Ermanaric Legends Given by Danish Chroniclers 123
Saxo Grammaticus 123
Chroniclers later than Saxo 143
Notes to chapter iii 145
IV. The Ermanaric Traditions Known to Anglo-Saxon Poets 149
Notes to chapter iv 169
V. The German Ermanaric Legends 177
Origins of the Ermanaric traditions current in Germany 183
Ermanaric as Roman Emperor 183
Ermanaric as an avaricious tyrant 184
Ermanaric's dual personality 186
Ermanaric as persecutor of Dietrich von Bern 190
Ermanaric as slayer of his son 195
Ermanaric as slayer of the Harlungs 205
The evil counselor 221
The sickness and death of Ermanaric 234
Swanhild 241
Development of the Ermanaric Legends in Germany 244
Notes to chapter v 263
VI. Transmission of the Legends: Conclusion 271
Notes to chapter vi 298
List of abbreviations 301
Bibliographies 303
Index 327

Citation preview

THE

L E G E N D S

OF

E R M A N A R IC

T h e Legends o f Ermanaric CAROLINE BRADY

University of California Press Berkeley and Los Angeles *943

U N I V E R S I T Y OF C A L I F O R N I A P R E S S B E R K E L E Y A N D L OS A N G E L E S CALIFORNIA

C AM BR I D G E U N I V E R S I T Y PRESS L ONDON, EN GL AN D

C O P Y R IG H T , 1 9 4 3 , BY T H E R E G E N T S O F T H E U N IV E R S IT Y O F C A L IF O R N IA

P R IN T E D IN T H E U N IT E D STA TES O F A M E R IC A BY

TH E

U N IV E R S IT Y

OF

C A L IF O R N IA

PRESS

TO

ARTHUR GILCHRIST BRODEUR A N D

FRANCES MAUD BRADY

Preface yms study wasfirst written as a dissertation in partial satisfaction of the requirementsfor the doctor's degree in English at the University of California. It has, how­ ever>been completely rewritten. Some explanation of the forms I have used for the per­ sonal names may be necessary. When I am speaking of the story in a particular text, I use the forms of the names as they appear in that text: thus jQrmunrekkr is theform ap­ pearing in the Norse texts, Ermrich theform in Dietrichs Flucht; but Ermanaric is the name of the character of the legends in general, without specific reference to any par­ ticular texts. In the course of my studies I have incurred numerous obligations, of which only afew can here be acknowledged. The greatest of these is to Professor Arthur Gilchrist Brodeur, of the University of California, who led me into the realm of Germanic heroes and has accompanied me upon my travels, guiding and assisting me at every step of the long and difficultjourney. Professor James Westfall Thomp­ son, also of the University of California, was likewise ever ready with encouragement and advice, references and books. To Professor Archer Taylor, of the University of California, I am very specially indebtedfor penetrating and most help­ ful criticism. Professor Gustave O. Arlt, of the University of California, Los Angeles, has read several sections of the present manuscript and has given many suggestions and much assistance. It is with gratitude that I recall my studies under Professors John S. P. Tatlock and Clarence Paschali, both of the University of California. Professors Margaret Schlauch, of New York University, and Lee M. Hollander, of the University of Texas, read sections of the manuscript

T

Cvii]

viii

P R EF A C E

in its earlier form and gave me invaluable criticism. Pro­ fessor R. W Chambers, of the University of London, has been most gracious in his encouragement. Professors David K. Bjork and Waldemar Westergaard, of the University of California, Los Angeles, generously placed at my disposal their private libraries. My sister has performed no small share of the chores, including the index, and has been my most constant, and severe, critic and adviser. And finally, without the cooperation and courtesy of the staffs of the li­ braries of the University of California, both at Berkeley and at Los Angeles, especially of Mrs. Enid F Tanner, it would have been difficult to complete this work. cß Davis, California

Contents C HA P T E R

I. The Gothic Ermanaric Legend Known to Jordanes Notes to chapter i II. The Norse Jörmunrekkr Legend The texts 26 The Hamðismal The Ragnarsdrapa Snorra Edda The Völsungasaga 48 The new relationships, characters, and motifs The story of the slaying of Svanhildr The nucleus: relationship between Jörmun­ rekkr and Svanhildr; motive for slaying; fig­ ures of son and co u n se lo r The names Randvér and Bikki The hanging of Randvér The lineage of Svanhildr The story of the vengeance for Svanhildr Change in protagonist: Guðrun becomes the avenger Jónakr E rp r 76 The stoning of Hamðir and Sörli The intervention of Óðinn Development of the Norse Version Notes to chapter ii 1 0 III. The Version of the Ermanaric Legends Given by Danish Chroniclers Saxo Grammaticus Chroniclers later than Saxo Notes to chapter iii

PAGE

i 18 23 26 37 44 51 51 51 54 62 65 68 68 76 85 91 99 9 123 123 143 145

CONTENTS

X CHAPTER

PAGE

IV. The Ermanaric Traditions Known to AngloSaxon Poets Notes to chapter iv V. The German Ermanaric Legends Origins of the Ermanaric traditions current in G erm an y Ermanaric as Roman Emperor Ermanaric as an avaricious tyrant Ermanaric’s dual personality Ermanaric as persecutor of Dietrich von Bern Ermanaric as slayer of his son Ermanaric as slayer of the Harlungs The evil counselor 2 2 1 The sickness and death of Ermanaric 234 Sw anhild 2 4 1 Development of the Ermanaric Legends in Germ any Notes to chapter v VI. Transmission of the Legends: Conclusion Notes to chapter vi 298

149 169 177 183 183 184 186 190 195 205

244 263

271

List of abbreviations

301

Bibliographies

303

Index

327

TH E

LEGENDS

OF

ERMANARIC

CHAPTER I

T h e G othic Ermanaric Legend Known to Jordanes us little about Ermanaric, King of jK e Goths. B yh is many warlike deeds he mäHe Tirnself formidable to all the neighboring nations; his territories were invaded by the Huns, who had con­ quered and allied with themselves those Álani who bor­ dered on the Ostrogoths (Greuthungi) and were called Tanaitaeplor aHong^ time he encleavored to hold his ground against the invaders; butat last^rumor exagger­ ating the impending dangers, he ended his fears by a voluntary death, about: a .d. 375. $0 much is history, as t^Iatedby the singularly trustworthy Roman historian and contemporary ofErmanari^AmmianusMarcellinus.1 Approximately one hundred and seventy-five years later the Gothic historian Jordanes, in his De origine actibusque Getarum (or Getica), wrote of the deeds and death of $\Hermanaricus, nobilissimus Amalorum, the I -Gothic Alexander.2By sagacity and valor Hermanaricus subjugated many peoples,, amojng„t,hein..theJJerulL the Venethi, and jthe Aesti of the Baltic coast; and by his own efforts he ruled over_all the nations of Scythia and Germany.3 The subjection of one of these tribes, the fleruli7.Tordanes especially emphasizes: Hermanaricus was not content until he had slaughtered a large part of the tribe and subjugated the remainder.4As a result of tliese many conquests Hermanaricus was ruling over a istory tells

H

T

'

1 For notes see pages 1 8-22.

ill

~ ÍT''

-

t

2

THE LEGENDS OF ERMANARIC

tremendous empire. extending apparently from the Black Sea to the Baltic, at the time when the Huns appeared on the eastern boundaries of his kingdom. While he was deliberating how he might meet the roe, one tribe which owed him homage, the:Rosomoni~seized the opportunity to betrayhimT^ Moved"Ey furyHiecausedtobel:orn~to pieces by wild*horses a certain woman of the Rosomoni, Sunilcfa by name, on account of the treacherous defec­ tion orner husband (pro mariti fraudulento discessu).Her brothers, Sarus and Ammius, in vengeance for the slay­ ing of their sister, attacked Hermanaricus and wounded him severely in the side, so that he lived on in infirmity oY body. Balamber, king of the Huns, taking advantage oftne infirmity of Hermanaricus, moved his battle array into the territory of the Ostrogoths, from whom the Visi­ goths, because of contention among themselves, had sep­ arated. Meanwhile, Hermanaricus, unable to endure the pain of his wound ~or the incursions of the Huns, died, full of days7 in the hundred and tenth year of his liieT6 Jordanes agrees with Ammianus as to the two main features of Ermanaric’s career: he made himself formi­ dable to neighboring nations by his warlike deeds; he died at the time when the Huns invaded Ostrogothic territory. But Jordanes differs on one point of major im­ portance: he makes no mention of suicide. And he gives a wealth of detail not to be found in Ammianus’ account: the names of the peoples Ermanaric conquered, the treach­ ery of the Rosomoni, the slaying of Sunilda, and the vengeance of her brothers on Ermanaric. In all probability, Jordanes’ recital of Ermanaric’s conquests is no more than an enumeration of those war­ like deeds by which, according to Ammianus, the great Ostrogoth made himself formidable to all the neighbor­

T H E G O T H I C L EGE ND

3

ing nations. To be sure, the point of view of the Getica is pro-Gothic. Jordanes was a Goth.7 And his source (of wmclThis own work isTIttle more thaiTan epitome8), the nowTost Gothicliistory of Cassiodorus. was intended to ^orífy tHgGothic people and the royal line of the Amali.9 Cassiodorus, by virtue of his long association with Theodoric the Ostrogoth, was definitely of Gothic sympathies; and his sources were not only various Greek and Latin literary documents,10 but also almost certainly Golhic oral tradition.11 Nonetheless, Jordanes’ glowing account oTErm anaric’s conquests is wholly consistent with Am­ mianus’ more concise statement; and this fact suggests that although Jordanes’ version may be colored by Gothic enthusiasm for the Gothic past, it is not story, but his­ tory .'I t is not at all unlikely, however, that in these deeds which he enumerates, especiallyinthe^onguestof the Heruli, lies the kernel of a legendary tradition which in the Middle Ages becomes especially well known in Germany and in England: that of Ermanaric as the “ colossal type of a ferocious and covetous ruler.” 12 Jordanes’ account of the circumstances of Ermanaric’s death likewise might be simply elaboration of the events reported more concisely by Ammianus. The two agree strikingly in their main outlines. In both, an act of a third and lesser tribe encourages the Huns to invade the territory of the Goths: in Ammianus, the alliance of the Alani of the Don with the Huns; in Jordanes, the treacher^ of the Rosomoni. In both, Ermanaric tries to ward off the invasion; in both, worry over the invasion and the evils resulting from it are responsible, in greater or less degree, for his death. In the details, however, the two accounts differ materially: Jordanes says nothing of suicide; Ammianus makes no mention of the great age of

4

T H E LEGENDS OF E R M A N A R I C

Ermanaric, the perfidy of the Rosomoni, Sunilda and her husband, her brothers and their attack on Ermanaric.This wide divergence immediately raises the question whether the details peculiar to Jordanes have been drawn from fact or fancy, whether they are history or fiction. Some critics have suggested that these details given by Jordanes but not by Ammianus could be, at least in part, historical;13 that Ammianus may have heard noth­ ing of them, that to his ears would have come only the events most important to a foreign historian: the Hunnish invasion and the fall of the Gothic kingdom.14Others have maintained that the entire Sunilda story is deliber­ ate invention, its only historical basis the undoubted fact that Ermanaric died; that the Rosomoni, Sunilda and her husband and her brothers, the attack on Er­ manaric— all these are pure fiction; and that even the names are “ fictiv-episch,” “ sinnvoll,” names based on the roles the characters play in the story.15 A careful consideration both of the tale as Jordanes tells it and of the circumstances which most certainly gave rise to that story, however, leads rather to the conclusion that the Sunilda story is neither pure history nor pure fiction, but legend. The assumption that it is not in its entirety pure his­ tory is well founded. Undoubtedly, Jordanes— or Cassiodorus before him— derived it from Gothic oral tradition; certainly there is no mention of it in any of the known written sources of his account of the Hunnish invasion: Priscus, Orosius^ and Ammianus.16 Presumably, there­ fore, Tt reflects The Gothic point of view. Presumably, also, ä story which for a hundred and fifty years and more had been handed down by word of mouth would admit some infiltration of nonhistorical elements. The

T H E G O T H I C L EGE ND

5

death of Ermanaric at the moment when the Huns were invading the Gothic realm presented two extraordinary HÍtuations, the political situation of the sudden break-up of the gigantic Gothic kingdom and the personal situa­ tion of the suicide of the Gothic king in the face of danger. Both would lend themselves to legendary ex­ planation: a great defeat always demands an explanation that will assuage the humiliation of the defeated people; suicide of a Germanic king, especially in an hour of na­ tional crisis, was so unprecedented that a Germanic people, and Germanic poets, would have found it inex­ plicable and would inevitably have sought and, in the course of time, have supplied some other, less unheroic reason for the great Ermanaric's death. The Sunilda story may very well have arisen as an explanation of these unparalleled situations.17 Not the least common exffianalmnadvanced by a people smarting under the sKamePof catastrophic defeat is treachery on the part of subjects or allies: in the Sunilda story we find the treach­ ery of thelCosomoni, a tribe which owed allegiance to the Goths. Again, as ordinary a manner of death as might be found in the age of the migrations was death at the hands of men avenging slain kin: the Ermanaric of Jor­ danes js struck down by the brothers of the slain Sunilda. It is not certain, however, that these explanations are wholly fictitious, that they were drawn solely from the imagination of Gothic storytellers, or, as some critics would have it, from myth.18On the contrary, the Sunilda story as Jordanes tells it gives every indication of con­ taining at leas t a kernel of historical truth.19 In the first place, it is political in tone, “ a political anecdote/'20 It presents both a political and a personal situation. The immediate motive for the attack of Sarus

6

T H E LEGENDS OF E R M A N A R I C

and Ammius upon Ermanaric is the personal one of ven­ geance for slain kin. But the prime motive, the detail which furnishes the basis for the entire story, the cd e ­ ceitful defection'' of Sunilda’s husband, is definitely po­ litical. T he slaying of the woman Sunilda bridge~s~~the gap, as itjwere, between the^two: it is political in that it is the result of the treachery of subject to overlord; it is personal in that it affords the motive for the attack of the brothers. This emphasis upon the political indicates that the Sunilda story is not pure fiction, but is, at least in part, rooted in history.'Torwhen, in an early version of a story, political aspects^are emphasize^ the chances are that here is a inflection of the actual political situation giving rise to thejstor n later versions political elements are modified, indeed sometimes disappear en­ tirely, and personal aspects— personal motives, personal conflicts, personal relationships of every kind— receive the emphasis.22 An excellent illustration is afforded by the Ingeld story. In the earliest of its extant versions, the Anglo-Saxon, the political events are represented as they must actually have been: a feud, lasting over several generations, between the two peoples HeaSobeardan and Danes, which was halted temporarily by the mar­ riage of the HeaSobeard prince Ingeld to the Danish princess Freawaru, only to break out afresh and to be ended at last by the slaying of Ingeld and the crushing of HeaSobeard power forever. In Scandinavian tradi­ tion, on the other hand, the HeaSobeardan have been forgotten, and the political aspects of the original story have been so far changed that Ingeld and his father Fröda have themselves become Danish kings. In Saxo political aspects have been completely transformed; the feud is represented as one between the Danes and their

T H E G O T H I C L EGE ND

7

Saxon neighbors, and Ingeld, originally leader of the de­ feated HeatSobeardan, appears as the victor.23 But in West Norse tradition the political has given way to the personal, and the feud has become one between two branches of the Danish 7oyaI line.24 In shortT as Axel Ölrík has demonstrated, in the Beowulf and in the Widstð “ de politiske begivenheder optræder med det virkelige livs mangfoldighed; sagnet er ikke lagt til rette til for­ herligelse af visse få yndlingsskikkelser; her kæmper også folkestammer, hvor den friere sagndigtning vilde i udføre enkelthelte.” 25Precisely because the Anglo-Saxon poems do represent tfie feud as “ a national business,” mey~are generally regarded as depicting Taithfully— much more faithfully than the later Scandinavian versions— the historical events that gave rise to the legend.26 Now, the development of this legend of the HeaSobeards Ingeld and Fröda parallels, in many respects, the de­ velopment of the Swanhild legend. Therefore, when in the earliest extant version of the Sunilda story we find a political motivation, when we find Jordanes deeply con­ cerned with the political consequences of the treachery of the Rosomoni, the chances are that this part of the story, at least, is rooted in history. The matter depends largely on the identity of the Rosomoni and on what constituted their treachery. Jor­ danes’ presentation of the story is perhaps not so clear as might have been desired: he states that the faithless race of the Rosomoni took advantage of the Hunnish invasion to betray Ermanaric (Rosomonorum gens infida . .. tali eum nanciscitur occasione decipere) and then pro­ ceeds immediately to tell of the slaying of Sunilda and the vengeance of Sarus and Ammius, bringing the defec­ tion of the husband almost parenthetically into his

8

T H E LEGENDS OF E R M A N A R I C

account of her slaying. The result has been a certain amount of speculation respecting what constituted the treachery of the Rosomoni, the identity of Sunilda’s husband, and the nature of his “ deceitful defection.” At first glance, the implication seems to be that the treachery of the Rosomoni was the attack of Sarus and Ammius; and so some scholars have taken it.27 This attack, however, is but the third in a series of events, the ultimate consequence of an incident which Jordanes specifically designates as deceitful— “ the deceitful de­ fection” {fraudulentus discessus) of Sunilda’s husband. This was the motive for the slaying of Sunilda, the moti­ vating incident of the whole story; for the tale clearly consists of three incidents: the “ deceitful defection” of the husband, the slaying of the woman, and the attack of the brothers on Ermanaric. Jordanes does not make dear the identity of the husband, but the most plausible conjecture is that he was prince or leader of his tribe.28 His “ deceitful defection” may have been revolt, as some critics have regarded it;29 or it may have been desertion to the Huns, as Boer has suggested.30The implication is that the man, owing support to Ermanaric, treacher­ ously betook himself and his support elsewhere, and, since he took advantage of the Hunnish invasion to do so, the presumption would seem to be that he took his support to the Huns. However, it does not seem likely that Tordanes meant to identify the “ deceitful defection” of the husband with the treachery of the Rosomoni. As Boer has pointed out, the subject of nanciscitur occa­ sione decipere is the entire Rosomonorum gens infida,31 The act of Sunilda’s husband is simply a specific and striking instance of the national treachery, just as the attack by Sarus and Ammius is another instance of it.

T H E G O T H I C L EGEND

9

The identity of the Rosomoni has never been satis­ factorily determined, although various identifications have been suggested. A number of critics regard the name as “ epic,” some even having tried to “ interpret” it.32 Others have attempted to identify the tribe with some historical people and have concluded that they were: the Roxalani;33Slavs;34the Heruli.35 The identification with the Roxalani, though now almost universally rejected,36 is nonetheless the most tempting. For these were, at one time at least, neighbors of the Goths, and they were a branch of a people which had long had relations with the Goths and which, like the Rosomoni, played a decisive role in the Hunnish invasion.37 Ammianus testifies that the Huns passed through the territories of those Alani who bordered on the Ostrogoths and were called the Tanaitae, slew many, and made an alliance with those who survived.The Alani themselves took an active part in the attacks on the Goths: Ammianus states (X X X I, 3. i) that in company with the subjugated Alani the Huns made a sudden in­ road into Ermanaric’s territory, and, further (X X X I, 3. 3), that Ermanaric’s successor, Vithimiris, with the aid of Huns whom he bribed, resisted the Alani but at last was slain in battle. A t the time of the Hunnish inva­ sion, then, the Alani were allied with the Huns against the Goths and took an active part against them. Ammianus emphasizes, too, the Hunnish conquest of those Alani called Tanaitae. Indeed, it was the alliance with them which encouraged the Huns to invade Ostrogothic territory and thus, indirectly at least, brought about Ermanaric’s death. Presumably, these were the Roxalani. According to Jordanes (XII, 74), ancient Gothia was bordered on the east by the Aroxolani;38

IO

T H E LEGENDS OF E R MA N A R I C

Strabo (VII, 3. 17) places the Roxalani between the Dnieper and the Don; Ammianus (X X II, 8. 31), around the marshes east of the Maeotie gulf. If the Tanaitae of Ammianus are the Roxalani, then the identification of the Rosomoni as the Roxalani gains strong support. There are objections to the identification. Possibly the most weighty39 is the difference in the names: Rosomoni is not an expected variant of Roxalani, for we have an s where we expect x, and m where we expect /. The occurrence of s for x is easy enough to explain; we have other instances of it. Jordanes in an earlier chapter of the Getica (XII, 74) mentions the Roxalani twice: in the first passage, two codices, the Canterbury and the Berlin, read arozolani; in the second, the Codex Ambrosianus reads arosolanis. The real difficulty is the m where we should have /, but of this there may be an explanation: in all probability Jordanes— or, rather, Cassiodorus— derived the story of the Rosomoni from oral Gothic tra­ dition; and, as Boer has suggested,40in oral tradition the name Roxalani (Aroxolani) could have been distorted to Rosomoni. In any case, despite the fact that the identi­ fication of the names is uncertain, even perilous, the difference in them cannot be held as sufficient reason for rejecting the identification of the two tribes if it can be supported on other grounds. There are other differences between Ammianus’ ac­ count and Jordanes’, aside from this fact that Ammianus speaks of the Tanaitae, who presumably but by no means certainly are the Roxalani, and that Jordanes speaks of the Rosomoni, who may or may not be the Roxalani. Ammianus tells that the Tanaitae were conquered by the Huns and made an alliance with them: Tordanes speaks of the treachery oX-the Rosomoni. Ammianus

T H E G O T H I C L EGE ND

II

writes in terms of tribes: the Tanaitae submit to the THuns^ and the Alani attack theTroths. Iordanes tells his story in terms of individuals^ having condemned the entire Rosomonorum zens for treachery, he places upon individuals the responsibility: Sunilda’s husband is guilty oT^Heceitful defection” ; her two brothers attack Ermanaric. The third difference is the personal motivation of the attack by Sarus and Ammius: the two brothers are avenging their sister. This of course involves the fourth difference: the slaying of Sunilda, an event not paralleled in Ammianus’ account of the Tanaitae. Every one of these differences, however, simply points to the conclusion that the Sunilda story is neither pure fiction nor pure history, but contains a kernel of histor­ ical truth modified and embellished by ä considerable nonhistoricald e m e ^ it is heroicTegend. The first is precisely what we should expect if the Rosomoni are to be identified with Ammianus’ Alani of the Don and if in fact Jordanes— or Cassiodorus— did draw the story of their treachery from Gothic oral tra­ dition, colored as it must inevitably have been by the Gothic point of view and the Gothic emotional reaction to the fall of the kingdom and the death of the king in a moment of national crisis. A diligent, responsible his­ torian will attempt to give solid history, an accurate and impartial picture of the events he is describing; but his success will depend in large part upon the adequacy of his source material. When the only accessible source is oral tradition, accuracy and impartiality become im­ possible. Faulty recollection, national feeling, rationali­ zation, the dramatic necessities of the poets who have handed the story down from generation to generation: all these elements lead to distortion of the events de-

12

T H E LEGENDS OF E R M A N A R I C

scribed. The point of view and emotions of a people which has experienced a catastrophe will inevitably be reflected by its poets in their stories of that catastrophe. And the distortion will become progressively greater as time passes and generations of storytellers exercise their increasing effect upon the representation of the national tragedy. To be sure, the formula “ legend is history colored by emotion” is no more true, without very considerable qualification, than is the formula “ legend is the sum of history plus fiction.” But emotion does play its part in transforming history into legend. Now, from the point of view of an impartial historian the submission of the Alani to the Huns would be simply a fact not to be judged. The Ostrogoths, however, were no detached his­ torians, looking at the matter in the cold light of reason; they were a people smarting under a great defeat, look­ ing around for someone to blame. That they should blame the Alani would have been natural enough: the “ Tanaitae, who bordered on the Ostrogoths,” stood as a buffer between Goths and Huns, and not until they sub­ mitted did the Huns attempt to invade the territories of Ermanaric; had they not given in, but held their ground a little longer, the Goths might not have been defeated, for Ermanaric was even then deliberating upon means of averting disaster. Furthermore, these subjugated Alani actually accompanied the Huns when they invaded Ostrogothic territory. And after the death of Ermanaric the Alani waged war against the Goths and were, so far as we can judge from Ammianus’ words, in no small part responsible for the ultimate defeat of the Gothic forces. From the Ostrogoths’ point of view the submission of the Alani, especially of the Tanaitae, could very well have been deceitful defection. This, then, is the first of

T H E G O T H I C L EGE ND

*3

several indications that in Jordanes’ Sunilda story legend has encroached upon history: the appearance of the point of view of one of the peoples affected by the event. Though in Jordanes’ account the emphasis is on the treachery of the Rosomoni, individuals exemplify by their behavior the national treachery. Again this is pre­ cisely what we should expect; it is simply another indi­ cation that such history as may lie behind the Sunilda story is well on its way toward becoming legend.41 “ One of the principles pointed out in Olrik’s great book [Dan­ marks Heltedigtning},” says Chambers, “ is precisely this: the earlier accounts represent the struggles of the time of the Migrations as being what they actually were— struggles between peoples >In later versions this political aspect is dimmed: the story is one of struggles between individuals, the nations are forgotten and confused.” 42 The Ingeld story, already referred to,43 furnishes an apt example of this “ law” of legend development. The sec­ ond part of the Nibelungen story affords another. About the year 437 the Burgundians under Gundaharius met the H unsjn battle, and Gundaharius and all his warriors were slain: this is the historical fact.4>d t is the Burgundians and Huns who are important here, not Gundaharius nor the Hunnish leader whose name the contempo­ rary historians do not even record. In the course of centuries, however, the story so develops as to glorify the individual heroes, in complete disregard of their po­ litical relations; of all the extant Norse texts recounting the tale, for instance, only the Atlakviða (18. 3) remem­ bers Gundaharius’ (Gunnar’s) Burgundian connections. This last stage, when the nations have been forgotten and the story is concerned solely with the individual heroes, has not been reached in Jordanes: on the con­

14

the legends of ermanaric

trary, the Rosomoni are still very well remembered. But individuals already play a distinct part. The remaining differences between Ammianus' account and Jordanes'— the attack of Sarus and Ammius, with its motive of vengeance, and the slaying of Sunilda— further illustrate the principle that in oral tradition con­ crete political events are, in time, interpreted in personal terms, and further indicate that the Sunilda story is neither pure history nor pure fiction, but legend.JT the story of the treachery of the Rosomoni is in fact an ex­ planation, in terms of Gothic national feeling, of the Gothic defeat, then it would as inevitably have passed ir ^ l K Prealm of Germanic heroic^poetry as did, lor instance, the story of that other ca^as^p^hicjdefeatof a Germanic people by the Hunnish invaders^ that of the Burgundians of Gundaharius. And once it did, it would unmistakably begin to take on the characteristics of heroic poetry; it would begin to borrow acts and mo­ tives, figures and themes from the common stock of that poetry; it would lose its political aspects; it would take on personal motives and personal relationships. “ It is an essential characteristic of heroic poetry," Chambers writes, “ t h a t . . . it gives the story modified almost past recognition by generations of poetic tradition."45 For, as Heusler has demonstrated admirably, heroic poetry is not simply history in verse: “ Der Hauptgrund ist das Bej ahende: dass germanische Heldendichtung den Wirk­ lichkeitsstoff in neue Gussformen bringt; dass ihr Motiv­ schatz auf das Privat-Menschliche eingestellt ist und sich abschliesst gegen Staatskurist und Kriegführung, gegen Völkergeschicke, Volkstum und Glauben."46 The Sunilda story as Jordanes sets it ,down for us, political in tone though it may be, contains one of the favorite

T H E G O T H I C L E GE ND

*5

themes of Germanic heroic poetry: vengeance.Jt con­ tains a dearly loved motive: vengeance for slain kin. I t contains typical characters of heroic poetry: young war­ riors seeking vengeance for slain kin« Furthermore, the death of Ermanaric presented not one unparalleled situation, but two, not only the political situation of the downfall of the Gothic kingdom, but also the personal situation of the suicide of a Germanic king. It may be that some Gothic poet saw all the tragedy of the old man, of his valiant attempt to hold his ground, of his grief at the evils impending over his country, of his weakness in slaying himself when his people most needed him. But heroic poetry did not deal with the sorrows of old men alone in their tents; it did not deal with the conflict in the soul of a man driven mad with fear. It dealt rather with action, with violent pas­ sions, with vengeance, with all the human relationships of a people in the midst of war; it dealt with the conflict between man and man upon the field of combat; it dealt with the conflict in the soul of a man forced to choose between two duties. The slaying of Sunilda in a moment of fury, the vengeance of her brothers: these are the stuff of poetry. Sunilda, Sarus, and Ammius are surely no his­ torical characters; they came into being because some Gothic jpoets were trying to account for the death of Ermanaric, nobilissimus Amalorum. That part of the Sunilda story, then, which deals with political aspects— the “ deceitful defection” of Sunilda’s husband— may very well have arisen to explain the Gothic defeat; the attack of Sarus and Ammius to serve as the reason for the death of Ermanaric; and the slaying of Sunilda to serve as the link between the two. This is, of course, an oversimplified statement of the development

16

T H E LEGENDS OF E R M A N A R I C

of the Sunilda story. But to determine precisely how it evolved, to decide just how and when this element and that was incorporated, is impossible.47 In the first place, the tale is preserved only in completed form, in which the two explanations already appear as a connected narra­ tive, that of the fall of the, kingdom serving as motiva­ tion for that of the death of the king. In the second place, the story did not grow up in a day; it is not the work of a single poet, plotting out in its entirety a good tale, but of generations of poets, each of whom indubi­ tably made additions and changes, slight or great— always, to be sure, in terms of other, similar stories and of the codes and modes of behavior in the heroic age. Thus, we may guess that to Gothic poets the suicide of the mighty Ermanaric, the Gothic Alexander, meant that he was no longer himself, that he must have been sick, or very old, or both; and that the cause of his sick­ ness was conceived of as an attack by enemies in the course of which he was wounded. We can be fairly cer­ tain that Sunilda, Sarus, and Ammius owe their existence to the fact that an act, to appear at all reasonable, must have a motivation and must be performed by an actor; for otherwise there could be no story, but only abstrac­ tions. In other words, the death— or sickness— of Er­ manaric had been explained as the result of a wound: someone must have given the wound, and he must have had a reason. Gothic poets, at some time in the course of development of the story, would inevitably associate the assassin with that “ perfidious race” whose treacherous adherence to the national enemy was credited with caus­ ing the national catastrophe. And they would quite na­ turally represent the attempted murder as an act of vengeance for kin; for this was a theme commonplace in

T H E G O T H I C L EGE ND

*7

Germanic heroic poetry and having its roots in Germanic social life. Perhaps the representation of the avengers as two brothers likewise was derived from the common stock of Germanic tradition; such pairs of avenging brothers (or kinsmen) are common at least in later legend: I Iróarr and Helgi, Sigmundr and SinfjQtli, Keto and Wigo. Farther than this we cannot go. Krom this investigation of Jordanes’ account of the death of Ermanaric we may state a few definite conclu­ sions and suggest various others. Certain it is that Jor­ danes knew a story which Ammianus did not know concerning the events which preceded the death of Er­ manaric. Probably as certainly, this story was not wholly historical, but a tale explaining the death of the Gothic king in a moment of national crisis and the downfall of a tremendous kingdom which extended from the Black Sea to the Baltic and embraced many tributary tribes.We may reasonably conjecture that this story was not pure fiction, but legend with a historical kernel— a kernel of history which became reshaped by Gothic emotions in the course of Gothic oral transmission, and embellished by the addition of themes and characters and motives drawn from the stock of Germanic heroic poetry. We may also conjecture— how reasonably must remain a matter of opinion— that the historical basis of the Sunilda story was the conduct of the Alani, particularly the Alani of the Don, at the time of the Hunnish invasion of Ostrogothic territory. And one final definite thing we know: fiction or legend, the Sunilda story told by Jordanes is political in tone, political in motivation; this story is the political tragedy of the nation,, not the personal tragedy of theJung_which we find when we meet_the Sunilda story in later versions of the Ermanaric legends,

N O TES TO C H A P T E R I 1 R e ru m gestarum lib r i q u i su p ersu n t, X X X I , 3 .1 - 2 . 2 G ^ V t f ,X X I I I - X X I V . 3 I b id ., X X I I I , 116-120 . 4 I b id ., X X I I I , 1 1 7 -1 1 8 . 6 I b id .) X X I V , 1 29. 6 Ibid .y X X I V , 1 29—130. 7 I b id ., L X , 316. 8 Ibid .y Prcef. Jordanes states that he wrote his abridgment o f Cassiodorus’ book largely from memory, and he maintains that he himself added certain co n v en ien tia from Greek and Latin histories besides an introduction, con­ clusion, and p lu r a in m edio mea d ictio n e. C f., however, Mierow, T h e G o th ic H isto ry o f J o rd a n es, pp. 13 if., 25 if., and the references there cited. 9 Cassiodorus, Variae, I X , 25. 10 C f. Mom m sen, P rooem ivm to his edition o f Jordanes, pp. x x x if.; Mierow, O p.

CÍt.y pp. 19 ff.

11 C f. Mierow, loc. c it. 12 M üllenhoff, Z fd A , X , 176 f. C f. below, chap, iv, pp. 164 ff., and chap, v, p p . 1 84 ff. 13 Cf. especially Heinzel, W S B , C X I X , 8; Panzer, D eutsch e H eld en sa g e im BreisgaUy pp. 24, 40; Boer, D ie Sagen von E r m a n a r ic h u n d D ie tr ic h von B e r n , pp. 4 ff.; Chambers, B eo w u lf: A n In tro d u ctio n to the S tu d y o f the P o e m , p. 444 (cf. also W idsithy pp. 17 f.). 14 So Boer, op. c it., p. 6. 15 Cf. especially Roediger, Z dV erf Volks., I, 247 ff.; Jiriczek, D H S , I, 62 ff.; Sijmons, Grdr.y I I I , 683 f.; and the references cited by them. ■ Especially to be noted, because o f its importance in analysis o f later ver­ sions o f the story, is the argument that the names S a ru s ( < G o t h . sarw a, “ w eapon,” “ armor” ) and A m m iu s ( < G o t h . *h am a : g a -ham on , “ to arm oneself” ) were given to the brothers because they had in their possession impenetrable (or magic) armor; thus Heusler, H o o p s , I, 627: “ D ie Ursprüng­ lichkeit dieses M o tivs [the armor] folgt daraus, dass die Nam en b eid er Helden au f die R üstung zielen {ham a-, sarw a- ) : ein klarer Fall sinnvoller N am en­ gebung.” C f. also Rassmann, D ie deutsche H eld en sa g e u n d ihre H e im a t, I, 262, n. 3; Sijmons, G rd r., I I I , 683; Detter-H einzei, Scem undar E d d a , II , 583; Jiriczek, D H S , I, 64 (not inconceivable); and others. This theory is w holly unfounded. Jordanes says not one word about the armor o f Sarus and Am m ius; the only texts which do mention it are the thirteenth-century Norse S n o rra E d d a and VQlsungasaga (cf. below, chap, ii, pp. 68 ff.). We sim ply can­ not use the evidence o f a thirteenth-century Norse text to demonstrate w hat was known to a sixth-century G othic historian. Indeed, any argument based on the names S a r u s and A m m iu s is perilous. A lthough the suggested deriva-

Hi8]

T H E G O T H I C L EGEND

X9

tions doubtless are correct, they are not evidence that the names are based on the roles played b y the characters: as Schneider has pointed out (G H S , I, 247), words pertaining to battle and weapons were common elements in the formation o f Germanic personal names. Æ sch e re, H eorogär, Hrö “ the masts of the sail of Naglfari” ) stand encircling the king’s couch44— stand “ without a seam” {saums andvanar^), that is, presumably, with locked shields,protecting theirmutilatedleader,46— andstraightway “ upon common agreement,” “ by common action,”

T H E NORSE L EGE ND

39

or “ as the result of counsel” (samraða)i7 HamcSir and SQrli were beaten to the earth with hard stones (str. 5). The “ impeller of the eddy of steels” (ståla stjkkvir flaumsY8 caused the brothers to be warded off, and all reward the sons of Jonakr with sword blows (str. 6).49 This account can scarcely be regarded as a continuous narrative; any possibility that it could be is precluded by the exigencies of the framework Bragi has adopted for his poem. The fall of HamSir and S q H í , he says (str. 7), is painted on the shield of Ragnarr LoSbrók, and he is simply setting down in verse what he has seen here. Now, the story of the attack of the brothers upon jQrmunrekkr consists, essentially, of three incidents: the mutilation of JQrmunrekkr; the Gothic attack with weapons, which prove useless against the two brothers; and the stoning of the two. These three incidents clearly were depicted on the shield, each in a separate picture; and each picture is described in a separate strophe of the Ragnarsdrápa (strs. 4-6). Bragi adds an introduction to explain the circumstances of the attack (str. 3), and a conclusion to make clear that he has, after all, simply described what he saw on the shield (str. 7). With such a framework it would have been impossible to add ex­ planatory matter which we might well wish for, or to establish the time sequence of the several incidents which took place during the battle in the hall. In the first place, the limitations of pictorial delinea­ tion might demand two, or more, separate pictures to depict action actually taking place simultaneously. The most striking example is the situation described in stro­ phes 4 and 5. The brothers have attacked, JQrmunrekkr has been mutilated, the house troops have rushed to the defense of their king and are now standing around him,

40

T H E LEGENDS OF E R M A N A R I C

protecting him from further assault and throwing stones at HamSir and SQrli. But all this could scarcely be de­ picted in a single picture: if the Gothic warriors stand “ without a seam” encircling the king’s couch, how are we (and the poet) to know that JQrmunrekkr has fallen on his head in a pool of blood and ale and that his severed hands and feet are lying on the floor ?Obviously, two pictures were necessary. Again, the sequence of the strophes presumably repre­ sents not the sequence of the action, but the sequence of the pictures on the shield, or even perhaps the order in which Bragi looked at them. Strophe 5, for example, de­ scribes the Goths as beating Hamc5ir and SQrli to the earth with stones; strophe 6 represents them as incited by stála stfkkvir flaums to attack the brothers with sharp-cutting weapons. Now, are we to assume that the attack with weapons followed the stoning, as the strophe order would lead us to believe? Certainly not: the Hamðismál, the Vqlsungasaga, and Snorra Edda all represent the Goths as resorting to stones only after they have found weapons useless against the brothers. Ordinarily, it is perilous to interpret one of the extant Norse texts in terms of the others: both Snorra Edda and the Vqlsunga­ saga were compiled centuries after the composition of the Ragnarsdrápa; and the Hamðismály although of ap­ proximately the same date as the Ragnarsdrapa, differs from it in several details. In this particular instance, however, ordinary scruples may be ignored: the Rag­ narsdrapa was the primary source for Snorri’s account of the attack; and we should therefore expect the two versions to agree more or less closely. According to Snorri, JQrmunrekkr awakened from sleep to find his hands and feet cut off; he cried out and roused his men, who fell on

T H E NORSE LEGE ND

41

HamSir and SQrli with weapons, but were unable to overmaster them; he then called out instructions to beat the brothers down with stones. Bragi agrees, on the whole, if we disregard strophe order:50 JQrmunrekkr awakened (str. 3), presumably to find his feet and hands lopped off (str. 4), and urged his men to attack HamSir and SQrli (str. 6; the kenning stála stjkkvirflaums is per­ fectly applicable to JQrmunrekkr, for to the skald a king is leader of warriors, director of the fray); under his di­ rection the Goths warded off (from his couch?) the two brothers (str. 6), but were unable to overmaster them with weapons (by implication); JQrmunrekkr plunged headlong into the gore on the floor, and his men stood encircling his couch and pelted HamSir and SqHí with stones (str. 5). The Ragnarsdråpa account is in no way inconsistent with the representation, in the other extant Norse versions, that the attack of HamSir and SQrli con­ sists of three incidents: the mutilation of JQrmunrekkr, the unsuccessful Gothic attack with weapons, the ston­ ing of the brothers. There does exist one major difference between Bragi’s version and Snorri’s: whereas Snorri specifically states that JQrmunrekkr instructed the Goths to stone HamSir and SQrli, Bragi makes no mention of instructions but describes the circumstances of the stoning simply by the term samråda.51 Unfortunately, this is an ambiguous word. It may mean “ by common action,” 52 that is, that the Goths with one accord reached out for stones when they found swords useless against the brothers. Or it may mean “ by common agreement,” but, thus inter­ preted, it fails to fit the situation which Bragi is describ­ ing: JQrmunrekkr has awakened, presumably, to find his couch streaming with blood, his hands and feet lying on

42

T H E LEGENDS OF E R M A N A R I C

the floor in the blood and ale, his hall in an uproar; the Goths are engaged in hot combat, vainly endeavoring to ward off the attackers. Strange circumstances, these, for JQrmunrekkr to be consulting with his warriors on the best course of action, or the Goths to be pausing to agree among themselves! Again, samråda may mean “ as the result of counsel” ; the element rdð implies some such meaning. Yet this interpretation would seem to imply also that there must have been present someone who was in a position to give counsel. And Bragi gives no indication who this individual may have been. He may have known the tradition which is set forth in both the Hamðismál and the VQlsungasaga: that OSinn inter­ vened in the battle in the hall and gave the advice to stone the brothers. But of this we cannot be certain. Moreover, the Ragnarsdrapa account differs in several essential details from that of the HamðismáL It tells of a surprise attack, while the Goths sleep, whereas accord­ ing to the Hamðismál watchers on the wall warn JQr­ munrekkr, who is sitting drinking with his men in the hall. HamSir’s taunt of the wounded JQrmunrekkr, the final dialogue between the two brothers, HamSir’s be­ lated recognition of Erpr’s vital importance to success in their vengeance: all these highly dramatic details of the Hamðismál account of the attack are wholly lacking in the Ragnarsdrapa. In spite of these differences between the two poems, however, a number of critics have argued that one of the two has borrowed directly from the other.53This opinion is based solely upon a certain similarity between three consecutive half-lines of the Hamðismál: styrr uarð i rannij stukko qlskálirj/í blóði bragnar lágo (23. 1-3), and three half-lines, not consecutive, of the Ragnars-

T H E NORSE L EGE ND

43

drápa: rósta varð t ranni (3. f)J e ll í blóði blandinn/brunn qlskála— runna (4. 5-6). As a general rule, however, we are justified in stating that a poet has borrowed from another a word or phrase when that word or phrase is identical in the two poems, when it occurs only in them, or when a specialized meaning or use of it occurs only in them. For the passages in question, these conditions do not exist. The half-lines rósta varð i ranni and styrr var older than the Guðrúnarhvqt, speaks of them as gumna, allz geirar ne bita, eggiar né tarn (25. 5-7). Once introduced, the ele­ ment of the armor could easily have fused with that older element of the invulnerability of HamcSir and SqtIi and the battle array of Gunnar and HQgni could have become impenetrable. Vqlsungasaga 44 (42) represents still another stage of development: the armor has become magical, rendered impenetrable by the spells of Guðrún. This magic armor undoubtedly has its inception in “ the invulnerable suits of armor” bequeathed to the sagas by foreign romance, together with “ unerring weapons, magic mirrors which re­ veal distant places, erotic talismans, bespelled garments, drinks of memory and forgetfulness, rich chessboards,

T H E NORSE L EGE ND

7

I

and náttúrusteinar which our heroes are constantly seek­ ing on quests, or receiving from the enamored giantesses and well-disposed witches who help them.” 118 It could have been introduced into the story only after magic had been “ legalized in literature” 119 under the influence of foreign romance; it belongs to the stuff of the fornaldarsgguTy the lygisQgur, and Saxo, not of heroic poetry.120 Likewise, the witchcraft of Guörún is a late innovation in the Norse legend, effected, by the various forces at work in the development of the latest Scandinavian versions of our story— foreign influences, misunderstand­ ing, and Christianity among others. These magic powers of hers are not known to the Eddie poets, nor yet to Snorri, but only to the sagaman and Saxo. In point of fact, all that Saxo does know of her is that she was a sorceress who aided the brothers in their vengeance. The Vqlsungasaga gives a detail particularly illuminat­ ing with regard to the part played by misunderstanding: paraphrasing the GuðrúnarhvQt, the sagaman erroneously interpreted the word ker (7. 4) as “ cup” instead of “ chest” ; consequently, in his account kumbl konunga or kerom valði becomes gaf þeim at drecka af storum kerum. From misunderstanding of a single word comes the first hint of witchcraft in an extant Norse text! And in Saxo GutSrun suffers the same fate as do the gods in his “ his­ tory” and in the lygisQgur: the heroic woman becomes a mere witch, just as ÓSinn the god of heroes becomes an amorous sorcerer.12111 It will be argued— it has been insisted— that already in the Hamðismál GucSrun is the sorceress giving magic armor to her sons.122 Such a conception is a direct con­ tradiction not only of the evidence of the poem itself, but of everything we know about the heroic life and

72

T H E LEGENDS OF E R M A N A R I C

heroic poetry. As Ranisch has emphasized, there is no room in the Hamðismál for magic of any kind.123 Even the compiler of the Vqlsungasaga does not succeed very well in incorporating it into his story: in capitulum 43 (41) he paraphrases the Guðrúnarhvqt very nicely, men­ tioning the storar bryniur ok godar ok aunnur herklqdi in their proper place; but he begins the next chapter pre­ cisely as though he had never before referred to GuSrun’s preparations: pat er nu ad segiafra sonum Gudrunar, ad hun hajdi sva butt þeirra herkledi . . . He is using two different traditions; that is clear. j f I can find in heroic poetry no trace of magic armor. To be sure, weapons of supernatural origin abound: Hrunting, Týrfingr, the sword given by OSinn to Sig^ mundr, the sword which Beowulf very opportunely finds hanging on the wall of the underwater lair. But there is a decided difference between byrnies and helmets hardened by a sorceress’ spells and weapons which are given to a warrior by Oc5inn god of warriors, or which are intended for a definite purpose, or which are simply well forged or tried in battle. Týrfingr must take a man’s life every time it is unsheathed; Hrunting was etched with poison and hardened by blood of battle. Beowulf’s breastplate was forged by Weland; that is, in another way of speak­ ing, it was exceptionally well forged. And the sword which Beowulf finds hanging on the wall of the under­ water lair, that which alone could take the life of Grendel’s mother, is another thing altogether; it is closely allied to the belief in the external soul, so common in folk tale. Invulnerable bodies, too, though not common, are known.124 But in their very essence they are a more primitive conception than is magic armor. Furthermore, nothing could be more foreign to the

T H E NORSE L E GE ND

73

nature of GuSrún, Gjuki’s daughter, than casting spells over helmets and byrnies so that her sons might safely carry out their vengeance for slain kin. This woman who slew two of her sons by her own hand and burned down the house over her husband Atli’s head is no Morgan la Fee or fairy princess. She is of the breed of that woman who could say of her own son who had shown fear: “ Tak þu hann þa ok drep hann. Eigi þarf hann þa lengr at Ufa” ;125 of the breed, too, of that daughter of Egill who took her sons all the way westward to Tunga to show them the house of their brother’s slayer. Indeed, J>orgerSarhvQt and GuftrúnarhvQt are one and the same. Says Guörún, “ You are little like Gunnar and HQgni; your mother’s brothers would soon achieve vengeance.” Says porgerSr, “ You are little like your noble kinsmen; your mother’s father would not have acted thus.” 126The Germanic woman might whet a laggard son or husband or brother to his duty; she might insult him into action by calling him a coward, a disgrace to the family. But she certainly did not make the insult reality by giving him magic armor to protect him against the blows that he was not man enough to ward off for himself. The magic preparation of her sons’ armor is but the last stage in the development of that tradition which represents GuSrún as doing all in her power to aid her sons in their vengeance for Svanhildr: she whets them to action (Hm) ; she gives them the armor of Gunnar and HQgni (Ghv); she gives them impenetrable armor (SnE ); she gives them armor which she herself has charmed

(.yss).\\ Linked with the armor moti f is that other tradition, like­ wise patently late in origin, of GuíSrún’s instructions to her sons before they set out on their journey. These are

74

THE

l e g e n d s

o f

e r m a n a r i c

known to neither the Hamðismál nor the GuðrúnarhvQt; they are to be found only in the prose texts. Moreover, and not insignificantly, they are to be found only in those passages which mention the impenetrable or magic armor. Equally significant is the fact that the instruc­ tions known to Snorri are not those known to the com­ piler of the Vqlsungasaga. According to the saga, Guðrún instructed her sons to avoid injury to stones and other large objects; according to Snorri, the brothers are to attack by night while the Goths sleep, and HamtSir and SQrli are to hew off the hands and feet of jQrmunrekkr, while Erpr lops off the head. These instructions re­ counted by Snorri are clearly based on the course of action which the brothers do— or should— carry out: the attack is made by night, while JQrmunrekkr and the Goths are sleeping; HamSir and SQrli do cut off the hands and feet, and Erpr would have cut off the head had he been present. Similarly, the instructions set forth in the saga appear to be based on what actually does happen: Hamftir and SQrli actually do injure stones— at least, in the opinion of the sagaman.127 As I see it, the explanation of the idea that Gu SnEyVss). Boer has demonstrated that this third brother came into the story as an answer to the question, Why did the brothers not slay the king?131 The mutilation of the king is an essential feature of the legend of the death of Ermanaric: Jordanes’ Sarus and Ammius merely wound him severely in the side, but do not kill him outright; the Norse HamSir and SqHí cut ofF his hands and feet, but fail to strike the fatal blow. Nevertheless, as the interest began to turn from Ermanaric to his attackers, people began to wonder why the attack failed; surely the avenging brothers would have succeeded in slaying the murderer of their sister had some unexpected occur­ rence not prevented them, especially since they came sufficiently close to the king to wound him. Some re­ sourceful poet, Boer thinks, concluded that there must have been a third brother who would have dealt the fatal blow, but who was not present at the critical moment. Thus did the third brother enter the legend of the death of Ermanaric.1 More difficult to ascertain is the source from which this third brother was derived. Here it is essential to dis­ tinguish clearly between the figure, the role, and the name Erpry for each of the three may have come from a different source. ' The role of Erpr bears a distinct resemblance to that of the third and youngest brother of folk tale who alone could have succeeded in the task imposed upon the three brothers.132 In the folk tale the elder two always fail, whereas the youngest always succeeds. The elder two are stupid and follow direc tions li ter ally, whereas the y ounges t

78

T H E LEGENDS OF E R M A N A R I C

is wise, frequently speaking in riddles which the others do not understand. The elder two are always jealous of the youngest, who usually is half brother or stepbrother, and try to cheat or kill him.133' i This folk-tale role of the third brother is quite recog­ nizable in the Hamðismál. In the first place, it is appar­ ent in the encounter between Erpr and his brothers. Hamöir and SQrli set out on their journey to the Gothic court and meet Erpr on the way. The very fact that they do meet him on the way and that he has not previously been mentioned in the poem suggests that in the en­ counter we have one of those casual appearances so common in folk tales.134They arrogantly ask him how he can aid them in their venture, and in true folk-tale fashion he answers that he will aid as one foot aids the other.135They do not understand, and he replies that “ it is ill to make known the way to cowards.” Infuriated, they slay him. Secondly, the division of the body of the king— hands and feet to be cut off by Hami?5 rikr), brother's son to Ermrich (Erminrikr), is forced to flee the enmity of Ermrich and seek refuge at the court of Etzel the Hun; years later he returns with his Hunnish allies and engages in battle with Ermrich in a battle before Ravenna (die Rabenschlacht); later still, thirty years after the beginning of his exile, hearing of the sick­ ness and subsequent death of Ermrich, he returns and takes possession of his kingdom. This course of events bears little resemblance to the actual career of the historical Theodoric who came to be known in German legend as Dietrich von Bern. Theo­ doric the Ostrogoth was born eighty years after Ermanaric had died and the Ostrogoths had fallen under the sway of the Huns. On the very day of his birth the news came that at last his people were free— free to harry, free to serve anyone who would pay them well. In 488, at the suggestion of the harassed Eastern Emperor Zeno, Theodoric led his Ostrogothic nation— the greater part of it, at least— into Italy to try his luck with Odoacer, who had set himself up as ruler of Italy. Four years of war­ fare followed, and in 493 Theodoric brought Odoacer to terms, treacherously slew him, and set up the Ostrogothic kingdom in Italy. During the ensuing thirty-three years, until his death in 526, Theodoric was virtually Western Roman Emperor. It is not within our province, however, to determine how the victorious Theodoric came to be regarded as the persecuted Dietrich von Bern; how the story of his exile and return arose; nor yet how tradition came to repre­ sent him as spending these thirty years of exile in the service of Attila. For this legend had already developed before Ermanaric came to be represented as the perse­

192

T H E LEGENDS OF E R M A N A R I C

cutor. We have evidence of this in the fact that the Hildebrandslied, a good two centuries older than the Annales™ knows of the exile of Theotrih. presumably at the Hunnish court; but represents him as fleeing the enmity not of Ermanaric, but of Otacher (Odoacer)^ his historical adversary: ßoh er [Hiltibrantl Otachres nid hina mit Theotrihhe^ enti sinero dezano filu (11. 18 f,). Since, then, in the Hildebrandslied it is Odoacer, in the Annales Ermanaric, who drives Theodoric into exile, the implication would seem to be that sometime between the date of the poem and the date of the chronicle, that is to say, sometime between the end of the eighth cen­ tury and the_end oLth^Lenth, Ermanaric took the place ^ O d oacer as j^rsecutor of Theodoric. And this is the opinion of the majority of the critics.24 The substitution is easily explained; indeed, certain traditions current in South Germany made it almost inevitable. In the first place, Ermanaric had long been known in South German tradition as an avaricious tyrant, willing to commit any nefarious deed in order to gain wealth and power; Theodoric had come to be re­ garded as driven from his hereditary kingdom by a foe who seized his lands. In the second place, Ermanaric was remembered in tradition as foe of the Huns:25 the sixth-century account of Jordanes represents him as struggling to defend his kingdom from Balamber; the seventh-century Wtdsið represents the innweorud Earmanrices as fighting the forces of Ætla the Hun in the Vistula woods. The Dietrich of German legend, on the other hand, was known as friend to Etzel the Hun, as peeking refuge at Etzel’s court. In the third place, both Ermanaric and Theodoric were remembered in legend as Amal kings, and had come to be regarded as contempo­

T H E G E R M A N LEGENDS

I93

raries. Actually, Ermanaric has a place in the Amal genealogy five generations above Theodoric; but legend delights in making its heroes contemporary, and Er­ manaric easily came to be regarded as brother to Theodoric’s father.26 Just so did German legend represent Vidigoia, leader of the Goths in the old days of the struggles with the Sarmatians,27 as retainer now to Er­ manaric, now to Dietrich; just so did Attila come to be regarded as foe of the one, friend of the other. Now, given two kings belonging to the samedynasty and rep­ resented as contemporary, thg one known as a tyrant and as foe of the Huns, the other as an exile and friend of the Huns, story would quite naturally have made the tyrant uncle responsible for the exile of the persecuted nephew.28 Thus, beyond doubt, did Ermanaric come to be regarded as persecutor of his nephew Dietrich von Bern. I Various critics have suggested further factors which they believe contributed toward the substitution of Ermanaric for Odoacer; but none bears too close scru­ tiny.29 One, however, is of particular interest, for it demonstrates the fallacy of basing important arguments about legend development on the sole evidence of per­ sonal names, a type of reasoning indulged in all too fre­ quently by students of the German Ermanaric legends. Jiriczek pointed out, without drawing any conclusions, that the name Sunilda-Sounigilda is known in connec­ tion with both Ermanaric and Odoacer.30 According to the seventh-century Byzantine chronicler Joannes of Antioch, after Odoacer was slain by Theodoric, his wife Sounigilda was imprisoned and his son Thela banished to Gaul; when Thela broke the terms of his exile and appeared in Italy, he was executed and his mother was

194

T H E LEGENDS OF E R M A N A R I C

starved to death in prison.31 That is, according to this account, Odoacer’s wife was named Sounigilda^ she was put to death, and her slaying was connected with the slaying of the son of Odoacer. There seems to be a partial parallel here to the Swanhild legend: according to the legend, a woman named Sunilda was put to death, and some versions represent this woman as wife to Ermanaric and her slaying as connected with the slaying of his son. But the agreement is purely superficial. The reasons for the slaying of the woman, the manner of the death, with all attendant circumstances, are different in the two stories: it is Ermanaric himself who orders the slay­ ing of Sunilda, the wife of his enemy, whereas it is Theodoric, enemy to Odoacer, who orders the death of Sounigilda; Sunilda is torn to pieces by wild horses, Sounigilda starved to death. The parallel exists in that Ermanaric and Theodoric each put to death in cruel fashion a woman named Sunilda-Sounigilda, whose ill luck it was to be the wife of an enemy of the Gothic king. Now, if Joannes’ account be truthful, we can merely note it as an interesting coincidence that Odoa­ cer’s wife and the woman of the Swanhild legend hap­ pened to bear the same name. But it is more likely that Theodoric did not order the slaying of Odoacer’s wife and, moreover, that Odoacer did not have a wife named Sounigilda: other sources relate the slaying of Odoacer,32 but no other mentions that of his wife. It is altogether probable that Joannes was merely inventing a pretty tale to blacken Theodoric’s character33 and that he got the idea, together with the name of the woman, from the Ermanaric story. In any case, truth or not, Joannes’ account affords no evidence that the name Sunilda was in any way connected with the substitution of Ermana-

T H E G E R M A N LEGENDS

J 95

ric for Odoacer as persecutor of Dietrich, and to base any such argument on it is not only fallacious, but dan­ gerous as well. Jiriczek pointed out that the tale could not be regarded as a factor in this question. Patzig, on the contrary, does see in the story a reason for the rep­ resentation of Ermanaric as persecutor of Dietrich,34 although, to be sure, he uses it as evidence for his wholly impossible theory that Ermanaric was substituted for Valentinian III and that the extant Dietrich legend arose through the preservation of memories of all the historic movements of the Goths at all times, mixed with literary Graeco-Roman sources, í j Ermanaric as slayer of his son.— Flodoard, in his His­ toria Remensis Ecclesiae (ca. 9 50), writes that Fulko, Archbishop of Rheims, admonishing King Arnulf to deal honestly with Charles the Simple, cited "from Teutonic books1' the* example of King Herrnenricus, who upon the impious advice of his counselor put to death all his race (qm omnem progeniem suam morti destinaverit impiis consiliis cuiusdam consiliarii su i)3hThis statement may be taken as the earliest extant hint of the existence in Germany of the tradition that Ermanaric slew his own son; for although no specific mentionlsTnadeliere oi^the son, the presumption is that he was included in theTftrogeniem of Hermen ric usJ6 Tfíe earliest specific mention of the slaying of the son is the statement in the Annales that Ermanricus of his own volition brought about the death ofhllTonTy son Fridericus (. . . mortenfFridertci umci filii sui, sua perpetrata voluntate . . .). The piðrikssaga (caps. 343-344) relates that Sifka, as a part of his vengeance for the seduction of his wife, suggested that Erminrikr send his eldest son, FriSrekr, to collect tribute from the Wilzi, and the second son, Reginballdr, to de-

196

T H E LEGENDS OF E R M A N A R I C

mand skatt from the king of England; he accused the third son, Samson, of seducing his daughter. Erminrikr, however, has no hand in the slaying of his two elder sons; their deaths are effected by Sifka himself, who has arranged with a jarl in Villcina borg that FriSrekr and his men be slain upon their arrival, and has given Reginballdr an unseaworthy ship. It is only Samson whom Erminrikr himself slays, and this seems to be rather the result of sudden impulse— possibly accident— than pre­ meditated murder: while Erminrikr, Sifka, and Samson are out riding one day, Sifka makes his accusation, and Erminrikr seizes his son with such fury that the boy falls off his horse and is trampled to death under the feet of his father's horse.37 Dietrichs Flucht (11. 2457 ff.) agrees with the þiðrikssaga that Friderich met his death among the Wilzi, but states that it was Ermrich himself who sent him there.38 Such are the extant German accounts of the slaying of Ermanaric’s son (or sons). 1 j The origin of this tradition in Germany is a disputed question, some critics suggesting that there may once have been an independent “ son legend,"39others holding that it is a trace of that version of the Swanhild legend which represented Ermanaric as slaying his wife and only son because of their reputed illicit love.40The first hypothesis is decidedly questionable, for it is difficult to explain how or why this “ son legend" should have arisen. Historically, there is no basis for it. Jordanes tells of only one son of Ermanaric, Hunimundus, and repre­ sents him as surviving his father; tradition appears to have remembered little about him other than that he was of great beauty and that he waged successful war against the Suavi.41 There is nothing here to lead to the legend that he was slain by his father.

T H E G E R M A N LEGENDS

I

97

The conjecture that the incident has been derived from the Swanhild legend, on the other hand, is de­ cidedly probable. The þiðrikssaga, in fact, affords almost decisive evidence in its favor. For this account represents Erminrikr not only as slaying his own son, but as doing so at the instigation of the evil counselor, and it gives him a motive which clearly is derived from the Swanhild legend: the counselor accuses the son of seducing a young girl. At the same time, the þiðrikssaga version differs in several important respects from that version of the Swanhild legend which we have seen in the Norse legend. In the first place, the girl whom Sifka reports to have been seduced is not wife to Erminrikr, but daughter of the counselor himself. Secondly, Erminrikr slays his son apparently by accident, not by intention. Again, there are three sons, only one of whom is accused of seducing the girl and only one of whom Erminrikr himself kills. And finally, the seduction motif has been extended: not only does Sifka accuse Samson of seducing his daughter, but his reason for making this accusation as well as for causing the deaths of the other two sons is that his wife has been seduced by Erminrikr. In short, the version of the Swanhild legend lying behind the piðrikssaga has lost the primary feature of the story, the slaying of Swan­ hild, together with the later feature of her marriage to Ermanaric; but it has retained such secondary develop­ ments as the role of the evil counselor, the slaying of the son, and the seduction motif. Even these details, however, have become distorted, and for reasons altogether clear. The tradition lying behind the saga obviously remembered, and emphasized, the seduction motif; but, with Swanhild gone from the story, the seduction was transferred to another girl.

I98

T H E LEGENDS OF E R M A N A R I C

Since this same tradition remembered that the counselor brought the accusation, the natural conclusion was that the girl must have been related to him, and thus the daughter was introduced into the story. There seems to have existed a variant of the son-slaying tradition, how­ ever, in which seduction as the motive was forgotten and in which the son’s name was Friderich.42 This tradition likewise seems to have been known to the sagaman, or his source; for in the piðrikssaga Erminrikr does have a son FriSrekr who is slain. To be sure, Erminrikr has no hand in his slaying; Sifka the counselor is wholly re­ sponsible for it. And the motive is, more or less remotely, seduction in that the slaying of FriSrekr is merely part of Sifka’s campaign of vengeance for the seduction of his wife. These details, however, are in all probability the result of rationalization in late tradition, possibly on the part of the sagaman himself. Sifka’s campaign of vengeance is clearly an attempt at an explanation of the dual personality of Erminrikr, the king who is now generous and good, now tyrant and oppressor and slayer of kin; seduction of Sifka’s wife as the motive of the vengeance is indubitably a doubling of the seduction motif derived from the Swanhild legend. If this is granted, then the slaying of FriSrekr is left unmotivated. The fact that it is Sifka, not Erminrikr, who actually causes the death presents no real difficulty. Clearly, the sagaman knew a tradition according to which Ermanaric slew only one son: this is constant in the Norse J(jrmunrekkr legend and in the Annales and Dietrichs Flucht alike. He knew further, as did the Norse legend, that the motive was the accusation of seduction falsely brought by the counselor. But he also had heard that other tra­ dition, preserved in the Annales and Dietrichs Flucht,

T H E G E R M A N LEGENDS

I

99

that Ermanaric slew his only son Friderich without any motive whatsoever. Rationalizing, he concluded that there must have been two sons, one named Friderich slain without reason, one slain by Ermanaric at the in­ stigation of the evil counselor. Further, to the sagaman Ermanaric was not a ruthless kinsman slayer, but a good king who committed evil deeds only at the instigation of the evil counselor. Consequently, he attributed that un­ motivated and unjustified slaying of FriSrekr to Sifka. This is precisely the sort of rationalization in which the sagaman has indulged with respect to other parts of his Erminrikr story. In the version of the son-slaying tradition preserved in the Annales and Dietrichs Flucht only the slaying of the son by Ermanaric remains, without any trace of the attendant circumstances. Both the role of the evil coun­ selor and the seduction motif have disappeared: the Annales states definitely that Ermanricus acted of his own volition, and Dietrichs Flucht cites the incident merely as an example of Ermrich’s innate baseness. Such a development was inevitable. The Annales preserves in recognizable form another incident derived from the Swanhild legend, which like­ wise is divorced from its original motivation: three brothers slay Ermanricus to avenge the murder of their father. Here the vengeance of the brothers is preserved without its original motive, the slaying of Swanhild; a slain father has been substituted for a slain sister. Just as the tradition known to the chronicler has remembered the ultimate consequence of the slaying of Swanhild, the vengeance of her brothers, without the proper motive, so this same tradition has remembered the deed attend­ ant upon the slaying of Swanhild, the slaying of her re­

200

T H E LEGENDS OF E R M A N A R I C

puted lover the son of Ermanaric, without its proper motive. Naturally, with Swanhild and her slaying lost from the story, both the vengeance of her brothers and the slaying of her lover were left unmotivated. For the one, the vengeance, tradition found a new motive in the slain father; for the other, the slaying of Ermanaric’s son, it found none. And its failure to do so is not surprising when we re­ member the conception of Ermanaric’s character which is preserved both in the Annales and in Dietrichs Flucht. The only reasons for introducing the evil counselor and the false accusation in the first place were, as we have seen,43 the necessity of explaining a cruel murder on the part of a good king and the fact that, had the accusation been just, the story would not have been sufficiently tragic to be the stuff of heroic poetry. When Ermanaric came to be conceived of as a heartless tyrant and a vi­ cious persecutor of kin, however, it was no longer neces­ sary to motivate the slaying of Friderich: the murder of an only son was no more than was to be expected of the man who could drive Dietrich von Bern into exile and seize his possessions. And it is this German conception of Ermanaric’s character that lies behind both the Annales and Dietrichs Flucht: to the chronicler he is astutior omnibus in dolo; to the poet, der ungetriuwist der ie von muoter wart geborn. In this tradition no new motive had to be supplied when Swanhild was forgotten, together with her unfortunate marriage to Ermanaric, her asserted love affair with his son, and her consequent death. On the contrary, the very conception of Ermanaric in this tradition led to the further loss of those other details which were remembered in the tradition drawn upon by the piðrikssaga: the seduction motif and the role of the counselor.

T H E G E R M A N LEGENDS

201

At the same time it is possible that, for a little while at least, this tradition preserved adumbrations of the orig­ inal motive just as the tradition behind the þiðrikssaga account of the slaying of Samson preserved the seduction without remembering who was seduced. The son of the Swanhild legend is accused of a deed which in actuality constitutes treachery to his father— seducing his father’s bride. Tradition may well have remembered for a time that Ermanaric slew his only son because of that son’s treachery, just as it remembered that three brothers slew Ermanaric to avenge slain kin, at the same time forgetting the identity of the kin. The one constant element of this second variant is the son’s name: Friderich (Friderich: DFl;Fridericus: Ann.; Friðrekr: þss.). The other factors are variable: the þiðrikssaga and the genealogy of the Amelunc line which precedes the actual narrative of Dietrichs Flucht agree that he met his death among the Wilzi, whereas the Annales states neither the method nor the locale of his slaying, and the main section of Dietrichs Flucht makes no mention of his death; the Annales and the Dietrichs Flucht genealogy agree that Ermanaric him­ self brought about the death of Friderich, the þiðrikssaga makes Sifka wholly responsible. Now, it hardly admits of dispute that, as has been suggested, this name has been derived from the historical Frederic the Rugian, victim of Odoacer and betrayer of Theodoric.44 Indeed, it would appear that in German tradition Frederic the Rugian was somehow identified with the slain son of Ermanaric, giving to him not only his name but also something of his colorful career. For Dietrichs Flucht gives an account of the part played by Ermrich’s son Friderich in the struggle between Ermrich and Dietrich

202

T H E LEGENDS OF E R M A N A R I C

von Bern which is strikingly similar to the historical account of the part played by Frederic the Rugian dur­ ing the struggle between Odoacer and Theodoric. In 487 Odoacer invaded the Rugian territory and carried cap­ tive to Ravenna the king, Feletheus (or Feva), and his wife; the following year the young prince Frederic re­ volted, but was forced to flee before the army sent by Odoacer and to seek refuge with Theodoric the Ostro­ goth.45 And with Theodoric he entered Italy and en­ gaged in the campaign against Odoacer. Then, suddenly, Frederic turned traitor, joining forces with the notorious Tufa, Master of the Soldiery in Odoacer’s forces, who some time before had deserted to Theodoric, become a general in the Ostrogothic army, and deserted again to Odoacer, taking with him a number of Goths. Together, Tufa and Frederic plundered all Liguria and Venetia, but they quarreled and engaged in a battle in which Tufa was slain. Frederic then shut himself up in Pavia, where he conducted a reign of terror for two years, until after the siege of Ravenna and Theodoric’s ultimate victory over Odoacer.47 What happened to him afterward, his­ tory does not tell us; after the Pavia episode he dis­ appears completely. According to Dietrichs Flucht (11. 3572 f.), Friderich is taken captive by Dietrich in the first battle between Dietrich and Ermrich,47 the ultimate outcome of which is the exile of Dietrich. He is held for some time in Bern but is finally set free, although Erm­ rich refuses to ransom him (11. 3835 flf.). Years later, when Dietrich returns with his Hunnish allies, he attacks Bádouwe, a city held by “ der junge künic Friderich” (11. 8233 ff.); after being defeated, Friderich flees (11. 8353 £)• He then drops out of the action of the story. The similarities between the two accounts are striking:

T H E G E R M A N LEGENDS

2O3

Frederic is for some time at Theodorici court, Friderich at Dietrich’s; Frederic deserts from Theodoric, Friderich is set free by Dietrich; Frederic is holding Pavia when Theodoric besieges Odoacer in Ravenna, Friderich is holding Bádouwe48when Dietrich returns with his allies to engage in battle before Ravenna; after the Pavia epi­ sode Frederic disappears from history, after the Bádouwe episode Friderich from the legendary account. There can be slight doubt that Frederic the Rugian gave to Fride­ rich son of Ermrich not only his name, but also the high­ lights of his career. It is to be noted that of all the MHG poems, only Dietrichs Flucht knows Friderich as son to Ermrich. In the others he appears simply as a retainer of Dietrich; he is the Friderich of Rabenschlacht (str. 261), the Fride­ rich der junge of Alpharts Tod (str. 76). In fact, Dietrichs Flucht itself well knows a Friderich Dietrich’s man (11. 2719fr.,9874) and a Friderich von Råbene (II.3012, 5730, 5850) besides Friderich son of Ermrich; there can be slight doubt that all go back to the same historical prototype. The implications are various and far-reach­ ing. It would seem that the tradition which made Frederic slain son to Ermanaric was not widely known in South Germany. This hypothesis is strengthened by the fact that even in Dietrichs Flucht the account of his death through the machinations of Ermrich his father is not organic:49 it occurs only in the genealogy of the Amelunc house, and finds no place, nor merits one, in the action covered by the narrative of the poem. Rather, South German tradition generally would appear to have remembered Frederic in a role very close to that which he played in history. The first step in the identification of Frederic the Ru-

204

T H E LEGENDS o f e r m a n a r i c

gian with the slain son of Ermanaric must have been the representation of Frederic as the victim of Ermanaric. Frederic was, essentially, a victim of Odoacer, an ally of Theodoric, and a traitor; and his story probably existed in tradition as a corollary to the story of the persecution of Theodoric by Odoacer. Consequently, when Ermana­ ric took the place of Odoacer as persecutor of Theodoric, he became persecutor of Frederic also. Then, since Er­ manaric was already known, from the Swanhild legend, as slayer of his son and, from the Dietrich story, as per­ secutor of kin, this Frederic whom he inherited from Odoacer as still another victim was now made his son. It is even possible that the Rugian was remembered in tradition as a traitor. In Dietrichs Flucht we have a vague reminiscence of this side of his story in the representa­ tion of him as first in Ermrich’s army, then a captive in Dietrich’s city of Bern, and finally as established with his own host in Bádouwe. Further, though there is no direct or definite evidence in Anglo-Saxon poetry for Frederic’s treachery, there may perhaps be a hint of it in the fact that in the JVtdsið Freoþeric, together with Wudga and Hama (who are characterized as wrceccan)> is listed separately from the other members of the innweorud.50We might not be far wrong, then, in assuming that early German tradition represented Frederic as victim of Ermanaric and as a traitor. I f this conjecture be correct, and if the slain son tradition also for a time represented the son as a traitor, then the identification of Frederic the Rugian with the slain son of Ermanaric is easily explained: tradition knew Frederic the traitor as victim of Ermanaric, the son accused of treachery as slain by Ermanaric, and concluded forthwith that Fred­ eric (Friderich) must have been the name of the son. If

T H E G E R M A N LEGENDS

2O5

by any chance any vague recollection of the false accu­ sation by the counselor remained in this variant as it did in that version which retained seduction as the mo­ tive for the slaying of the son, it would have disappeared once the son had come to be identified with Frederic the traitor: Frederic was not falsely accused of treachery, but was in reality guilty of it. With the false accusation gone, the person who made it was no longer essential to the story: Ermanaric would easily come to be regarded as slaying his only son Friderich of his own volition. ?We may conclude, then, that the figure and slaying of Ermanaric’s son have been derived from the Swanhild legend. At least two variants of the tradition de­ veloped in Germany, one of which remembered as the motive the counselor’s statement that the son had se­ duced a young girl, the other of which retained as a motive only a vague recollectión of the son’s treachery. The first of these variants is preserved in the piðrikssaga account of the slaying of Samson; the second fused with the German tradition of Ermanaric as persecutor of Frederic the Rugian and gave rise to the tradition that Ermanaric of his own volition slew his only son Friderich. Ermanaric as slayer of the Harlungs.— The first men­ tion of the Harlung legend in Germany is the statement in the Annales that Ermanricus hanged his nephews Embrica and Fritla Frithla: W). Dietrichs Flucht (11. 2467 flf.) tells that Diether, Ermrich’s eldest brother, had three young sons whom Ermrich slew; and (11. 2543 ff.) that Ermrich hanged the Harlunge and seized their gold and their land (presumably Brisache: 1. 2436). The piðrikssaga (cap. 344) relates that Sifka accused— or, more precisely, caused his wife to accuse— the neph­ ews of Erminrikr, EgarcS and Aki, the sons of Aki

2o6

T H E LEGENDS OF E R M A N A R I C

Aurlinga trausti, of wishing to seduce the queen; Erminrikr attacked and hanged them. According to the Vorrede, Ementrich hangs his brother Harlung’s two young sons and, at the suggestion of Sibiche his counselor, seizes their lands in Prisszgowe and around Brisach.51 Heinrich von München tells that Erntrich hanged the three Harlungs, the sons of Diether.62Such are the extant accounts of the slaying of the Harlungs. Other German texts men­ tion them, but not their deaths at the hands of Ermanaric. One Latin chronicle knows them as his nephews, although it says nothing of his slaying of them: accord­ ing to the Annales Pegavienses, Emelricus, rex Theutoniae, had two brothers, Dietmarus of Verdun and Herlibo of Brandenburg; and Herlibo had three sons, Emelricus, Vridelo, and Herlibo, who are known as the Harlongi.63 Other texts mention the Harlungs, but tell neither their slaying nor their relationship to Ermanaric: Beatus Rhenanus notes that from Harelus is derived the name of the Harlingi, whose city was Brisaeum;64 Ekkehard of Aura reports that Brisahc, with the adjacent country called Brisahcgowe, is said to have belonged formerly to the Harelungi;65 Rosengarten D (str. 63) speaks of Eckehart as guardian of the Harlungs, and in Rosen­ garten F (III, 16) Dietrich speaks of them as “ mines vetern kint” ; the MHG Biterolf represents Fritele and Imbrecke, the Harlungs, as young companions (and rela­ tives, 1. 9808) of Dietleib. The entire Harlung tradition, in short, is confused; no element remains constant through all the versions, nei­ ther the names of the brothers nor that of their father, neither their relationship to Ermanaric nor their fate, not even the number of the brothers. To be sure, with the exception of the Annales Pegavienses, which is really

TH E G E R M A N LEGENDS

2 07

giving only a genealogy and not a version of the Ermanaric legends, wherever Ermanaric is known as uncle to the Harlungs he is known also as their slayer. And the method of slaying them is invariably hanging. Further, all those documents which give more than a genealogy, with the single exception of the Annales, know another figure of the story, the foster father or guardian of the brothers. It is he who, according to Dietrichs Flucht and the Vorrede, endeavors to avenge their deaths. The name of this guardian, indeed, is one of the more stable factors of the story. In the South German texts and in the Vor­ rede it is Eckehart {Eckhartt: Vorrede); only the þiðrikssaga, which gives a decidedly distorted version of the Harlung story, differs by giving him a name which ap­ pears in the majority of the texts as that of one of the brothers, Fritila. However, neither the earliest extant German text to refer to the legend, the Annales, nor that text which makes the earliest extant mention of the Harlungs among any Germanic people, the Widsiðy mentions this figure at all. The names of the brothers are also comparatively con­ stant. The earliest of the extant German texts, the Annales, names them Embrica and Fritla Frithla: W). The Biterolf, the latest of the texts with which we are concerned, agrees; they are Imbrecke and Fritele. And these are the names which we found in the fVidsið, pos­ sibly as those of the Herelingas: Emerca and Fridla. The Annales Pegavienses preserves one of these, Vridelo; the other two names given in this chronicle are clearly improvisations— Emelricus, from the uncle; Herlibo, from the father— because the chronicler, or his source, remembered the name of only one of the brothers.56The piðrikssagay however, differs radically in giving to the

2 o8

T H E LEGENDS OF E R M A N A R I C

brothers the names which in South German tradition occur as those of the guardian and his father: Egarð {Eckehart) and Aki {Hache). There are scholars who re­ gard the piðrikssaga as preserving the genuine old names of the brothers and the Emerca and Fridla (and their variants) of the other texts as the contribution of later tradition. The most recent exponent of this view is Kemp Malone, who seems, like Boer before him, to regard the piðrikssaga as not only the oldest of the extant German accounts after the Annales, but also, presumably, the most reliable.57 Malone’s strongest argument, however, is that “ since the English poet [of the Wtdstð\ associates Hehca with the Herelings in line 112 it seems obvious that the tradition he knew was that represented by the piðrikssaga” ; and he con­ cludes, “ in other words, the Herelings of Widsith are Aki and Egari slay four hun­ dred and fifty men, but lose only one of their own num­ ber, Blödelinck, the twelve-year-old son of the widow. Certain striking similarities between this story and the Hamðismál are immediately apparent. In both, men whom Ermanaric has wronged undertake an expedition to his court to avenge their wrongs; in both, the attackers see on the roadside a gallows; in both, a watcher (or watchers) on the walls warns Ermanaric of the approach of the attacking forces; in both, the king announces that the attackers will be bound and hanged when they arrive in his hall; in Erminrikes Dot one of the attackers is very young and very wise and another is eyn Hörninck, while in the Hamðismál one who should attack the king is the youngest of three brothers and is called a hornungr by his brothers; in both, the youngest attacker is slain. And on the strength of these similarities between the two stories the majority of the critics have argued that the ninth-century Norse Hamðismál and the fifteenthcentury Low German Erminrikes Dot are based on the same old Low German original.113 All too little attention has been paid to the points of difference, which are even more striking than the points of resemblance. The identity of the attackers, the motive

238

T H E LEGENDS OF E R M A N A R I C

for the attack, the scene in the hall at the time of the attack, the outcome of it, the name and role of the youngest warrior and his relationship to the other mem­ bers of the attacking force, the use of the hanging motif: all are different in the two poems. In Erminrtkes Dot the attacking forces are not the two brothers of the wife of the Köninck van dem Armentriken, but the king’s nephew, Dirick van dem Berne, and his twelve cham­ pions. The motive is not the fact that the king has slain the sister of the attackers, but that he has wrested Dirick’s kingdom from him. The attackers do not fall under a shower of stones, after having succeeded only in cutting off the king’s hands and feet, but succeed in cutting down the entire “ veerdehalff hundert” men of the king and themselves escape with a single casualty, the youngest champion. This youngest warrior, Blödelinck, is not the brother of the other attackers, but a hero whose services must be bought with “ süluer vnd ock rodt Goldt.” He actually accompanies the warriors to the king’s hall and accomplishes marvelous deeds, whereas the youngest champion of the Hamðismál does not arrive. Blödelinck is slain by the king’s men in the battle in the hall; the Erpr of the Hamðismál is slain by his own brothers on the way to the king’s hall.114 Still, the differences between the two poems, no less than the similarities, indicate a close connection between the two: obviously, the details in Erminrtkes Dot are distortions of details in the Hamðismál.uh That the two poems could have been based upon a common original, an Old Saxon lay, however, is alto­ gether impossible. On the contrary, the points of differ­ ence indicate very clearly that the story preserved in the Hamðismál can never have been widely known in Ger-

TH E G E R M A N LEGENDS

239

many: for had it been widely known the changes would hardly have taken place. Certainly, there must be good reason for the old tale of the vengeance expedition of HamSir and SQrli to have come to be attached to Dietrich von Bern. Schneider gives a most plausible explana­ tion: the poem which reached Germany treated of events and heroes unknown to German audiences; they were delighted to hear of the death of Ermanaric, however, and, knowing nothing of the Swanhild story, but know­ ing very well the story of the persecuted Dietrich, they transferred the deed of which they wished to hear to a hero whom they knew to have a motive.116 The similarities do not demand that we postulate a common original. Since Erminrikes Dot is at least six centuries later in date than the Hamðismál, it would not be at all improbable that Erminrikes Dot goes back, per­ haps through several intermediaries, to a Norse version substantially like that preserved in the Hamðismál. We could assume with good reason that the Hamðismá! or a poem very much like it found its way to North Germany and there in the hands of the folk deteriorated through the centuries into the ballad which has come down to us.117 The nature of similarities and differences alike proves that we have in this ballad an excellent example of the fate of a heroic poem when it comes into the hands of a people who did not themselves develop the legend, but who received it from a people with a very different knowledge of legend: the main incidents of a story (the slaying of Swanhild and the vengeance of her brothers) may be lost entirely and a new story (the vengeance of Dietrich) substituted; but minor details (the gallows on the roadside, the fact that one hero was a bastard) will be remembered vividly.

24O

T H E LEGENDS OF E R M A N A R I C

A third version of the slaying of Ermanaric is pre­ served in the Vorrede: Eckhartt, guardian of the Harlungs, slays him to avenge his young lords.118 The case here is analogous to that of Erminrikes D6t: the folk were delighted to hear of Ermanaric’s death, but they knew nothing of HamtSir and SQrli, the avenging brothers; consequently, they transferred the deed to a hero whom they knew to have a motive.119 Similarly, the þiðrikssaga attributes— or seems to at­ tribute— the slaying of Erminrikr to still another indi­ vidual who was known to have a motive. For although this text represents Erminrikr as dying of an illness, it strongly hints that it was Sifka, the wronged counselor, who saw to it that the illness was fatal: he advised that the fat of Erminrikr’s paunch be cut and pulled out (cap. 417). It would appear that the sagaman, like other North German storytellers, had heard of the tradition that Ermanaric died at the hands of men he had wronged and is once again trying to reconcile conflicting traditions, that Ermanaric died of an illness and that he died at the hands of men he had wronged. Thus we have three German texts, all North German, which know of the attack on Ermanaric by men whose motive was vengeance, and a fourth, also North Ger­ man, which, though it makes no mention of a violent attack, likewise seems to know that his death was moti­ vated by vengeance; but no one of them retains the old motive of the slain sister Swanhild. The oldest of these accounts, the Annales, preserves the fact that three brothers attacked and that their motive was vengeance for slain kin; but in Erminrikes Dot and the Vorrede the attack has been transferred to figures of other Ermana­ ric legends, in the þiðrikssaga the vengeance is trans-

T H E G E R M A N LEGENDS

24I

ferred to the counselor. Two of the versions, the Annales and Erminrikes D oty show clearly that behind them must lie the Norse tradition. The other versions might as well be derived from Jordanes. But on the whole it seems fairly safe to conclude that the North German tradition of Ermanaric’s death has, in the main, been derived from the Norse version of the legend. Swanhild.— One text written in Germany gives a full account of the slaying of Swanhild and of the vengeance of her brothers. This is the Chronicon universale of Ekkehard of Aura, and the version of the legend it re­ cords is that of Jordanes— practically verbatim™ But elsewhere in extant German literature there is no men­ tion of Swanhild, nor of her slaying. And in all German story there is no account of her marriage to Ermanaric. Even her name is to be found only as a personal name in historical documents. One document has received special attention from students of legend because it happens to mention not only the name of Swanhild, but also that of one of her brothers: a deed of the year 786, found in the archives of St. Gall, states that a certain Heimo and his daughter Svanailta bequeath various properties to the monastery of St. Gall, and it is witnessed by a Saraleoz,121 It has been asserted, repeatedly and emphatically, that the occurrence of these names side by side in this single will affords incontrovertible proof that the Swanhild legend was known in South Germany in the eighth century.122 Actually, of course, it is merely a happy circumstance that a man by the name of Heimo happened to have a daughter by the name of Svanailta; and it is just as fortuitous that one of the witnesses happened to be named Saraleoz. Another man, whose wife was named

242

T H E LEGENDS OF E R M A N A R I C

Swanahilt (Svanihilt), in 807 gave St. Gall his properties in Brisagauge; one of the witnesses was Hecho.123 In 838, according to another document in the St. Gall archives, a certain Svanihilt gave all her property in Turgau to the church of St. Mary Mother of God in Ricchinbach; among the witnesses of this will was Amalunc.l24 The his­ torical documents, indeed, afford several such combina­ tions of names which recall grouping of heroes in Gothic (and other Germanic) story and some which find no counterpart in legend: the will of 786 lists also among the witnesses an Eghiart; another, from Nibalgauia (Legau?) and dated 766, lists Hiltibrandi and Heribrandi side by side among the witnesses and was written by a clerk named Hamedeos;Uh one from Ostrach in 851, Dioterih and Heimo;126 Wasserburg, in 827, Erffo and Hiltiprant and Cundram;m St. Gall, 832, Vurmheri and Witagouvo;128 Goldach, 856, Witagauo and Hacco and Waltheri;129 St. Gall, 864, Witagouo and Hagano;lZ0 Was­ serburg, 807, Hiltibrant and Hadabrant and Erfo;ul Bure, 843, an especially fine group— Witagouvo, Ketto> EikaehartyHilteyWiolant, Vago, and Wito.132Perhaps these combinations had best be explained as the workings of the laws of chance. Besides, the name Swanahilt together with its variants is common among the benefactors and lay sisters, especially of St. Gall, but also of Constance and other monasteries and churches of the region; it occurs constantly in the books of the Confraternities of St. Gall and Reichenau.133And, to digress from wills and mere lists, a Bavarian lady named Swanahildis (or Sonu hildis)y of the stock of Odilo, duke of Bavaria, was car­ ried off as a concubine by Charles Martel in 725, to become the mother of his third son, Grifo.134 Now, all these occurrences point to only one conclusion: that the

T H E G E R M A N LEGENDS

243

name Swanahilt was known in German Switzerland in the eighth century, and was very popular in a large part of southern Germany in the ninth; the St. Gall docu­ ment of 786 is no more important, from the standpoint of the student of the Ermanaric legends, than, let us say, the document of 807. Presumably, of course, knowledge of the legendary figure lies behind the popularity of the personal name; presumably at some time before 786 some story about Swanhild— and her brothers, for they, especially Ammius, are known, too136— came into German Switzerland. This may well have been the Gothic version, substantially that set down by Jordanes, which represents Swanhild as a political pawn, not as wife to Ermanaric. For from the Ostrogoths of Italy Cassiodorus, and thus Jordanes, learned the Swanhild story: with the Ostrogoths of Italy, German Switzerland stood in a particularly intimate re­ lation. It is in the land of the Alamanni that we find a heavy concentration of Gothic names, it is here that we find some of the earliest traces of Gothic legend;136 it is here, we think, that remnants of the tattered Gothic nation sought refuge when they were driven out of Italy forever. The Alamanni had come to the aid of the Goths in their final stand against the Empire, and it is no im­ probable guess that when at last the Goths accepted defeat many of them went home with their Alamannic allies to their Swiss fastnesses. If they did, they cer­ tainly took their legends with them.137 The personal names in German historical documents point to precisely the same conclusion as does Ekkehard’s chronicle: the early Gothic, or Jordanean, version of the Swanhild legend was known at one time in Ger­ many. At the same time, the testimony of the personal

244

THE

legends

of

erm an aric

names points to a conclusion wholly different from that of the learned Ekkehard. He, a historian of acumen, copies the account of a man whom he believes to be an authority; he is using accredited source material. In short, Ekkehard of Aura affords no indication whatso­ ever that the Gothic tale of the slaying of Swanhild ever became a part of German story,138 that it ever was known to anybody in Germany except chroniclers ac­ complished in Latin. The personal names, on the other hand, indicate that a story about Swanhild and her brothers, presumably the Gothic version, actually entered into the genuine oral traditions of South Germany. Yet there is in German literary texts (aside from Ekkehard) no story of Swanhild at all, neither the early Gothic nor the later, more personalized, version. The seduction motif, the slain son, the evil counselor, the, slaying of Ermanaric: these are to be found. But of Swanhild, there is the learned Ekkehard’s copy of Jor­ danes; there is her name hidden away in archives. Of her marriage to Ermanaric, of her tragic love for his son, of her consequent death: of these there is no account in all Germany. D E V E L O P M E N T O F T H E E R M A N A R IC LE G E N D S IN G E R M A N Y

Although various theories of the development of the Ermanaric legends in Germany have been advanced, we may, in general, discern three prevailing schools of thought: that of which Jiriczek perhaps has given the most thorough and detailed presentation, which sees in Alamannia the cradle of the German Ermanaric legends and in the “Harlung myth” the most efficacious factor in their development; that of Boer, who regards the

T H E G E R M A N LEGENDS

245

Elbe district as the birthplace of all the most important details of the legends and the piðrikssaga as preserving the oldest and most genuine German version; and that of Schneider, who places less emphasis upon locale than upon hypothetical “ heroic lays” which traveled from one end of Germany to the other and which coalesced and became remodeled in the course of their journeys. All these views, despite the wide differences between them, agree upon certain major points: first, that the kernel of the Ermanaric cycle in Germany was the Swanhild story in the form in which it was known to Jordanes; and second, that from this kernel there evolved, in the hands of German poets, not only the German Ermanaric legends now extant, but also a version of the Swanhild legend closely approximating that preserved to us in Norse texts, which died out in Germany at a comparatively early date. Following Müllenhoff, Jiriczek believes that into Alamannia there came the version of the Swanhild legend set down by Jordanes, that is, that version which re­ tained the political relationship between Swanhild and Ermanaric and the political motive for her death. Here it influenced, and was in turn influenced by, the Harlung my th.The seduction motif of the Harlung story was trans­ ferred to the Swanhild legend, and Swanhild became the wife of Ermanaric. The necklace of Frija became the hoard of the Harlungs, and Ermanaric’s motive in slay­ ing them was made greed for this gold rather than vengeance for the attempted seduction of Swanhild. As a result of this change in the motive for the slaying of the Harlungs, a new seducer of Swanhild had to be found; consequently, the son of Ermanaric was in­ troduced. The evil counselor also was transferred to

246

T H E LEGENDS OF E R M A N A R I C

the Swanhild legend, while Eckehart, his adversary, re­ mained in the Harlung story. This new, more inclusive Ermanaric legend then journeyed through all Germany and thence to England and to Scandinavia, although the Harlung story, which by now had become a mere episode in the greater story, did not reach Scandinavia. In Ger­ many Swanhild completely disappeared, the only weak trace of her tale being the þiðrikssaga account of the death of Samson. With the loss of Swanhild there was lost also the motive for the slaying of the son, and that incident became merely an instance of the counselor’s wickedness. And with Swanhild lost, the vengeance for her death, though retained for a time, was lost, and Ermanaric’s death came to be explained in other ways. The connection with the Dietrich legend made it im­ possible to represent Ermanaric as dying by the hand of two foreign heroes; so in High German territory he came to die a bloodless death of age or sickness. On North German soil traces of the old legend remained longer; so the vengeance was transferred to Dietrich (Erminrikes Dot) or Eckehart (Vorrede).139Obviously, Jiriczekregards Dietrich as the last victim whom legend attributed to Ermanaric, for he believes that the legend of Ermanaric’s persecution of Dietrich had not developed— or had not been carried through completely— by the eighth century, two centuries later than the revision of the Swanhild and Harlung legends.140 Boer, on the other hand, contends that the Jordanean, or more political, version of the Swanhild legend trav­ eled directly from the Black Sea region to the Elbe dis­ trict and there underwent certain changes: the woman Sunilda became Swanhild, wife to Ermanaric; the adul­ tery motif was brought in; the figure, guilt, and hanging

T H E G E R M A N LEGENDS

247

of the son were introduced. The Harlung story, of North German origin, was connected with Ermanaric because he was now well known as a slayer of kin. The seduction and hanging motifs were transferred from the Swanhild legend, which died out in Germany, to the Harlung story. And, still in North Germany, the “ old” Ermanaric legend was connected with the Dietrich legend: OdoacerSifka, the ancient enemy of Dietrich, came into the Ermanaric story and became the evil counselor.141 In short, on every major point, Boer's theory is directly contradictory to Jiriczek's. Schneider postulates the evolution of three lays. The first, Gothic in origin, and composed in the fifth century, a lay of which Jordanes' account is merely a summary, told of the execution of Swanhild and the death of Er­ manaric from a wound. Later, it passed from the Goths to the Germans and was well known in Alamannia in the eighth century, as is evidenced by the St. Gall names. A second lay was composed in Germany in the seventh century. It represented Ermanaric as a cruel tyrant and especially as a murderer of his kindred; it told of the murder of Friderich, son of Ermanaric; it attributed all Ermanaric's kinsman slaying to the insti­ gation of the evil counselor. Possibly this lay told also of the hanging of the Harlungs; of the quarrel of Wudga (Witege) and Hama (Heime) with Ermanaric over his treatment of his nephews, of Hama's flight with the hoard of the Harlungs, and of the adventurous life of Wudga and Hama thereafter. It may even have attrib­ uted to the evil counselor the motive of vengeance for slain kin (as Bicco in Saxo, Odoacer in the Annales). This is the lay that went over to England about the year 700 and was used by the JVídsið poet. Meanwhile,

240

T H E LEGENDS OF E R M A N A R I C

the first lay continued its journeyings in Germany. It became reshaped; it borrowed unmistakably from the German lay. Swanhild became wife of Ermanaric the known murderer of relatives, who was under the in­ fluence of the evil counselor. Thus there arose a third lay, which traveled to Scandinavia and became the basis of the Norse version. The first lay had lost its identity in the third; the second remained in Germany to dominate the Ermanaric legends there.142 Each of these views displays certain glaring weak­ nesses. Jiriczek's entire case rests upon a false premise: the contamination of the Jordanean version of the Swan­ hild legend by the “ Harlung my th.” For so far as anyone has ever been able to discover, no Harlung myth ever existed.143 Boer depends in large part upon the late and highly romantic þiðrikssaga and upon the other North German texts, and discounts wholly the evidence of the MHG poems, which he regards as “ young tradition” resting upon the representation in the þiðrikssaga.Ui Early or late, however, the MHG sources must be taken into consideration, the more especially since they are, for the most part, very little later than the extant þiðrikssaga. Schneider's theory can stand only if the existence of his postulated “ second lay” can be proved. And such a lay, with its encyclopaedic array of themes and incidents, could never have existed: the prime requi­ sites of the true heroic lay are unity of plot and of inci­ dent, conciseness.145 But perhaps of greatest importance is the fact that no one of these critics has taken sufficient cognizance of the very real and very striking differences between the Ermanaric traditions current in North Ger­ many and those in the South. For the most noticeable, and significant, characteristic

TH E G E R M A N LEGENDS

249

of the German Ermanaric legends is the difference between North and South German texts. Of highest im­ portance is the difference in the conceptions of Ermanaric’s character. In the South he is an avaricious tyrant, persecutor and slayer of kin, “ der ungetriuwist . . . der ie von muoter wart geborn.” In the North he is astutior omnibus in dolo, largior in dono; or again, great and kindly king and loving kinsman who is transformed through the machinations of his trusted counselor into a tyrant and a persecutor and slayer of kin. Again, in the South German texts the emphasis is on the Dietrich legend; the Harlung story and the slaying of the son are purely incidental, instances of the infamy of this most consummate of villains. The focal point of South Ger­ man story is the avarice and tyranny of Ermanaric which led him to oppress his illustrious kinsman Dietrich. In the North the persecution of Dietrich, although in the piðrikssaga recounted at greater length, is of no greater importance than the hanging of the Harlungs and the murder of the son; all are simply parts of the vengeance of Sifka. The focal point here is the betrayal of the good king by his evil counselor. And third is the difference in the knowledge of the Swanhild legend. In North German texts, vestiges of the Swanhild legend are comparatively heavy: Flodoard, the Annales, the piðrikssaga all mention the slaying of the son; Flodoard and the piðrikssaga, and possibly also the Annales, know that at the instigation of his evil counselor Ermanaric slew his kin; the Annales, Erminrikes Dot, and the Vor­ rede relate his death at the hands of men whom he has wronged; the piðrikssaga knows the seduction motif. On the other hand, traces of the story are few in South Ger­ many: the reference in Dietrichs Flucht to the sickness of

250

T H E LEGENDS OF E R M A N A R I C

Ermanaric; the personal names Svanaiha, Hamadeos, and Saraleoz in historical documents; the allusion in Dietrichs Flucht to the slaying of the only son, Friderich; the rep­ resentation in this same poem of the part played by the evil counselor, and the vague reminiscences of it in the other Middle High German poems. The first of these vestiges of the Swanhild legend in South German story, the death of Ermrich as it is re­ counted in Dietrichs Fluchtyshows that it has come from the Gothic version known to Jordanes, which represents Ermanaric as dragging out a miserable existence in bodily pain before he died. From the same version may very well have been derived the personal names in historical documents. And since these personal names not only are preserved in the St. Gall deed of 786, but are especially prevalent throughout German Switzer­ land from the eighth century on, the evidence amply supports the initial premise of the Jiriczek theory: that the Gothic version of the Swanhild legend traveled to Alamannia very early, possibly already in the sixth century. But that Alamannic poets made Swanhild wife to Ermanaric and introduced into the story the son and the evil counselor there is not the slightest shred of evi­ dence. On the contrary, there is no trace in all Germany of Swanhild’s unhappy marriage and tragic death. Fur­ thermore, the Dietrichs Flucht account of the slaying of the son argues strongly that this incident was not an inherent feature of South German story at all, but was derived from North German tradition at a compara­ tively late date. In the first place, it is not an organic part of the Dietrichs Flucht story: the poet simply cites it as an instance of Ermrich’s wickedness, giving no de­ tails other than that Ermrich sent his only son Friderich

T H E G E R M A N LEGENDS

25I

to his death among the Wilzi. There is no hint of a motive, no suggestion that Sibiche had a hand in the deed ;146the poet merely pauses a moment in his genealogy of the Amelunc line to disclose the fate of the only son of Ermrich and to emphasize Ermrich’s base character. Nowhere else in the poem, which covers the whole of Ermrich’s lifetime, is there mention of this murder. And it is this fact that the poet mentions the slaying of Friderich only in the genealogy and does not again refer to it that is particularly significant: for he knows and relates in detail that story of Friderich’s part in the struggle between Ermrich and Dietrich which had its roots in the treachery of Frederic the Rugian and which, like history, mentions Friderich no more after the fall of Bádouwe (Pavia). The presumption would seem to be that the poet knew two stories of Friderich, the one the genuine South German tradition of Friderich the deserter in the struggle with Dietrich, the other the tale that Ermanaric slew his only son Friderich. And that he derived his knowledge of this second story from North German tradition the poet of Dietrichs Flucht discloses by the statement that Friderich met his death among the Wilzi.147 For a South German poet would have known nothing of the Wilzi, the most formidable of the Baltic Slavs, in the eighth century allied with the Saxons against Charlemagne, in later centuries carrying on continual warfare against the Saxons. Indeed, the very name Wilzi apparently had no meaning to Middle High German poets: Heinrich von München, inserting the Dietrichs Flucht account of the slaying of Friderich into his chron­ icle, states that Erntrich sent his only son Friderich to ein wildez lant.u8 North German poets, on the other hand, not only would have been familiar with the name,

252

T H E LEGENDS OF E R M A N A R I C

but would have known that sending a man to the Wilzi was a subtle but certain way of insuring his death. In­ deed, since Dietrichs Flucht is the only MHG text to mention this story of the slaying of Ermanaric’s only son Friderich, and since we know the poet to have used a variety of sources, written as well as oral, we might be justified in assuming that it was he who borrowed the North German story of Friderich’s death to lend color to his own tale, and that the incident never, even at a late date, became a part of South German tradition. Likewise inorganic in South German story is the part played by the evil counselor. Ermrich is the blackest of tyrants, capable of committing any nefarious deed. He slays the Harlungs out of pure lust for land and gold; he sends his only son to death for no reason at all. It is only in the persecution of Dietrich that the counselor plays any part, and here his role is rather that of the guileful diplomat encouraging his lord in evil ways than that of actual instigator of the deed. Not his evil counselor’s advice, but his own lust for land and power, his ambi­ tion to be Roman Emperor, is Ermrich’s real motive in persecuting Dietrich von Bern. The counselor is wholly superfluous. And because he is, it is difficult to see why he should have been introduced to play this none too clearly defined role in a story which already portrayed Ermanaric as completely base. It is therefore only reasonable to suppose that the counselor has been bor­ rowed by South German tradition from a source in which his figure was an integral feature of the story and his role definitely evil. That this source was North Ger­ man tradition we may safely conclude from the fact that we find the evil counselor playing an important part in the North German Ermanaric story from a compara-

T H E G E R M A N LEGENDS

253

tively early date: already in the tenth century Flodoard knew that it was at the instigation of his evil counselor that Ermanaric slew all his kin; the Sifka of the þiðrikssaga is the true villain of the piece, with Erminrikr a mere pawn in the counselor's game of vengeance. In all probability, then, the story of the evil counselor filtered down from the north at a date earlier than that of the composition of any of the extant MHG poems, but after the South German story of Ermanaric the tyrant and slayer of the Harlungs and persecutor of Dietrich had attained an advanced stage of development. On the evidence of the extant texts, then, we may conclude that Jiriczek is unquestionably correct in main­ taining that the Gothic version of the Swanhild legend traveled to Alamannia, probably in the sixth century, but that he is wholly unjustified in assuming that Alamannic poets were the first to represent Swanhild as wife to Ermanaric and to add the slaying of the son and the figure and role of the evil counselor. On the other hand, Boer's assertion that these latter details were trans­ mitted to South Germany from the north is amply sub­ stantiated by the evidence of North and South German texts alike. But Boer’s presumption that the Gothic legend known to Jordanes traveled directly to the Elbe district and was there remodeled into the “ personalized" version is open to challenge. There is every indication that the story current in North Germany was a version of that which represents Ermanaric as slaying his wife and son at the instigation of his evil counselor and as slain by his wife's brothers. The version of that story which was known in North Germany, however, was decidedly corrupt. The various incidents as they are related in extant texts are

254

T H E LEGENDS o f e r m a n a r i c

distorted; the role of the evil counselor is by no means clear-cut, for he,like Ermanaric, has a dual personality, at first faithful, then bent upon a scheme of vengeance; the seduction motif is confused; Swanhild, her marriage, and her slaying have all been lost entirely. And Boer’s thesis becomes even more impossible when we recall that a number of the traces of the Swanhild legend in extant North German texts can only be distortion of details peculiar to the Norse version. The þiðrikssaga gives an account of the death of Samson which is obviously de­ rived from that tradition of the slaying of Swanhild which is set forth by Snorri; it retains the seduction motif without its proper application; Ermimikes Dot can only be a distorted version of a poem very much like the Hamðismál, for it retains, isolated from their proper context and deprived of their proper meaning, details which have import only in the Norse story; the Annales account of the death of Ermanaric can only be based upon a distorted version of Norse tradition, for it pre­ serves the folk-tale element of the third brother who should have cut off the king’s head, while at the same time it has lost the role which that brother was to have played and the jealousy of his brothers which prevented him from playing it. In short, the only hypothesis which can reasonably be postulated on the basis of the extant notices in North German texts is that the only version of the Swanhild legend ever to reach North Germany was the Norse, and that German poets distorted, rather than developing more fully, the story which came to them. This assumption renders explicable not only the loss of Swanhild from the story, not only the distortion of the several other elements of the legend, but also the

T H E G E R M A N LEGENDS

255

retention of those elements which do remain. The situa­ tion is analogous to that in Denmark, where once again we find confusion of Norse tradition. Danish tradition preserves clearly its affinity with Norse tradition, but it has lost certain important incidents which were essential to the Norse version: specifically, those incidents which arose from the connection of the Ermanaric legend with the SigurSr-Burgundian-Atli legend. This loss we can attribute to the fact that the VQlsung-Nibelung saga stuff was never popular in Denmark. Where the back­ ground of knowledge necessary to an understanding of certain incidents is lacking, those incidents will either disappear or be reinterpreted— this is an inflexible “ law of change.” Now, the necessary background for an un­ derstanding of the Norse jQrmunrekkr story is a knowl­ edge of Swanhild, and so far as we can tell from our extant texts Swanhild was never known— or, if known, was never very popular— in North Germany. Conse­ quently, when Norse tradition came there, the connec­ tion with Swanhild would have tended to be lost. One might, perhaps, expect the events connected with her, the slaying of her lover the son of Ermanaric and the vengeance of her brothers, likewise to have been lost. But there were, arising from German Ermanaric tradi­ tions, certain checks to the loss of those elements. The slaying of the son would be likely to be retained because it represented Ermanaric as slayer of kin. German tra­ dition, as formulated in the Dietrich legend, knew him as persecutor of kin. Therefore, since it is no long step from persecutor to slayer of kin, the slaying of the son would tend to remain, and it would tend to be retained in just such a form as that in which it has been retained in, let us say, the Annales: isolated from its original

25 6

T H E LEGENDS OF E R M A N A R I C

context, divorced from its original motive, as an instance of the wickedness of Ermanaric. The vengeance attack of the brothers, too, would tend to be retained in just such a form as that in which it has been: that is, as ven­ geance carried out by men who had been grievously wronged by the great villain, the tyrant and kinsman persecutor of German tradition. These traces of the Swanhild legend in German texts, then, point to certain definite conclusions. First, the version known in South Germany was that tale with a political tone which was known to Jordanes; all traces of the other version to be found in this region filtered through, isolated from the original context, from North German tradition. The version known in the north was the “ personalized” story, transmitted to North Ger­ many from Scandinavia, distorted in the hands of people and poets who had not the background to understand the meaning of all its elements, and influenced by the traces of the Jordanean version and specifically German Ermanaric traditions which poured up from the south. Secondly, these striking differences between North Ger­ man knowledge of the Swanhild legend and the South German demonstrate that German poets cannot have remodeled the Jordanean version into that story which portrays Ermanaric as the foolish old man who is duped by his evil counselor into slaying his wife and only son. On the contrary, such an assumption is absurd on its face, since it cannot explain these phenomena: the fully developed story is found in Scandinavia in the ninth century; it is not found in any extant South German text; only distorted fragments are found in North Ger­ man. In short, all the evidence points to the Scandina­ vian origin of the personalized Swanhild story in the

T H E G E R M A N LEGENDS

257

form in which we know it. The personalized version, like the political, came full-grown to Germany and underwent only deterioration, in no sense development, in the hands of German poets. And finally, these vestiges of the Swanhild legend, together with the differences in the conceptions of Ermanaric’s character and the focal point of his story, reveal that the Swanhild story must not be selected as the point of departure in any attempt at an analysis of the development of the Ermanaric legends in Germany. The most efficacious factor in the development of the Ermanaric legends in Germany is the same as that effecting the development of the Gothic, Anglo-Saxon, Norse, and Danish versions: the point of view of the people telling the story, influenced by its own traditional lore, its knowledge of geography and of history— all of which are directed by that people’s position on the map of Europe, its own history, its relations with its neigh­ bors. In Germany, however, we have to do with two peoples, with very different cultural, geographical, and historical backgrounds. The one dwelt in the south, close to Italy, the kingdom of Theodoric’s Goths; it maintained more or less friendly relations not only with those Goths of a century and more after Ermanaric’s death, who themselves had developed new points of view about Ermanaric and about his enemy the Huns, but also with members of those tribes which the great con­ queror had completely crushed. The other lived in the north, near peoples whose opinion of Ermanaric had been derived from the Ostrogoths. To the one people Dietrich von Bern was the legendary hero of prime im­ portance; and Ermanaric was actually, in the last analysis, drawn into his cycle, precisely as were various

258

T H E LEGENDS OF E R M A N A R I C

other figures of Germanic legend. To the other, the Germans of the north, Dietrich was a tragic hero of the first magnitude, but Ermanaric’s own tragedy was of no small importance.ThesetwopeoplesinhabitingGermany, however, did not maintain a state of isolation one from the other; their cultural backgrounds, if not their geo­ graphical and historical, influenced and were influenced by one another. The versions of the Ermanaric legends which they have handed down to us all show an amalga­ mation of North and South German tradition. In extant South German story Ermanaric is known primarily as persecutor of Dietrich von Bern. The slay­ ing of the Harlungs is simply the lesser event prelimi­ nary to the long campaign against Dietrich, revealing Ermanaric’s motives in persecuting his kin and, inci­ dentally, providing the funds for the campaign; the slaying of the son is still less important, serving merely as a further, and even more dreadful, example of Ermanaric’s baseness, and probably was never an integral part of the South German legend. This development of the Ermanaric legends in the south was inevitable. To South Germany and Italy in the fifth and sixth centuries were carried the traditions of Ermanaric as a tyrant over subject peoples, and this tradition remained paramount in the south throughout the Middle Ages. Not much later came the Swanhild story in the form in which it reached Jordanes (or Cassiodorus before him), a tale which enhanced Ermanaric’s reputation as an avari­ cious tyrant and an oppressor. Here, too, was spread his renown as foe of the Huns. Then, sometime after the eighth century but before the year 1000— that is, some­ time between the date of composition of the Hildebrandslied and that of the Annales,— under the influence of

T H E G E R M A N LEGENDS

259

these now well-known traditions about him, Ermanaric took the place of Odoacer as persecutor of Frederic the Rugian and of Theodoric the Ostrogoth, whose exile and return had come by the eighth century to be celebrated in song. And with this substitution the conception of Ermanaric’s character and the tenor of his legend be­ came firmly established in South Germany. He was not only the oppressor of his subjects but also the persecutor of his kindred. And thus was the way paved for the con­ nection with him of the originally independent Harlung story, that tale of two young princes put to death and their lands and gold seized by an avaricious king. All these traditions are South German in origin, for all have evolved from the conception of Ermanaric as a tyrant and a usurper of lands. In the south they lived on, to dominate the Ermanaric legends there throughout the Middle Ages. And from the south these traditions spread into North Germany, there to fuse with other Ermana­ ric traditions and thus to give rise to still others which are distinctly North German in their inception. Into North Germany from Scandinavia had come a version of the Swanhild legend which portrayed Er­ manaric not as a king in any way evil himself, not tainted with tyranny or wickedness, but as a foolish old man who was duped by his trusted counselor into slay­ ing his nearest of kin, and as slain by the two avenging brothers, whose names were traditionally preserved— a story which retained a dim recollection of a third and nameless brother. From this legend German story in­ herited certain traditions which dominated the North German version of the Ermanaric legends: the concep­ tion of Ermanaric as a good king; the figure and slaying of the only son; the figure and role of the evil counselor;

2 Ó0

T H E LEGENDS OF E R M A N A R I C

the slaying of Ermanaric himself, and the cutting off of his hands and feet, by men whom he had wronged; se­ duction as the reason for the slaying of his nearest of kin; hanging as his favorite means of executing relatives. Possibly from Scandinavian story, possibly from HreicSgothic, came that characterization of Ermanaric which had its ultimate origin among the Ostrogoths of his own day, and which was known also in England and in Den­ mark: as most generous of ring givers. And in North Germany these traditions fused with those which had come up from the south: tyrant, persecutor of Friderich and of his nephew Dietrich, slayer of the Harlungs. As a result, by the year 1000, in the Annales, we find the persecuted Friderich of the south identified with the slain son of Scandinavian story; Ermanaric bringing about the slaying of this only son by his own wish, not at the instigation of the treacherous counselor; the Har­ lungs made nephews of Ermanaric, and hanged; and possibly, though not too probably, the attribution to the counselor of the persecution of Dietrich. And by this date we find Ermanaric conceived of as astutior omnibus in dolo> largior in dono. Perhaps already by the end of the tenth century (although not recorded in the Annales), perhaps later, the role of the evil counselor was extended to the Harlung legend as well as to the legend of Ermanaric’s persecution of Dietrich von Bern; other details, such as the seduction motif, which were derived from the Scandinavian tradition, must have been known in North Germany by the date of the com­ position of the Annales although they are not mentioned in that text. Possibly, too, it was about this time that in North Germany the evil counselor was given the name Sibiche, but of this we cannot be certain. If the

T H E G E R M A N LEGENDS

261

name was derived from the Harlung legend, then per­ haps South German poets of a later time identified the counselor who induced Ermanaric to slay his nephews with the troublemaker of the Harlung story, the antith­ esis of the foster father Eckehart. If, on the other hand, the name has been taken over from HreiSgothic story, then presumably it was given to the evil counselor by the North German storytellers, who were the first, in Germany, to know of the figure and role of the counselor in the Ermanaric legends. All these traditions are dis­ tinctly North German in inception, for they have evolved from the fusion of traditions indubitably German in origin and development with other traditions imported from Scandinavia. But they did not remain isolated in North Germany; they began to trickle south. And thus in extant MHG texts we find the evil counselor made responsible for the persecution of Dietrich von Bern as well as, presumably, the slaying of the Harlungs, despite the fact that there was no real place for an evil counselor in the South Ger­ man tradition of the tyrannous Ermanaric who would stoop to any infamy in order to glut his greed for gold and land and power. We find in South German texts also the hanging motif, and, after 1100, at any rate, the slaying of the son and the name Friderich applied to this son. In Germany, then, we find a number of conflicting traditions, some of them South German in origin, some of them North German, and some resulting from the mingling of the two. Some— Ermanaric as slayer of his son, the evil counselor, the death of Ermanaric— have been derived from the Swanhild legend, that legend which arose to explain the death of Ermanaric in a mo-

2

Ó2

T H E LEGENDS OF E R M A N A R I C

ment of national crisis; others— Ermanaric as a tyrant and as the persecutor of Dietrich von Bern— have their roots in the historical traditions of Ermanaric as mighty conqueror of many peoples and foe of the Huns. We have plunged “ down into the multitude of tra­ ditions and the chaos of contradictions . . . to find which elements change and which remain” ; we have found that the element which undergoes the greatest change is the character of Ermanaric himself, and that the cause of that change is the difference in point of view of the peoples telling the story. It is the conflicting traditions about his character which determined the current of his legends in Germany and gave rise to the chaos of the extant versions.

N OTES TO C H A PT E R V 1 T h e H e r o ic L e g e n d s o f D e n m a rk , p. i. 2 D ie tr ic h s F lu c h t, 11 . 2414 f. 3 C h ron ico n W irzib u rg en se, M G H , S S , V I , 23. 4p id r ik s s a g a , udgivet ved H enrik Bertelsen, caps. 342 ff. A lthough this text was w ritten down in Norse, in N orw ay, it is a compilation o f N orth Germ an stories; hence its version o f the Erm anaric legends is to be regarded as N orth Germ an rather than Norse. 5 So named in the Strassburg manuscript o f the H eld en b u ch , from c a . 1450; printed b y Friedr. Heinr. von der H agen, H eld en b u ch , I,

c x i- c x x v i .

6 C f. Schneider, G H S , 1 , 60 f. 7 On the date and composition o f the saga cf. esp. Unger, Saga p i d r ik s K o n u n g s a f B e r n , pp.

hi

f., x n ; Storm , S a g n kred sen e om K a r l d en S tore og

D id r ik a f B e r n , pp. 88 ff., esp. 93 ff.; Aarb^ger, 1877, 315 f f ; Bertelsen, p i d r ik s Sag a a f B e r n , I,

x l v ii

ff., esp.

l iv

ff.; Jonsson, L it . H is t ., II , 840 ff.;

Schneider, G H S , I, 67 f. Th e most im portant literature is cited by Jonsson, p. 840, n. 3. 8 A n n . § u e d ., M G H , S S , I I I , 3 1 ; C h ron . W ir z ., M G H , S S , V I , 23. 9 So Lorenz, G e rm a n ia , X X X I , 149 f. 10 C f. Bresslau, N e u e s A r c h iv , X X V , 32 f. 11 C f. Jiriczek, D H S , 1 , 71 ff., esp. p. 73, n. 1. 12 Unless otherwise noted, I quote from the A n n a le s ^ u ed lin b u rg en ses. 13 First b y Schröder, Z f d A , X L I , 27 f f ; and more recently b y H oltzm ann, S a ch sen u n d A n h a lt, I, 94 (cf. his entire discussion, pp. 92 ff.). T h e flaws in Schroder’s arguments have frequently been pointed out: e.g., by K oegel, G esch ich te der deutschen L itte ra tu r, I, ii, 381; Jiriczek, D H S , I, 71 ff. C f. also Sijmons, G G A , 1900, 339, n. 1; Cham bers, W id s ith , n. 4 to p. 29. 14 C f. esp. Jiriczek, D H S , I, 71 f.; Sijmons, G rd r., I I I , 635 f. 16

C f. below, pp. 226 f., 235 f.

16 M G H S S , X X I V , 222. I t m ay be noted that the p s s ., too (cap. 15), rep­ resents Erminrikr as king over twelve cities in Spain. 17 For a different view cf. Heinzei, W S B , C X I X , 30. 18 C f. Cham bers, W id s ith , p. 37. 19 A bove, chap, iv, pp. 165 f. 20 Jordanes, G etica , X X I V , 130. 21 C f. below, pp. 197 ff., 229 f., 234 ff. 22 Cf. below, pp. 197 ff. 23 T h e H ild e b r a n d s lie d was written down in its present form ca. 800, but it was composed a t an earlier date. 24 So, e.g., Rieger, Z f d M , I, 234; Heinzel, W S B , C X I X , 56 (cf. also 61 ff.); Koegel, op. c it., I, i, 231 (but cf. ii, 214); Jiriczek, D H S , I, 101, 143 f.; Sij­ mons, G rd r., I l l , 629, 691; Heusler, B S B , 1909, 925; H o o p s, I, 465 f. For a

264

T H E LEGENDS OF E R M A N A R I C

contrary view cf. esp. Storm, S ag n kred sen e, pp. 72, 77; Boer, D ie Sa g en von E rm a n a r ic h u n d D ie tr ic h von B e r n , pp. 54 ff., 133 ff. (both Erm anaric and Sibiche took the place o f O doacer); H au p t, P a la e s tr a , C X X I X , 251 ff.; Wesle, P B B , X L V l , 264 f. 26

C f. Cham bers, W id sith ,, pp. 253 f.; for a contrary view, Heusler, B S B ,

I 9 ° 9 >9 2 5> H o o p s, 1,13 8 . 26 Heinzei { J V S B , C X I X , 56), possibly correctly, regards the representa­ tion o f the two as contemporaries as, in part at least, a result o f the represen­ tation o f Erm anaric as foe, Theodoric as friend, o f A ttila . 27 C f. Jordanes, G etica , X X X I V , 178. 28 Some critics— n otably Jiriczek, Sijmons, and Heusler— believe that E rm anaric’s reputation as kinsman slayer was a prime factor. B u t it is b y no means certain that this tradition was known at so early a date in South Ger­ m any, where the substitution m ust have taken place; cf. below, pp. 250 ff., esp. pp. 258 f. 29 E .g., those o f Rieger, op. c it., p. 235; Kauffm ann, P h ilo lo g isch e S tu d ie n . Festg abe f ü r E d u a r d Sievers, p. 156. 30 D H S , 1,14 4 . 31 Published, w ith a German paraphrase, b y M ommsen in H erm es, V I , 323 ff32 E .g., A n o n y m i V a le s ia n i p a r s p o sterio r, M G H , A u c t. a n t., I X , 320 ff.; C h ron ico n C u s p in ia n i (F a s ti V in d o b on en ses p r io r e s ), M G H , A u c t . a n t., I X , 320 ff.; A g n e lli . . . L ib e r p o n tific a lis , cap. 39, M G H , S crip t, rer. L a n g o b a rd ., p p - 303

f33 M om m sen, however {op. c it., pp. 332 ff.), believes that Joannes gives

an unpartisan account o f the fall o f Odoacer, whereas most o f the other sources give the official Gothic version. 34 D ie tr ic h von B ern u n d sein Sa g en kreis, pp. 11 f. 36 H is t. R e m en sis eccl., IV , M G H , S S , X I I I , 564. 36 Indeed, Boer argues {D ie S a g en , p. 54, n. 1) that the reference here can be only to sons o f Ermanaric. H owever, although the usual significance o f p ro g e n ies is “ offspring,” “ child,”

the meaning “ race,” “ fam ily,” occurs,

though rarely, even in classical usage. 37 T h e resemblances between this and Snorri’s account o f the death of Svanhildr would seem to suggest that the sagam an has drawn either upon tradition related to that used by Snorri or upon tradition which itself was de­ rived ultim ately from S n E . C f. also Sijmons, G G A , 1900,344 n. 1. 38 C f. also Heinrich von M ünchen, unpublished, but quoted b y Grim m , D H S , p. 225. 39 Sijmons, G G A , 1900, 343 f. (by im plication); Heusler, H o o p s , 'I, 628 (possibly); Kienast, Z f d A , L X I I I , 70 (possibly). 40 E .g., Jiriczek, D H S , 1 , 105; Boer, D ie S a g en , pp. 62 ff. For a still different view cf. Schneider, G H S , I, 239 ff., esp. p. 248; 378 f.

T H E G E R M A N LEGENDS

265

41 G etica , X L V I I I , 250. 42 Boer, too (D i e Sag en , pp. 62 f.), recognizes the existence o f the two variants and the combining o f them in the p s s . 43 A bove, chap, ii, pp. 52 f. 44 So, e.g., H einzei, W S B y C X I X , 5 (possibly); M atthaei, Z J d A y X L I I I , 326; Cham bers, W id s ith y pp. 222 f. Questioned b y Jiriczek, D H S y I, 134, n. 1; Panzer, D e u tsch e H elden sa ge im B r e isg a u y p. 46. 45 Eugippius, V ita S . S e u e r in iy p. xliiii. 46 On T u fa ’s treachery cf. Ennodius, V ita b e a tissim i viri E p ip h a n i e p is c o p i T ic in e n s is ecclesia e, C orp . s c r ip t. eccles.y V I , 359 ff. On all this cf. Schm idt, G eschichte d er deutschen Stä m m e, 1 , 121 ff., 288 ff., esp. 297 f. 47 According to the V orrede (p. c x x iii) , der Bernner takes captive one o f Em entrich’s two sons, who, however, is not named. 48 B a d o u w ey o f course, is not to be identified with P a v ia y but w ith P a d ova . B u t it must be remembered that the story has undoubtedly been handed down in oral tradition. 49 For a different explanation cf. Jiriczek, D H S y I, 76. 60 C f. above, chap, iv, pp. 158 ff. 51 Von der H agen ’s H eld en b u ch , I,

c x x iii.

62 Grimm, D H S y p. 225. 63 M G H y S S y X V I , 234. 54 R erum G erm an icarum lib r i tresy p. 98. 58

C h ron ico n un iversa le, M G H y S S y V I , 185 (cf. also the A nnalist Saxo,

ib id ., p. 603). W hether this chronicle, or this recension o f it, is the work o f Ekkehard o f Aura or o f F ru to lf o f M ichelsberg (the view advanced by Bresslau in N e u e s A r c h iv y X X I , 197 ff., and accepted b y such authorities as W attenbach) does not really concern us; since in studies o f Germ anic legend the author has always been designated as Ekkehard, perhaps for the sake of consistency it is well to continue the practice. 56 For a different view cf. Panzer, op. c it.y p. 69, n. 11. 57 Boer, D i e S a g en y p. 67; cf. his entire discussion of the developm ent o f the H arlung legend (pp. 67 ff.), which leaves almost com pletely out o f account the South Germ an texts; M alone, W id s ith y pp. 162 f. 58 O p. c it.y p. 163. 59 I b i d ; pp. 90, 161, 162 f.; cf. also S a g a -B o o k y X I , 28 f. 60 W id s ith y p. 1 61. 61 I b id .,p p . 126 f. 62 Grim m ’s discussion o f this point ( D H S y pp. 290 f.) is especially pene­ trating; cf. also Jiriczek, D H S y 1 , 78 ff. 63 Cf. above, chap, iv, pp. 156 f. 64 Z f d A y X X X , 221 ff.; he had already anticipated these views in Z f d A y X I I , 303 ff, C f. also Bugge, P B B y X I I , 71 ff.; Koegel, op. c it.y I, ii, 213 f. (also i, 149); Jiriczek, D H S y I, 100 ff.; Sijmons, G rd r.y I I I , 685.

26 6 66

T H E LEGENDS OF E R M A N A R I C E .g., b y Panzer, op. c it., pp. 85 f., n. 114; pp. 89 f., n. 135; Boer, D ie

Sa g en , pp. 65 ff.; Cham bers, W id s ith , p. 31 (but cf. pp. 32 f.). 66 C f., however, MüllenhoflF’s explanation, Z f d A , X I I , 304. 67 I t is mentioned for the first time by Ekkehard o f Aura {d. 1125). A number o f instances are listed b y Panzer, op. cit., pp. 60 ff., and the accom­ panying notes; cf. also Förstem ann, A ltd e u tsch es N a m e n b u ch , II, i, 1255 f. 68 832: H erilu n g o b u rg , Harlanden, near Pöchlarn, Förstemann, II, i, 1255; 853: H e rilu n g e u e ld e , near Harlanden, Förstem ann, I I , 1, 1256; 893: H erle n sd o rp h t, Hersdorf, near Wallersheim, Förstem ann, loc. cit.; 967: H erlin g eh em , near Oostakker, Förstem ann, loc. cit.; 1050: H e r ilin c h o v in , Härlkofen, Förste­ mann, loc. cit. 69 B e o w u lf, p. 178. 70 T h e relation tof r f f i r and q m u rlig r had already been suggested b y K oegel, o p . c it., I, ii, 214. 71 O p . c it., pp. 48 ff. M alone (W id s ith , pp. 162 fif., esp. 165; E n g lis h S tu d ie s, X V I I , 141 ff., esp. 143) sees a certain connection between the Herelingas and the W ild H ost. C f. also Flasdieck, A n g lia , L X I , 316 ff., but esp. 324 f. 72 C f. below, pp. 216 f. 73 F r ite le m ay have had originally a short vow el: cf. Boer, D ie S a g en , p. 67, n. I. 74 C f. Hieronymus, E p is tu la e , C V T. a letter to Sunnia and Fretela. 75 D ie Sa g en , p. 80. Cf. also those cited b y Rieger in favor o f his Heruli theory, Z f d A , X I , 201 flF. 76 H e regards (pp. 54 f.) as a trace o f E ck eh art’s m ythical role the fact that according to D F l . and R a b . E ckehart seizes the fleeing Sibiche (R ibstein), kills him w ith a sword, and slings the body in front o f him on his horse; bu t this argument seems a little strained. 77 G D S , 1 848, p. 472. 78 Z f d A , X L I I I , 318 ff. Cf. also M arquart, O steu ro p ä isch e u n d osta sia tisch e S treifzü g e, pp. 379 flF. (but cf. his more cautious statem ents in F e s ts c h r ift V ilh e lm T h om sen , pp. 105 f.). 79 C f. Panzer, op. c it., p. 47; Boer, D ie Sa g en , p. 80; Chambers, W id s ith , P- 3 1, n. 3 (cf. p. 216, n. to i. 87). 80 L o c . cit. 81 C f. Steenstrup’s penetrating criticism { A r k iv , X I I I , 134 flF.) o f A xel O lrik’s heavy use o f place names in analyzing S axo’s sources. 82 L o c . cit. 83 W id s ith , pp. 139 flF., 146. 84 I b id ., p. 140. 85 G etica , V , 43. 86 O p . c it., p. 146. 87 V ery unlikely, however, is M u c h ’s identification {H oop s, I I , 450; Z f d A , L X I I , 147) with the H arii o f Tacitus, the Charini o f Pliny.

T H E G E R M A N LEGENDS

267

88 So also Boer, D i e Sag en , pp. 78 f. 89 T h is is, I believe, telling evidence against the theory that Ostrogotha was no historical personage, but sim ply an eponymous hero. C f. also Cham ­ bers, W id s ith , pp. 13 ff. 90 G etica , X V I I , 94 ff. 91 So also Boer, D i e Sag en , p. 79. 92 Th is is one obstacle in the w ay o f accepting M alone’s explanation o f the hanging o f the Harlungs, W id s ith , p. 165; E n g lis h S tu d ie s, X V I I , 143. Th e Harlungs m ight have had "a particular association with Woden” ; but w hy did they come to be associated with Ermanaric? 93 Ja h rb u ch f ü r V olkslied fo rsch un g , IV , 41. 94 L o c. cit. 95 T h e passages in which this enm ity is hinted are cited b y Panzer, o p . c it., pp. 21 f. and the notes thereto. 96 Von der H agen ’s H eld en b u ch , I, c x x n f. 97 D ie S a g en , pp. 53 ff., esp. pp. 56 ff. 98 C f. esp. Lorenz, op. c it., pp. 144 f.; Jiriczek, D H S , 1 ,1 0 1 f., 143; Sijmons, G rd r., I l l , 691; G G A , 1900, 339, 344. 99 Since the forms T h eodericu s and T h eo d o ricu s appear quite consistently in Latin chronicles, I do not quite follow Schroder’s argument, Z f d A , X L I,2 6 . 100 T h e authorship and meaning o f this passage have been much disputed; however, there can be slight question that the Quedlinburg annalist himself composed the sentence: e.g., Lorenz, op. c it., pp. 145 f.; Grim m, D H S , p. 36 (cf. n. 3); Jiriczek, D H S , I, 158; Sijmons, G rd r., I l l , 623, and the refer­ ences he cites in n. 3; W attenbach, D e u tsch la n d s G esch ich tsq u ellen , I, 377. B u t cf. H oltzm ann, op. c it., pp. 94 ff. 101 B u t cf. B oer’s attem pts to explain this, D ie Sag en , pp. 59 f. 102 M üllenhoff, Z f d A , X X X , 226; Jiriczek, D H S , I, 101 ff.; Sijmons, G rd r., I l l , 685 f.; G G A , 1900,346. C f. also H ertz, D e u tsch e Sage im E ls a s s , pp. 87 ff. and n. 106; Cham bers, W id s ith , pp. 32 f. Schneider, too, although he does not subscribe to the M öllenhoff theory, derives the evil counselor from the H arlung story, Z f d A , L I V , 348; G H S , I, 239 ff., esp. p. 248. For his view o f the original content o f this story cf. below, p. 247. 103 On the personal name cf. above, chap, ii, p. 58 and nn. 88 and 89. 104 Cf. above, chap, iv, p. 156. 105 E ver since Binz suggested the identification, P B B , X X , 207 f. A dis­ senting voice has been heard from time to time, e.g., Cham bers, W id s ith , p. 34, n. 2. 106 So Boer, A a r b jg e r , 19 1 1, 42 f., 69; M alone, W id s ith , p. 188; P M L A , X L , 808 f. For a contrary view cf. Schneider, G H S , I I I , n o . 107 A a r b jg e r , 1 9 1 1,4 3 , 69. 108 P M L A , X L , 809. Still less can I agree w ith M alone { ib id ., p. 808; W id -

268

T H E LEGENDS OF E R M A N A R I C

sithy p. 1 88) that S ife c a seems to have been a stock name for an evil counselor in Germanic story. 109 T h e poet at no time actually says so in so m an y words, but hints it in

11. 2558 ff., 3501

ff., 4284 ff., 9846 ff.

110 Boer { D ie S a g en , pp. 57, 61; cf. the chart on p. 62) and Schneider (G H S , I, 249, 379) suggest that Odoacer here m ay be represented as the evil coun­ selor, in a role resembling that o f the Bicco o f Saxo. 111 T h a t the representation o f Odoacer as the third brother results from learned combination on the part o f the chronicler has already been recognized b y H einzei, W S B y C X I X , 3 f.; Jiriczek, D H S y I, 7 1 , 114 f. (but w ith an explanation very different from mine); Sijmons, G rd r.y I I I, 691 n.; G G A y I 9 °°> 3 3 9 Boer, on the contrary (D ie Sa g en y p. 56), wonders whether we are justified in drawing such a distinction between “ gelehrter com bination” and “ sagenbildung.” I t is to be noticed that he makes this comment im m ediately before his statem ent that the A n n a le s represents Odoacer as the true enem y o f Theodoric. 112 C f. de Boor, B eiträ g e z u r D e u tsch k u n d e , p. 38. For very different views o f its date and place o f composition cf. M eier, J a h rb u ch f ü r V o lk slied /orschu n g y IV , 46 f., 56; D e u tsch e V olkslied er m it ih ren M e lo d ie n y I, 24 ff. I use M eier’s edition in D eutsch e V olkslied er, I, 21 ff. 113 E .g., Jiriczek, D H S y I, i n ; Sijmons, Z f d P h y X X X V I I I , 163 f.; and in Gering-Sijmons, K o m m en ta r zu den L ie d e r n der E d d a , II , 426; Kienast, Z f d A y

Lxm,49ff.

V e ry few have dared speak ou t against this prevailing view: Boer {D ie S a g en y pp. 23 ff. [cf. also p. 20]) and Schneider { Z / d A y L I V , 350 ff.; but cf. G H S y I, 252), de Boor {op. c it.y pp. 35 ff.) and M eier {Ja h rb u ch , IV , 53 ff.; D e u tsch e V olkslied ery 1 , 26). 114 M eier {J a h rb u ch y IV , 53 ff.) sees even further points o f difference. 115 D e Boor and M eier, however, repudiate any attem pt to establish any close relationship. 116 Z fd A y L I V , 351 f. (but cf. G H S y I, 252). 117 C ertainly, the changes cannot have been effected in any such manner as that postulated b y Kienast, op. c it.y pp. 57 ff., esp. pp. 65 and 71 f.: “ als im laufe des 12 jh .s das andenken H am dirs und Sörlis immer mehr verblasste, entschloss sich ein spielmann dazu, das alte H am dirlied gehörig aufzu­ frischen” (p. 7 1), as a twentieth-century playw right m ight sit down and “ freshen up” a Shakespearean play by m aking it into a musical comedy. 118 C f. also Agricola, Sprichw örter, no. 667, quoted b y Grimm, D H S y pp. 326 f., who regards this notice as probably derived from the Vorrede; Jiriczek { D H S y I, 82), as an independent variant. 119 HeinzeFs comments on these later versions o f Erm anaric’s death { J V S B y C X I X , 63 f.) are worth noting.

THE

GERMAN

LEGENDS

269

120 M G H , S S , V I , 123. 121 Wartmann, U rk u n d en b u ch der A b te i S a n c t G a llen , 1 , 104. m M ost em phatically by Müllenhoff, Z f d A , X I I , 302 ff. C f. also Jiriczek, D H S , I, 68; M atthaei, Z f d A , X L I I I , 322; Schneider, G H S y I, 379. B u t cf. Sijmons, G G A , 1900,342 f.; Cham bers, W id s ith , p. 29, n. 1. 123 Wartmann, op. c it., 1 , 186. 124 I b id .. 1 , 344 f. ™ I b id ., I , 49 f. 126 I b id ., II, 37. 127 I b id ., I , 285 f. 128 I b id ., 1 , 316 f. 189/tó /., II , 68 f. 130 Ibid.y II, n o . 131 I b id .y 1 , 182 f. 132 I b id ., II, 6 f. 133 C f. the lists in Piper, L i b r i C o n fra tern ita tum S a n c ti G a lli A u g ie n s is F a b a r ie n s is . 134 A n n a le s q u i d ic u n tu r E in h a r d iy S c r ip t, rer. G erm .y IX , 3; A n n a le s M e tten ses, S c r ip t, rer. G erm ., V I I , 26,32 f. 135 Hamedeos,"]G(y (Wartmann, op. c it.y I, 50); H am a d eo h c, 799 { I b id ., 1 , 148); H a m ed eoh, 807 { I b id ., I, 187); H a m a d h io , 855 { I b id ., II, 6 1); H a m m i, 866 { I b id ., II, 132); Sara leoz, 786 { I b id ., I, 104). 136 C f. Uhland, S ch riften , V I I I , 334 ff. 137 On all this cf. Chambers, W id s ith , p. 37. 138 E kkehard’s comment that the Sarus and Am m ius who wounded Erm enricus vulgariter Sarelo et H a m id iech d ic u n tu r { M G H , S S , V I , 130) is certainly not evidence that a story about the slaying o f Swanhild was current in G er­ man tradition. H e is here objecting to the stories that Ermenricus drove Theodericus to refuge with A ttila at the instigation o f Odoacer, stories quod non solu m v u lg a rifa b u la tio n e et ca n tilen a ru m m od u lation e u sita tu r , verum etiam in q u ibusda m cr o n icis annotatur; and he points out, b y w ay o f correction o f these erroneous tales, that Ermenricus reigned in the time of Valens and Valentinian and was killed a du obu s fr a tr ib u s Saro et A m m io , quos co n icim u s eos f u i s s e q u i vulgariter Sarelo et H a m id ie c h d icu n tu r . T h e specific chronicle which he here attacks is the W ürzburg, which, as we have seen, likewise knew o f the death o f Ermanaric at the hands o f the brothers Ham idus and Sarilus, but nothing o f Swanhild. T h is passage suggests a tem pting conjecture. Jiriczek argues { D H S , I, 69) that had tradition known the third brother mentioned in the W ürzburg chronicle, Ekkehard would certainly have mentioned it as another “ untruth.” Perhaps if tradition had represented Swanhild as wife to Erm anaric, E kk e­ hard would have mentioned it! B u t, as Boer warns { D ie S a g en , p. 55), it is dangerous to argue from w hat Ekkehard does not say.

27O

THE

139D H S y 1 , 9 9 fF. 140I b id ., pp. 143

LEGENDS

OF

ERM ANARIC

(in the H ild e b r a n d slie d Theodoric’s opponent is Odoacer),

157 (in D e o r peodric’s enemy cannot be Eormanric). 141 D ie Sa g en , pp. 15 ff., 25, 51, 62, 8o, and p a s s im ; cf. esp. the various charts appearing at intervals throughout the book.

142G H S y I,3 7 8 ff. 143 C f. pp. 166 f., and Sijmons* criticism o f Jiriczek’s views, G G A y 1900, 3 4 1 ff144 D ie Sa g en , pp. 137 ff. and p a s s im . C f., e.g., his com m ent on R a ben sehla cht (p. 147): “ Auch hier ist es klar, dass die darstellung der p S zu grunde liegt.” 145 Heusler, D ie altgerm an ische D ic h tu n g , pp. 158 f. Some of Olrik’s “ Episke Love i Folkedigtningen” {D a n sk e S tu d ie r, 1908, 69 ff.; also cf. Z f d A y L I , 9 ff.), too, apply to the heroic lay. 146 Jiriczek appears to have overlooked this point when he states (D H S y I, h i

) that when the seduction of Swanhild as the m otive for the slaying o f the

son was lost in German story, the slaying became merely an instance o f the counselor’s wickedness. 147 Cf. also Jiriczek, D H S y 1 ,1 8 1 ; Schneider, G H S y I, 235 f. 148 LI. 295 ff.; cf. Grim m, D H S , p. 225.

CHAPTER VI

Transmission o f the Legends: Conclusion preceding chapters the several versions of the Ermanaric legends current among the Germanic peoples have been analyzed, and the origin and de­ velopment of each of these versions have been discussed at length. A glance in retrospect at the several legends may afford a clearer view of the history of each of them in the Middle Ages. The first of the stories about Ermanaric to develop was the Swanhild legend, which arose, in'all probability, as a result of the Ostrogothic emotional reaction to the downfall of the great Gothic kingdom and the death of the Gothic king in a moment of national crisis. In its extant versions we can discern with a fair degree of accuracy all the stages of legend development, from the historical event giving rise to its growth in the heroic age, through a highly dramatic, highly individualized treatment of the theme, which has received its stamp from the hands of highly skilled poets, to its ultimate deterioration in popular ballad when the Middle Ages had run their course. Ermanaric died, ingloriously, if we may believe Am­ mianus Marcellinus, an eminently trustworthy histo­ rian. We have no tale here, but the simple account of a warlike king, disheartened at seeing his extensive and fertile fields overrun by the squat Huns riding furiously from the east, and crushed at last by fear and grief. But

isr the

I

C2713

272

T H E LEGENDS OF E R M A N A R I C

in this simple account is a germ of tragedy: the tragedy of an illustrious prince unable to save his people from the flaming ferocity of a cruel and horrible invader, the tragedy of a great kingdom deprived of its leader in its hour of need, and the greatest tragedy of all— that of a Germanic king who brought death upon himself. This germ of tragedy served as the nucleus of a legend which became more and more elaborated and more and more distorted as the tale spread from Germanic people to Germanic people. Each new poet added something of his own to the lay; he supplied that which he felt was lacking when the story came to him; he interpreted. Supplying missing details, he drew, naturally enough, upon the stock of legendary tradition which was the heritage of the poets of his people, upon his own histor­ ical and geographical knowledge, upon the cultural background of his race. And interpreting, he was in­ fluenced, equally naturally, by the point of view from which his people regarded the mighty Ermanaric, no­ blest of the Amal kings, the Gothic Alexander. A very early stage of the Swanhild legend has been preserved to us by Jordanes. Here the political apsects, characteristic of the earlier stage of Germanic legend, are retained: the Goths are fighting against the Huns and are betrayed by the Rosomoni. Individuals play their part, yes; just as in the Beowulf version of the Ingeld story individuals play their part. HröSgär offers Freawaru, the pawn in the political game, to Ingeld; the marriage is about to be consummated; the unnamed but nonetheless individualized warrior precipitates the battle; Ingeld takes up the sword once more.1 Just so, the un­ named but nonetheless individualized husband of Sun1 For notes see pages 298-299.

CONCLUSION

273

ilda precipitates trouble by his “ deceitful defection” ; Ermanaric slays Sunilda, the pawn in the political game; Sarus and Ammius take vengeance. But just as “ the feud between Danes and Heathobeardan is represented as a national business,” 2 so does the Sunilda story of Jordanes savor of national business in that the motive for her slaying, the motivation of the entire story, is the political deed of treachery of subject to overlord. In the Wtdsið we may have a version of the Swanhild legend. But since we do not know what story about Ealhhild was known to the poet, we cannot draw upon this poem for evidence concerning the development of the Swanhild legend. We can merely note that we may have here a stage of the legend at which Swanhild has come to be represented as Ermanaric’s wife. The second stage of the legend which has been pre­ served to us is that set forth in the Norse version. Here the political aspects have been lost entirely: no longer have we struggles between peoples, but rather struggles between individuals. Ermanaric remains king of the Goths, but Swanhild and her brothers are no longer members of the tribe of the Rosomoni (a historic tribe if the identification of them as the Roxolani holds, or at least the surrogate of a historical tribe); Swanhild is daughter to SigurSr the VQlsung and GutSrun Gjuki’s daughter, Swanhild’s brothers are sons to GutSrun and her third husband, Jonakr. No longer is the motive for the slaying of Swanhild the political one of the treachery of her husband to Ermanaric; she is wife to Ermanaric now, and the motive for her slaying is her own asserted treachery to him involved in her supposed love for his son. No longer do we have the political tragedy of a na­ tion, the downfall of the mighty Gothic kingdom and

274

T H E LEGENDS o f e r m a n a r i c

the death of the king in a moment of national crisis; the tragedy now is the personal tragedy of an old king duped by his evil counselor into slaying his wife and only son, and the consequent personal tragedy of all those upon whom his act recoils. In this second stage of the Swanhild legend we have to do with the work of poets who were directed by no national impulses. The ignominy of subservience to the Huns abated in Gothic breasts as the years passed and Goths were still ruled by their own kings, albeit owing allegiance to Hunnish kings from Balamber to Attila. But the Swanhild story lived on: as the Rosomoni and the nature of their treachery faded from recollection, and when once the political results of that treachery were no longer keenly felt, the personal elements of the story inevitably grew in significance and were realized with ever greater intensity. Tradition still remembered treachery as the motive for Swanhild's slaying, and the dramatic theme of double vengeance— Ermanaric's ven­ geance on her, and her brothers' vengeance on him. The tragedy of Ermanaric's death, no longer linked with na­ tional tragedy, assumed new vividness as the final act in a tragedy of fate. The shift in the focus of dramatic interest from national struggle to conflict between indi­ viduals involved a shift in the motive for Swanhild’s murder, and a consequent reinterpretation of the rela­ tionships and circumstances which lay behind it. In this process new characters entered the story, bringing new conflicts with them, and the plot became richer and more complex. The most pregnant moment in this proc­ ess of poetic interpretation was that at which Swanhild was made wife to Ermanaric: this new relationship gave rise to the new motive of love between stepmother and

CONCLUSION

275

stepson, the new characters of son and evil counselor, and the new theme of the false accusation. Thus did the second stage of the Swanhild legend come into existence. Once this stage of the legend came into the hands of Norse poets, they began to expand and develop it. They answered the two great questions left unanswered in the story which came to them: Why did the brothers fail to slay Ermanaric when they came close enough to him to wound him? What was the fate of the brothers? Folk tale gave the answer to the first: the third and youngest brother alone could have slain the king, but he was slain by his jealous elder brothers before he could accomplish the deed. Primitive belief of one sort or another afforded the answer to the second question: the brothers were stoned to death because they were invulnerable, or be­ cause they had insulted stones by spilling on them the unhallowed blood of the murdered third brother, or for some other reason no longer known to us. Further details were probably added from the OcSinn cult: the hanging of the son of Ermanaric; the dedication of Swanhild’s brothers to O’Sinn; the intervention of OSinn in the battle in the hall. The legend was connected with the extremely popular Norse SigurSr-Burgundian-Atli leg­ end: Swanhild was made daughter to SigurSr and GuSrún, and her brothers sons to GuSrun and Jonakr; GuSrun the avenging woman became the true avenger of Swanhild; the name Erpr was given to the third brother, and Bikki to the evil counselor. Other new mo­ tifs were added; among them, that of the impenetrable (or magic) armor. The king’s son was given the name Randvér. New incidents were added: GuSrún’s whetting of her sons to action; the journey of the brothers to Ermanaric’s court. And as a result of the addition of

276

T H E L EGENDS OF E R M A N A R I C

these new characters, incidents, relationships, and mo­ tifs the legend attained the height of its development in the hands of Norse poets. But we see also in the later Norse prose versions, Snorra Edda and the Vqlsungasaga, the beginning of the deterioration of legend in the hands of later rationaliz­ ing and systematizing storytellers who no longer under­ stood the true meaning of various details of the story. Systematizing, these storytellers made the third brother, like the other two, son to Guðrún and Jonakr; and there­ by they lost the original motive for his slaying, the hatred of the elder brothers for the youngest, and bastard, half brother. The original meaning of the ston­ ing was lost, and various new explanations of it were introduced. Nevertheless, although the legend loses a good deal in the hands of these storytellers, the later versions stand close to the older. Deterioration has set in, but has not become advanced. A wholly different type of deterioration is to be ob­ served in the Danish version. Here important incidents, such as GuSrún’s whetting, are lost entirely; others, like the slaying of the third brother, are retained, but in dis­ torted form; and significant details, such as the cutting off of Ermanaric’s hands and feet, are remembered without their original meaning. The causes of this type of deterioration are to be sought in the fact that the legend has passed from West Norse poets to Danish storytellers with a wholly different mode of thought and national spirit, with different ideals and interests. The course of the development of the legend by West Norse poets had been determined by impulses originating in the mode of thought, in the whole cultural background, of that people. When the legend, full-grown, was trans-

CONCLUSION

277

mitted to Denmark, however, the background was no longer there, and change was inevitable. The West Norse interest in the VQlsung-Nibelung story, which struck its deepest roots in Norway and Iceland, had made for the linking with it of the Swanhild legend and had caused this linking to become in itself one of the greatest factors in the development of the Norse version. The VQlsungNibelung story seems never to have been popular among the Danes, however, who, from first to last, were most concerned with the legends of their own national heroes, the SkjQldung kings. Consequently, when the Norse ver­ sion of the Ermanaric legend traveled to Denmark, the connection with the VQlsung-Nibelung legend disap­ peared, leaving as its only traces the names Guthruna and Buthlus and the enmity of Bicco to the house of Swanhild. The Danish version which has been preserved to us shows, in addition to this variety of deterioration, a definite distortion, the work of the learned medieval chronicler who did not distinguish clearly enough the difference between legend and history, but tried to rec­ oncile the two. Saxo makes Ermanaric king of the Danes, and gives him a typical viking career, with some deeds drawn from the careers of historical Danish kings. As a result, certain details in the Swanhild legend must be changed. Later chroniclers level out the legendary elements so far as possible, emphasizing the “ historical” side of Ermanaric’s story. Thus in the hands of medieval Danish historians and annalists legend which has de­ veloped from history becomes “ history” once again. In extant German texts we find various stages of de­ terioration of legend. The first of these is represented in the Quedlinburg and Würzburg chronicles. Here, as in

27 8

T H E LEGENDS OF E R M A N A R I C

the Danish version, some important incidents are lost, others are remembered only in distorted form, and details are recalled without their original meaning. Just as the Danish version retained the detail of the cutting off of Ermanaric’s hands and feet without its original meaning, the fact that the third brother who alone could have cut off the head had been slain, so these German chronicles retain this detail without its meaning. But whereas the important incidents lost in the Danish ver­ sion are incidents which had been added by Norse poets as the result of the connection of the Swanhild legend with the SigurcSr-Burgundian-Atli legend, the most im­ portant incident lost in the German version had been an essential feature of the Swanhild legend from its very beginning: the slaying of Swanhild. These chronicles preserve only two incidents of the Swanhild story: the slaying of Ermanaric’s son and the attack of the brothers on Ermanaric. But these two incidents are not con­ nected, and the motives for both have been lost. Another and yet more distorted version of the Swanhild legend is preserved in the þiðrikssaga. Here, too, Swanhild, her marriage, and her slaying are missing; and here, further, the vengeance attack upon Ermanaric is missing. But various incidents and other elements of the Norse ver­ sion of the legend are preserved, in more or less distorted form: the slaying of the son at the instigation of the evil counselor; seduction as the motive for the slaying of the son; the counselor’s statement that the wife of Ermana­ ric has been seduced. But the son has been tripled, and only one of the three sons does Ermanaric himself slay; this slaying seems to be accidental, and the motive for it is the counselor’s assertion that this son has seduced his daughter, not Ermanaric’s wife. The seduction motif

CONCLUSION

279

has likewise been tripled. And the counselor accuses not the son, but the nephews of Ermanaric, of seducing Ermanaric’s wife. Almost complete deterioration of the legend is represented in the late Low German ballad, Erminrikes Dot. Here only one incident of the Swanhild legend is retained: the attack on Ermanaric. And this one incident is greatly confused: the deed has been trans­ ferred to Dietrich von Bern; the motive, the attendant circumstances, the outcome of the attack: all have been changed. Indeed, it is chiefly through such traces of the Norse version of the legend as the role of the third brother and the gallows motif that we are able to recognize here a version of the old Swanhild legend. Thus in Germany no complete, or nearly complete, version of the Swan­ hild legend has been preserved, but only various dis­ torted traces of it. Various are the reasons for the deterioration of the legend in Germany. In the first place, since no German text even mentions the marriage of Swanhild to Er­ manaric, it is not at all improbable that the version of the legend which reached Germany did not know it; since none refers to her slaying, this incident must have been forgotten very early. Consequently, German poets would naturally have tried to interpret and explain the vengeance attack in terms of the legends which they knew. In the second place, the legend has passed from one Germanic people to another which has not only a wholly different legendary and cultural background, but also a wholly different conception of Ermanaric’s char­ acter. Norse poets, who developed the Swanhild legend and from whom German tradition obviously derived it, had the Ostrogothic conception of Ermanaric as the great king who slew his wife and only son only because

280

T H E LEGENDS OF E R M A N A R I C

he was foolish enough to trust implicitly the treacherous counselor. German poets, on the contrary, knew Ermanaric as an avaricious tyrant and the persecutor of his nephew Dietrich von Bern. This tale was of prime importance: all other deeds committed by Ermanaric were subordinate to it. And the final reason, the reason for the stage of deterioration represented in Erminrikes Dot, is that as the Middle Ages advanced, heroic legend passed from skilled court poets into the hands of the folk. Thus we can trace the progress of the Swanhild legend from its historical kernel, the death of Ermanaric and the fall of his kingdom in the year 375, through its develop­ ment in the hands of Norse poets, to its ultimate end in a ballad of the fifteenth century. The legend has lived long and traveled far, and both benefited and suffered by its travels. We come now to one of the most important and at the same time most controversial problems in the study of the Swanhild legend: the form in which the legend traveled and the paths which it followed as it spread from Germanic people to Germanic people. The story undoubtedly arose among the Ostrogoths in the Black Sea region shortly after the death of Er­ manaric. The emphasis on political aspects, on the treachery of the Rosomoni, indicate^ that the legend must have arisen while the Ostrogoths were still smart­ ing under the ignominy of defeat. Moreover, the whole story in the earliest extant version is told from the Ostrogothic point of view; the interest is in the fate of Ermanaric, not in that of the avenging brothers of Swanhild. We find our first extant version not in the Black Sea

CONCLUSION

281

region, however, but in Italy. Undoubtedly, it was car­ ried by the Ostrogoths when, under Theodoric, they mi­ grated from the east and set up the Ostrogothic kingdom in Italy. And it must have been from Ostrogothic tradi­ tion that Jordanes— or, rather, Cassiodorus before him— derived his version of the Swanhild legend. That the Swanhild story came into Cassiodorus’ hands in the form of a heroic lay is decidedly probable.3 An essential characteristic of the heroic lay is its unity and conciseness; it deals with a single incident or group of very closely related incidents, all other material being merely alluded to.4 This is precisely what we have in Jordanes’ account: a simple account of double ven­ geance, with no extraneous details. But the passage as we have it in Jordanes is not a close prose paraphrase of the lay itself. In the first place, it lacks dialogue, an essential characteristic of the heroic lay, so far as we can judge from the Hildebrandslied, the Finnsburg fragment, and the lays incorporated into their accounts by Paulus Diaconus and Saxo Grammaticus.5 In the second place, it lacks the dramatic qualities of the heroic lay. We must remember that Jordanes is giving little more than an epitome of Cassiodorus’ long history, and has omitted a great deal. Cassiodorus’ version of the Swanhild legend may, in form, have been very much like Saxo’s version of the Ingeld story or of the BjarkamaL Jordanes’ account is just what we should expect; he is clearly writing a summary of a story, and he omits no more than would anyone who is writing a synopsis: all the main outline is there, but the dramatic speeches we cannot expect. Obviously, the lay which lies behind Jordanes’ ac­ count, however, cannot lie behind the Norse version: had the story which came into the hands of Norse poets

282

T H E LEGENDS OF E R M A N A R I C

represented Swanhild as wife to a treacherous subject of Ermanaric and as justly punished for her husband’s treachery, they would scarcely have made her the daughter of the illustrious SigurSr and GuSrun. True, we might suppose that an earlier generation of Norse poets made her the wife of Ermanaric, and that later poets, composing after the Sigurc5r-Burgundian-Atli legend had become popular in Scandinavia, connected the two leg­ ends by making her the daughter of SigurSr and GuSrún. But there is an even more decisive indication that in the immediate source of the Norse version we do not have to do with the lay behind Jordanes’ account: the fact that there is no mention of the Huns in the Norse jQrmunrekkr legend. Since in Norse tradition the Huns are fiends, it is certainly not too much to suppose that had Norse poets ever known of Ermanaric’s enmity toward them, some trace of it would be preserved in the Norse story. It would seem, then, that we must assume the change from political to personal to have taken place before the legend ever reached Scandinavia. The question is, Where did this change take place? On this point critics have been perhaps more in accord than on any other phase of the development of the Ermanaric legends: with very few exceptions they agree that the story was remodeled in Germany and transmitted from there to Scandinavia. The fact that no complete version of the Swanhild legend is preserved in German texts, that only distorted traces of it occur in them, is regarded as a great pity; but that it must once have been widely known in Germany is regarded as proved beyond all doubt by the fact that it is known in Scandinavia. Thus Lawrence states: “ The Sunilda motive seems to have early faded out in German territory, though it seems

CONCLUSION

283

necessary to postulate its existence to account for the presence of the Swanhild story in Scandinavian.” 6 11Now, if the Swanhild legend could have reached Scan­ dinavia only by way of Germany, then, of course, we should have to assume that it did pass through Germany and that for some deplorable reason all traces of its pas­ sage have vanished. But such an assumption is not nec­ essary, for there are other routes the legend could have taken.! [ It could have been transmitted, as Boer has sug­ gested,7 directly from the Goths to North Germany and Scandinavia by way of the so-called “ culture stream” from the Black Sea to the Baltic, that is, the old trade routes up the Dnieper and the Vistula. This stream, by which all manner of cultural influences and cultural objects were sent by the Goths back to the old home­ land,8was presumably broken oif in the last quarter of the fourth century when the Huns, coming from the east, invaded the lands west of the route. But with the death of Attila and the end of the Hunnish domination it was reopened. Arne points out that after the death of Attila all the Goths did not settle down in Pannonia and later follow Theodoric into Italy, but a number of them settled around Kiev, where we find manifold traces of them. He speaks of especially rich archaeological traces of the Goths around GaluScinoborg from about the period 450-550. And these Goths were maintaining contacts between the Kiev district and East Prussia, and trading north along the Dnieper and the Niemen, be­ tween 450 and 550: fibula types found around Galu§cinoborg and elsewhere in the Kiev district appear also in Masuria in East Prussia.9 It is even possible that the trade was followed by actual northward movements of

284

t h e

l e g e n d s

o f

e r m a n a r i c

the Goths about the middle of the sixth century.10 Further, during this period from 450 to 550 there was open a route between the Goths of the Danube lands and Scandinavia which went from Carnuntum, between Pressburg and Vienna, up the March, down the Vistula to the Baltic, and thence along the coasts of East Prussia and Kurland.11 According to Shetelig, after the Hunnish advance, plunder and imperial subsidies cre­ ated a golden age for the northern peoples from about the middle of the fifth to the middle of the sixth centuries. He points out that gold coins minted by Theodosius, Leo, Zeno, and Anastasius were carried northward from the Goths to related peoples, from the Danube lands to the Baltic.12 Finds of these coins are especially rich on the lower Vistula and Gotland, Oland, Bornholm, and to a lesser degree on the Swedish mainland. Of the 500 from the period 395-565 which by 1920 had been found in Scandinavia and North Germany, the majority date from 450-550; of the 111 gold coins from the period 395565 found on Gotland, 42 are coins of Anastasius (491518).13 Lindqvist emphasizes that the presence of this gold in the southeastern parts of the north, especially on the three great Baltic islands, is usually regarded as a result of communication along the Vistula with the Goths after the fall of the Hunnish dominion during the third quarter of the fifth century.14 Summing up our archaeological evidence, then, what we find is this: specific indication of the existence of a Gothic settlement in the Kiev district during the period 450-55°; definite evidence of relations between South Russia and East Prussia in the period 450-550; abun­ dant evidence of relations between the Danube lands and the Swedish islands and mainland in the period

CONCLUSION

285

450-550. All this points clearly to the fact that though the old “ culture stream” was broken off by the Hunnish invasion in the last quarter of the fourth century, it was very obviously reopened with the end of Hunnish domi­ nation and remained open for approximately a century; and, further, it shows that there was also open another “ stream” to the north which merely skirted Germany. In short, this archaeological evidence definitely estab­ lishes the existence of relations between the Ostrogoths and East Prussia and Scandinavia a century after the death of Ermanaric. It would therefore have been quite possible for the Swanhild legend to have traveled di­ rectly from the Goths to Scandinavia by way of the old trade route through Russia or that up the March and the Vistula. Certainly, at least one other Ermanaric tradition would seem to have reached the north by one of these routes: the Ostrogothic conception of his character. The Wídsíð, North German tradition, and Saxo all know the Ostrogothic conception of the great king ruling over many tribes and of the generous dispenser of treasure. And we know that this tradition cannot have reached England, North Germany, and Denmark by way of Ger­ many, for it is altogether inconsistent with the German conception of his character. Another tradition, too, that of Ermanaric as foe of the Huns, probably traveled northward the same way. For although the Norse leg­ end knows nothing of his enmity toward the Huns, the JVtdsið poet makes it the central theme of his tale.15And certainly this tradition cannot have traveled northward by way of Germany, for the German attitude toward the Huns was decidedly sympathetic. Now, certain tradi­ tions may possibly have spread to the north before the

286

T H E LEGENDS OF E R M A N A R I C

old culture stream from South Russia was completely broken off. The Ostrogothic conception of Ermanaric’s character could have been transmitted even while he was still ruling. His reputation as foe of the Huns could have begun to travel northward even while he was mak­ ing his stand against them, but before they overran the whole of his territories and drove him to take his own life— the invasion, after all, was not accomplished in a day, but seems to have lasted over a number of years.16 The Swanhild legend, on the other hand, cannot have been transmitted until sometime after 450; for, obviously, if it did arise as a result of the reaction to the death of Ermanaric at a moment of national crisis, then it cannot have begun to develop until after his death in 375, which is likewise presumably the date of the interruption of the culture stream. It is here, on the question of the date of the transmission to the north, that Boer has erred. For he believes that the stream was broken off once and for all with the coming of the Huns at the end of the fourth century, and he must therefore assume that the story of the slaying of Swanhild and the consequent wounding of Ermanaric by her brothers passed up the culture stream before that time, that is, almost imme­ diately after Ermanaric’s death.17 This seems impossible to believe, for it means that the story must have trav­ eled to the north before it arose; some time must have been necessary for the tale to develop. A Goth hastening toward the old home on the Vistula within a short time after the Hunnish attacks would have come breath­ lessly in with the news of the great king’s suicide, of the atrocities committed by the invaders, of all the horrors of the invasion, rather than with the fully developed Swanhild legend. Furthermore, Boer proceeds to attrib­

CONCLUSION

287

ute to the story an order of development which is by no means usual in heroic poetry. According to his theory, those details which in the Jordanean and the Norse versions are identical— that is, the wounding ot Ermanaric and the slaying of Swanhild— must have passed up the culture stream before it was completely broken off; but the motive for the slaying of Swanhild which is given by Jordanes— the treachery of her hus­ band— must have developed after the stream was broken through. This, argues Boer, is the reason that the Norse and the Jordanean versions differ upon the motive: men did not know it, and Norse and Gothic poets indepen­ dently supplied one.18 It is, however, difficult to believe that heroic poets would first have invented the romantic tale of the slaying of Swanhild and the vengeance of her brothers, and then, some time later, have added the purely political motive: heroic poetry does not as a rule proceed from romantic to political and factual, but from political to romantic. It is much more reasonable to suppose that soon after Ermanaric's death the Ostrogoths began to develop their own story of it: the more or less political version of the Swanhild legend. Then, later, there arose a second stage, intermediate between Jordanes' version and the Norse. It had undoubtedly lost the political relationships of the figures, and dealt with conflicts between individuals rather than between nations. It represented Swanhild as Ermanaric's wife and as slain because the counselor falsely accused her of adultery with the king's son; but of the vengeance it told only that the brothers made an unsuccessful attack upon Ermanaric, nothing at all ot the circumstances of their attack nor of their own fate. The Ostrogoths could scarcely have developed this ver­

288

T H E LEGENDS OF E R M A N A R I C

sion, however, either while they were still under Hunnish dominion or while they were dwelling in Pannonia after the battle with Attila’s sons near the river Nedao. For it was the Jordanean version which they carried with them to Italy. In the first place, it must have been from the Goths in Theodoric’s household that Cassiodorus derived the story. In the second place, in all southern Germany, in those regions where Ostrogoths and their close associates dwelt, there is no trace of that version which represents Swanhild as wife to Ermanaric. This latter version, however, might well have grown up among those Goths who went over to Kiev and the sur­ rounding districts after Attila’s death. These Goths no less than Theodoric’s would have recovered from the ignominy of subservience to the Huns; just as much as Theodoric’s, they would have lost all recollection of the Rosomoni and the Huns in the Swanhild story. We might, then, attribute to these Gothic poets of the first generation after Attila’s death that tragic story of Ermanaric’s unhappy marriage, of the luckless love of his young wife and his only son, of the tragic misunder­ standing which caused him to kill them upon the unsup­ ported word of his counselor. And this story we might assume to have traveled northward sometime between, let us say, 475 and 550, directly from the Kiev Goths to East Prussia. There exists also another possibility, one which is per­ haps preferable. Both the trade routes, the one up the Dnieper from the Kiev Goths and that along the March and the Vistula, led to the same district in the north: East Prussia. And there was inhabiting East Prussia one people which would have had every reason to have taken an interest in the Ermanaric legends, a people, indeed,

CONCLUSION

289

which was long and intimately associated with the Goths: the HreiSgoths.19 It might have been the poets of these “ stay-at-homes” by the Vistula mouth, and not Ostrogothic poets at all, who remodeled the Swanhild legend. The Jordanean version could have been trans­ mitted to them either from the Ostrogoths in Pannonia or from those others who settled around Kiev; probably the latter, since they maintained especially lively rela­ tions with their kinsmen in East Prussia. These Vistula Goths, even more than the Ostrogoths, would soon have lost the political aspects of the story, the political rela­ tionships of the characters. Living as they did in the farthest reaches of Ermanaric’s empire, they would not have felt so keenly the downfall of the kingdom; they would have been the more likely to have realized more fully the personal tragedy of the Gothic king. They would have forgotten much sooner the connection of the woman Swanhild and her avenging brothers with the Rosomoni, for that faithless tribe would never have meant anything to them. If it is to these poets that we are to attribute the change from political to personal in the Swanhild legend, then we must assume the Jor­ danean version to have reached them very shortly after 450, the approximate date of the reopening of the culture stream; for a certain amount of time must be allowed for the development of the new story, the new relationship between Ermanaric and Swanhild, the new motive, the new conflicts brought about by new characters, before all this was passed on to Scandinavia. If we look at our Norse version, we see that 450-550 is decidedly the most suitable date for the arrival of the story in Scandinavia. The Norse legend appears to be heavily influenced by OSinn worship.20 The son of jQr-

29O

T H E LEGENDS OF E R M A N A R I C

munrekkr is hanged, the jQrmunrekkr of the Hamðismál vows that he will hang HamSir and Sc^rli: hanging is the characteristic method of sacrifice to OSinn. In the battle in the hall OSinn intervenes and saves the Goths from complete destruction at the hands of the two brothers: OSinn is god of the Goths. He does not intervene until after the two have cut off the hands and feet of JQrmun­ rekkr: OSinn often allows his chosen warriors to be mortally wounded that he may receive them into Val­ halla. Hence it would be only reasonable to assume that the Swanhild legend must have reached Scandinavia at a date when the OSinn cult was especially flourishing. This date, as the Danish bog deposits bear witness, was precisely the period from 450 to 550.21 Especially valu­ able is the evidence of the Kragehul spear shaft. The inscription, dating from about 550,22contains the proper name AnsugisalaR, “Hostage of the As,” that is, of OSinn.23 This implies that the story of the war between the Æsir and the Vanir, and its compromise and ex­ change of hostages, was known at the time of the inscrip­ tion. It implies, in fact, that the story of the coming of the Æsir to Scandinavia, their war with the native gods of the Vanir line, and the exchange of hostages, must have been known in the sixth century very much as it was to Snorri. That the Swanhild legend was known in England can­ not be conclusively proved: the Widstð poet has some knowledge of a wife of Ermanaric, but there is no defini­ tive evidence that she is to be identified with Swanhild; nor, indeed, any real literary evidence that the Swanhild legend was known to Anglo-Saxon poets. If it was, how­ ever, then with respect to Anglo-Saxon tradition, too, sometime between 450 and 550 is the best possible date

CONCLUSION

29I

for the arrival of the legend in the north. It was precisely during the sixth century that the stories related in the Beowulf were transmitted to England, probably from the continental home of the Angles.24 As Chambers has pointed out, the poem shows a full knowledge of those events which took place in Scandinavia during the period from 450 to 530, whereas it contains no reference to any event to which we can with certainty assign a later date.25 Now, if the stories related in the Beowulf could have been carried over to England by the Angles in the sixth century, certainly the Ermanaric traditions could have been transmitted at the same time and in the same way. And it is reasonably certain that the Ostrogothic conception of Ermanaric’s character and the representa­ tion of him as foe of the Huns, both of which were known to the Wtdstð poet, were carried thus. The Swanhild legend also could have been. On the other hand, it could have reached England not from the continental home of the Angles, but directly from Scandinavia. In this case, too, it could very well have been transmitted in the course of the sixth century, for there is abundant archaeological evidence of lively connections between England and Norway in the latter part of that century. Bent animal bodies, with uplifted necks and reversed heads, charac­ teristic of the type of ornamental design called by archaeologists Style I,26 and certain indications of the later type known as Style II,27 are especially marked in Anglo-Saxon ornamentation.28 There is evidence from certain buckle types that this influence continued into the seventh century.29 To recapitulate: the version of the Swanhild legend which represented Swanhild as Ermanaric’s wife and her reputed love for his son as the motive for her slaying

292

T H E LEGENDS OF E R M A N A R I C

may have grown up among the Ostrogoths of the Kiev region in the first generation after Attila’s death and have traveled northward to Scandinavia by way of the old trade route through Russia sometime between 475 and 550; or it may have been developed by HreiSgothic poets from the Jordanean version, which had reached them by way of one of the trade routes to East Prussia in the second half of the fifth century. Sometime be­ tween 475 and 550, more probably between 500 and 550 (especially if the HreiSgoths were the ones to remodel it), the legend took root in Scandinavia and began to de­ velop in the hands of Norse poets. Possibly in the second half of the sixth century or the beginning of the seventh it was carried to England either from Norway or, together with the Ostrogothic conception of Ermanaric’s character and his reputation as foe of the Huns, from the continental home of the Angles. Thus far we have dealt with the transmission of the Swanhild legend in the Age of the Migrations. But there was also a second period of transmission when in the Viking Age the Norse version was carried into North Germany. \ 1Certain traces of the Swanhild legend in Germany can be explained only on the basis of the assumption that the Norse version traveled to North Germany from Scandinavia. If we found in German texts only the slay­ ing of the son, the figure of the evil counselor, and the vague recollection of the seduction of Swanhild in the þiðrikssaga, then possibly we might assume that the second stage of the story might have spread directly from East Prussia through northern Germany. But we find, in addition, certain incidents which can be explained only as distortions of that version of the legend which

CONCLUSION

293

had been developed by Norse poets: the Annales account of the attack on Ermanaric by the three avenging brothers; the obvious distortion of the Hamðismál story in Erminrikes Dot; the circumstances of the slaying ot Samson, the youngest son, in the þiðrikssaga. 11 Those critics who have argued that the Swanhild legend was transmitted from Germany to Scandinavia have emphasized the commerce between Germany and Scandinavia in the ninth century. This factor must be considered, of course: that German merchants several centuries later did carry North German legendary tra­ dition to Scandinavia is attested by the piðrikssaga, written down in Norway, but setting forth German tra­ dition; and presumably eighth- and ninth-century mer­ chants likewise carried their stories with them. On the other hand, we have strong indication that as potent influence was exerted by Scandinavia upon Ger­ many as by Germany upon the North.30 For the late eighth, the ninth, and the early tenth centuries were dis­ tinguished by lively political relations between Scandi­ navia and not only Germany but the entire Frankish Empire. That in these relations the Scandinavians took the more active part is made very evident by the pres­ ence of Scandinavian mercenaries in the service of Charlemagne, and by the constant assaults of hordes of vikings upon the Frankish Empire, in the course of which several long-continuing Scandinavian settlements were made in Frankish territory. These settlements were mainly of two kinds: camps established at strategic points to serve as bases of operation and as winter quar­ ters; outright grants of land offered as bribes to viking chieftains by badly frightened emperors. They were in various parts of the empire, but we are concerned pri­

294

T H E LEGENDS o f e r m a n a r i c

marily with those in Frisia and Saxony. Frisia particu­ larly was for a long time in a sense little more than a colony of the Danes, sometimes only partially, some­ times almost completely under viking dominion. The occupation began in the early years of the century, when in 807 a chieftain named Halfdan (exact provenience un­ known) with a considerable following took service with Charlemagne and received a grant of lands in Frisia. In 826 Harald, an unsuccessful claimant to the Danish throne, was baptized in Mainz, and a little later he settled in lands at the mouth of the Weser which he had received from Louis the Pious. Later, this same Harald was granted Walcheren, and his brother Rörik received Dorstad. This Rörik in 857 won for himself land near Hedeby and seems to have made himself lord of a large and important realm which included most of Schleswig and North Frisia. It was not until 885, when the viking Godfred, whom Charles the Fat had bought off with a grant of Frisian lands (probably Rörik’s old fief), made impertinent demands of Charles for the wine districts of Coblenz and Andernach, that Frisia was freed from Danish dominion.31 Surely the Scandinavian settlers in Germany carried their legends with them. And that they did do so is sug­ gested by the appearance in German legend of such figures as Fruote von Tenemarken {Rabenschlackt) and in the Charlemagne cycle of Ogier the Dane, and of such stories as that of Hetel (HeSinn), Hild, and Hagen (HQgni) in the MHG Kudrun. It can, then, hardly be denied that the version of the Swanhild legend current in Scandinavia in the early ninth century could have been carried to North Germany and there distorted into the forms which have come down to us in the German texts.

CONCLUSION

295

The other Ermanaric legends, of Ermanaric as perse­ cutor of Dietrich von Bern and as slayer of the Harlungs, had a history very different from that of the Swanhild legend. Neither of them grew up around the figure of Ermanaric, but undoubtedly both were originally inde­ pendent legends which came to be attached to him at a comparatively late date. That of the persecution of Dietrich we know to have been originally independent, for the Hildebrandslied preserves a version of the story which represented Odoacer, not Ermanaric, as the persecutor. It came to be attached to Ermanaric because of the tra­ ditions about him current in South Germany in the eighth century: that he was an avaricious tyrant and a foe of the Huns. Probably the Harlung legend, too, was once independent, perhaps connected with some other Ostrogothic hero; and it came to be attached to Erman­ aric because he was known as an avaricious tyrant, driven to all manner of evil deeds by his lust for gold and land, and as a heartless slayer of kin. In all probability it was not attached to him until after he had come to be well known in Germany not only as the persecutor of Dietrich von Bern, but also as the slayer of his son. The pattern once established, he was able to pick up both sons and nephews, whom he slew. The evidence of the Wídsið indicates that at its date the Harlungs had not yet come to be regarded as his slain nephews. Further, these legends are purely German, German in origin and not widely known outside of Germany. The Harlung story was known to Saxo, but that it ever came to have a Danish version is unlikely; it is more likely that Saxo derived it directly from German tradition.The version known in England seems to have been the origi­ nal story, before it was connected with Ermanaric. The

296

T H E L EGENDS OF E R M A N A R I C

legend of Ermanaric as the persecutor of Dietrich von Bern seems never to have been known outside of Ger­ many. Anglo-Saxon tradition may have known the story of Dietrich’s exile, if the J)eodric of Deor is to be taken as the Goth, but, so far as can be determined from the evi­ dence, did not know Ermanaric’s connection with it. From this investigation of the several versions of the Ermanaric legends we have seen that three historical facts of the career of Ermanaric the Ostrogoth exerted a heavy influence on the three legends which grew up around him in the Middle Ages. The historic Ermanaric ruled over an extensive kingdom, embracing many sub­ ject tribes; when the Huns invaded his kingdom, he for some time maintained a stout resistance to them; at last, worn out with grief and fear, he committed suicide. So much is history, as we learn it from Ammianus Marcel­ linus. The legendary Ermanaric ruled over an extensive kingdom which, according to the German and Danish versions of the legends, embraced the lands of those he had conquered; he hanged his nephews the Harlungs and seized their land and gold (German); he was a tyrant over his subjects (German, Anglo-Saxon) and an evil truce breaker (Anglo-Saxon); he persecuted his nephew Dietrich von Bern, friend of Attila the Hun (German); he waged war against the Hunnish forces (German, Anglo-Saxon); he was a great and generous king (AngloSaxon, Ostrogothic); he was a great and generous king who degenerated into an evil tyrant and kinsman slayer and was brought to his ultimate downfall through the machinations of his evil counselor (Danish, North Ger­ man); he was at one and the same time “ more generous than all in gifts, more cunning than all in guile” (North German, Anglo-Saxon); he was the most noble of all the

CONCLUSION

297

kings of his line, but he brooked no treachery from subjects (Ostrogothic); he slew his wife and only son at the instigation of the evil counselor (possibly Ostrogothic, possibly HreiSgothic; Norse; Danish; possibly AngloSaxon). Such is the character and such are the deeds of the Ermanaric of legend. In this study I have tried to show how the vast body of conflicting tradition grew up from the three historical facts. The fact of Ermanaric’s conquests gave rise to the two conflicting traditions that he was a great and gener­ ous king, the most noble of his line, and that he was an evil tyrant and a truce breaker. The reason for the great discrepancy here is the point of view of the people telling the story: the peoples whom he crushed would naturally have had an opinion of him very different from that held by his own people, whose fortunes and glory the great conqueror had increased. I have tried to show that these conflicting tradition's concerning Ermanaric’s character were, perhaps more than any other single factor, re­ sponsible for the tenor of the later Ermanaric legends: where only Ostrogothic tradition was known he was a great and generous king who committed evil deeds only at the instigation of the counselor whom he trusted im­ plicitly; where only non-Ostrogothic (subject-people) tradition existed he was a cruel tyrant and oppressor; where the two traditions came together various attempts at reconciliation were made, one of them, Saxo’s, fairly convincing because it was the work of a great “ histo­ rian,” the others distinguished by all sorts of inconsist­ encies. We may conclude that the point of view and the legendary, cultural, and historical background of each of the peoples telling the story were the most influential factors in the development of the Ermanaric legends.

N O TES TO C H A P T E R VI 1 B e o w u lf, 11 . 2024 ff. 2 Cham bers, B eo w u lf: A n In tro d u ctio n to the S tu d y o f the P o e m , p. 442. 3 C f. Heusler, D i e altgerm anische D ichtu n g,, p. 153: “ Eine Heldensage ist a ls L i e d entstanden, hat sich im L ie d e verbreitet und w eitergebildet.” C f. also pp. 20 f., 150, 1 52, and p a s s im ; H o o p s, II , 494. 4 Heusler, D i e altgerm an ische D ich tu n g ,t pp. 158 f.

5I b id ., p. 147. 6M P , IV , 3 5 4 , n . 2 . 7 D i e Sagen von E r m a n a rich u n d D ie tr ic h von B e r n , pp. 11 ff. M o gk has suggested (in Festgabe f ü r R u d o lf H ild e b r a n d , pp. 5 ff.; G rd r., I I , 624) that the Heruli, returning to Scandinavia after the overthrow o f their kingdom on the Danube by the Lombards shortly after 500, carried G othic legends w ith them. I t is difficult to understand, however, how they could have transmitted to the north the conception o f Erm anaric’s character which is preserved in the Norse version o f the Swanhild legend. For the H eruli had every reason to regard Ermanaric as a cruel and vicious tyrant. 8 C f. in particular Salin, S tu d ie r tillä g n a d e O scar M o n te liu s , pp. 136 ff.; D i e altgerm an ische T h ie r o rn a m en tik, esp. pp. 12 ff., 354 ff., but sl\so p a s s im . Schück has suggested ( U U Å , 1918, pr. 3:2, 14 f.) that this m ay have been the route by which the tales o f the G othic-H unnish wars— another legend which is known in Scandinavia but not in G erm any— were transm itted to the north. 9 S tu d ie r tillä g n a d e O scar A lm g r e n , p. 268 (R ig , 1920, 84); D e fo r h is to r is k a T id e r i E u r o p a , I, 437; cf. also p. 443; D e t stora S v itjo d , pp. 13 f., 27 ff. 10 Arne {D e fo r h is to r is k a T id e r i E u r o p a , I, 437) suggests that after the fall o f the little G othic kingdom on the Dnieper in the 550^ a number o f G oths m ay very possibly have sought refuge among their kinsfolk in the north, w ith whom they had even earlier maintained relations. In D e t stora S v itjo d (p. 35) he suggests the return o f certain southern G oths to Öland c a . 480, to G otland in the beginning o f the sixth century. 11 Von Friesen, U U Å , *924,68. 12 D e fo r h is to r is k a T id e r i E u r o p a , II, 345 f. 13 On these cf. esp. Nerm an, D e t sven ska rik e ts u p p k o m st, pp. 181 ff. and the references there cited; D ie V erb in d u n g en zw ischen S k a n d in a v ien u n d dem O st­ b a ltik u m in d er jü n g e r e n E is e n z e it, pp. 31 ff.; L in d q vist, F o rn v ä n n en , 19 17, 140 f.; Arne, F o rn v ä n n en , 1919, 107 ff.; A c ta A re h a e o logica, II, 1 ff., esp. 19 ff.; Bolin, F y n d e n av rom erska M y n t i d et f r i a G e rm a n ia , p a ssim . 14 S tu d ie r tillä g n a d e O scar A lm g ren , p. 66 {R ig , 1919, 66). 16 C f. m y article in U n iv. Calif. Publ. English, I I I , 225 ff. 16 C f. H odgkin, I ta ly a n d H e r In v a d ers, I, i , 247, n. 2. 17 D i e Sag en , p. 12.

[>98 3

CONCLUSION

2 99

18 I b id ., pp. 10 ff. 19 On them cf. esp. von Friesen, R ö k sten en y pp. 108 ff., esp. pp. 122 f f , and the references there cited. 20 C f. m y article in P M L A , L V , 910 ff. 21 On the relation o f the deposits to the 0 I (Iowa C ity , 1922), 128-136. ---------- . “ Vic5ga in Scandinavian Hero Legend ,” S S N , V I (M enasha, 1920), 7 5 -8 1. L a r s e n , S o f u s . “ Saxo Gram m aticus, hans Værk og Person,” Aarbger, 1925,

1-286. L a w r e n c e , W i l l i a m W i t h e r l e . “ Structure and Interpretation o f W id s ith ,”

M P , I V (Chicago, 1906-1907), 329-374. ---------- . “ T h e S o n g o fD e o r ,” M P , I X (Chicago, 1 9 11 -19 1 2 ), 23-45. L each , H en r y

G o d d a r d , and S c h o e p p e r l e , G e r t r u d e .

“ Haraldssaga

H ringsbana and the Tristan and Svanhild Rom ances,” S S N , II (Urbana, 19 14 -19 15 ), 264-276. L e i t z m a n n , A l b e r t . “ Dietrichs Flucht und R abenschlacht,”

Z fd P h , L I

(Stuttgart, 1926), 46-91. L o r e n z , H e r m a n n . “ D as Zeugniss für die deutsche Heldensage in den

Annalen von Quedlinburg,” G e r m a n i a ,K X K I (Wien, 1886), 13 7-15 0 . M a l o n e , K e m p . “ Aki Qrlungatrausti,” S a g a -B o o k o f the V ik in g S o cie ty , X I

(Coven try, 1936), 26-29. ---------- . “ Becca and Seafola,” E S t , L X X I I I (Leipzig, 1939), 180-184. ---------- . “ E alh hild ,” A n g lia , L V (Halle [Saale], 1931), 266-272. ---------- . “ Herlekin

and Herlewin,”

E n g lis h

S tu d ie s , X V I I

(Am sterdam ,

J9 3 5 )> 141-14 4 ---------- . “ Hliþe and HlQ(Sr,” A P h S , V I (København, 19 31-19 32), 328-331. ---------- . “ A N o te on W id s ith 9a,” B eib la tt z u r A n g lia , X L I X (Halle [Saale], i

9 3 8 ) ,3 7 5 - 3 7 6 -

---------- . “ Secca and B ecca,” S tu d ia germ a n ica tillä g n a d e E r n s t A lb i n K o c k d en 6 D ecem ber 1 9 3 4 , pp. 192-199. Lund, 1934. ---------- . “ Th e Tale o f G ea t and Mæfthild,” E n g lis h S tu d ie s, X I X (Amster­ dam, 1937), 193-199. ---------- . “ T h e Theodoric o f the R ö k Inscription,” A P h S , I X (K øbenhavn, i

9 3 4 - I 9 3 5 ) } 7 6" 84 .

---------- . W id s ith . London, 1936.

B I BL I OG R A P H I ES

311

M a l o n e , K e m p . “ W id s ith and the C ritic,” E L H , V (Baltimore, 1938), 49-66.

---------- . “ W id s ith and the H erva rarsa ga,” P M L A , X L

(Menasha,

1925),

769-813. M a r q u a r t , J o s e f . O steu rop ä ische u n d osta sia tisch e Streifzü g e. E th n o lo g isch e

u n d h istorisch -top og ra p h isch e S tu d ie n z u r G eschichte des 9. u n d 10 . J a h r ­ hu n derts (ca. 8 4 0 -9 4 0 ). M it U n terstützung der Königl. Akadem ie der Wissenschaften zu Berlin. Leipzig, 1903. —

-— . “ Studien zum W idsfö,” F e s ts c h r ift V ilh e lm T h om sen z u r V ollen d u n g des siebzig sten L eb en sja h res pp. 9 8 -110 . Leipzig, 1912.

M a t t h a e i , G e o r g . “ Die

bairische H unnensage in ihrem Verhältnis zur

Am elungen- und N ibelungensage,” Z f d A , X L V I (Berlin, 1902), 1-60. — ■— — . “ R üdiger von Bechlaren und die Harlungensage,” 7f d A , X L I I I (Berlin, 1899), 305-332. M e i e r , J o h n . “ Drei alte deutsche B alladen,” J a h r b u c h fü r V olkslied fo rsch u n g y

IV (Berlin und Leipzig, 1934), 1-65. M i k k o l a , J. J. “ D ie Nam en der Völker H ermanarichs,” F in n is c h -U g r is c h e

F o rsch u n g en , X V (Helsingfors, 1915), 56-66. M o g k , E u g e n . N o rw e g isch -islä n d isch e L ite r a tu r , G ru n d riss der germ a n isch en

P h ilo lo g ie , herausgegeben von Hermann Paul, II , 1, 555-923. 2. verbesserte und verm ehrte Auflage. Strassburg, 1901-1909. M o l b e c h , C . “ Saxo Gram maticus, hans Historie og Charakteer, og hans

lærde Sam tidige i Danm ark, under Valdemarernes Tidsalder,” H is to r is k T id ssk rifty 2 R ., 5 B. (Kjøbenhavn, 1854), 1-12 4 . M u c h , R u d o l f . “ Widsith. Beiträge zu einem Com m entar,” Z f d A , L X I I

(Berlin, 1925), 113-15 0 . ---------- . “ W lstlaw udu,” H o o p s , IV (Strassburg, 19 18 -1919 ), 555. M u d r a k , E d m u n d . D ie deutsche H elden sa ge.

Berlin, 1939. Jahrbuch für

historische Volkskunde, V I I. M ö l l e n h o f f , K a r l . “ Die austrasische D ietrichssage,” Z f d A y V I (Leipzig,

1848), 435-459. ---------- . B eo v u lf. U n tersuchun gen über d as a n g elsäch sische E p o s u n d d ie älteste G eschichte der g erm a nischen Seevölker. Berlin, 1889. ---------- . “ F rija und der H alsbandm ythus,”

Z fd A , X X X

(Berlin, 1886),

217-260. ---------- . “ Zeugnisse und Excurse zur deutschen H eldensage,” Z f d A , X I I (Berlin, 18 6 ^ 18 6 5 ), 253-386; 4 i 3 “ 4 3 6* ---------- . “ Zur Geschichte der N ibelungensage,” Z f d A , X (Berlin, 1855-1856), 146-180. ---------- . “ Zur K ritik des angelsächsischen Volksepos,” Z f d A y X I

(Berlin,

1856-1859), 272-294. M ü l l e r , P e t e r E r a s m u s . “ Om Kilderne til Saxos ni forste Böger og deres

Trovæ rdighed,” D e t kongelige d a n sk e V id en sk a b e rn es S elsk a b s p h ilo s o p h is k e og h isto risk e A fh a n d lin g e r , 4 R ., 2 B. (Kiöbenhavn, 1824), 1 -1 7 4 .

312

T H E LEGENDS OF E R M A N A R I C

M ü l l e r , W i l h e l m . M y th o lo g ie der deutschen H eld en sa g e. Heilbronn, 1886. N e c k e l , G u s t a v . B eiträ g e z u r E d d a fo rs ch u n g m it E x k u r s e n z u r H eld en sa g e.

Dortm und, 1908. ---------- . “ Eddaforschung,” Z e its c h r ift f ü r den d eu tsch en

U n terrich t, X X X

(Leipzig und Berlin, 1916), 1—16; 81-98; 162-170 . ---------- . “ D er W ert der isländischen Literatur besonders für die Erkenntnis germanischer F rühzeit,” D eutsch e Is la n d fo r s c h u n g , I, 1-62. Breslau, 1930. Veröffentlichungen der Schleswig-Holsteinischen Universitätsgesellschaft, N r. 28, i. ---------- . “ Zu den Eddaliedern,” A r k iv , X L I I I (Lund, 1927), 358—373. N e r m a n , B i r g e r . “ Försök till datering av Reidgoternas konungaätt,” F e s t-

s k r ift t i l F i n n u r J ó n sso n 2g. M a j ig 2 8 y pp. 206-212. København, 1928. N o r e e n , E r i k . “ Studier i fornvä s tnordisk diktning, I I I ,”

U U A , 1923.

Filosofi, språkvetenskap och historiska vetenskaper, 3. N o r m a n , F . “ ‘D eor’ : A Criticism and an Interpretation,” M L R , X X X I I

(Cam bridge, i 9 3 7 )» 3 7 4 “ 3 81O l r i k , A x e l . “ E ddam ythologie,” N e u e J a h rb ü ch e r f ü r das k la ssisch e A lte r ­

tu m G esch ich te u n d d eutsche L ite ra tu r , X L I (Leipzig, Berlin, 1918), 38-48. ---------- . “ Glem som hed,” D a n s k e S tu d ie r , 1912 (K øbenhavn), 1 4 5 -1 5 1. ---------- . K ild e r n e t i l S a k ses O ld h isto rie, en litera tu rh isto risk U n dersøgelse. 2 Bde. København, 1892-1894. ---------- . “ M ärchen in Saxo G ram m aticus,” Z d V e r f V o lk s ., II (Berlin, 1892), 117 -12 3 ,2 5 2 -2 5 8 ,3 6 7 -3 7 4 . ---------- . “ Tvedelingen a f Sakses kilder, et genm æle,” A r k iv y X I V

(Lund,

1898), 47-93. O l r i k , J ø r g e n . “ Studier over Tilblivelsen a f Sakses Væ rk,” H is to r is k T id s -

s k r ift, 8 R ., 2 B. (K ø b e n h a v n , 1909-1910), 211-2 6 3 . O l s o n , E m i l . “ Ragnarsdrapa 5. 1-4 . E n textkonjektur,” S tu d ie r tillä g n a d e

A x e l K o c k ypp. 525-530. A rk iv för nordisk filologi tilläggsband till band X L n y följd. Lund, 1929. P a l u d a n - M ü l l e r , C . “ Studier til B enyttelse og Bedømmelse a f nogle K il­

deskrifter til nordisk Historie. I I I. Bidrag til K ritik a f Saxo’s Historie­ væ rk,” H is to r is k T id s s k r ift, 4 R ., 5 B. (K jøbenh avn , 18 75 -18 77), 333-381. P a n z e r , F r i e d r i c h . D eutsch e H eld en sa g e im

B r e is g a u . Heidelberg, 1904.

N eujahrsblätter der Badischen Historischen Kommission, N eue Folge, V II. P a t z i g , H e r m a n n . D ie tr ic h von B e r n u n d s e in Sa g en kreis. D ortm und, 1917. P e t e r s e n , N . M . B id r a g t i l den d a n sk e L ite ra tu rs H isto r ie . 5 Bde. 2. U dgave.

K jø b e n h a v n ,18 6 7-18 71. P e t s c h , R o b e r t . “ D ie tragische Grundstim m ung des altgermanischen H el­

denliedes,” A u fs ä tz e z u r S p ra ch - u n d L itera tu r-G esch ich te W ilh e lm B r a u n e zu m 20. F e b r u a r ig 2 0 dargebracht, pp. 36-46. Dortm und, 1920. P i p p i n g , H u g o . “ Eddastudier I I ,” S N F y X V I I . 3 (Helsingfors, 19 26),1-13 8 .

BI B L I OG R A P HI ES

3 X3

P r i c e , R . ‘‘Additional N otes to the Song o f the Traveller,” in John Josias

Conybeare, I llu s tr a tio n s o f A n g lo -S a x o n P o e tr y , pp. 275-281. London, 1826. R a n is c h , W il h e l m . Z u r

K r i t i k u n d M e t r ik der H a m þ ism á l. Inaugural-

Dissertation. Berlin, 1888. R a s z m a n n , A u g u s t . D ie deutsche H eld en sa g e u n d ihre H eim a t. 2 Bde. 2.

Ausgabe. H annover, 1863. R e i c h a r d t , K o n s t a n t i n . “ Beiträge zur skaldenforschung I und I I ,” A r k iv ,

X L V I (Lund, 1930), 32-62; 199-257. ---------- . “ R a g n a rsdrapa 5,” Z f d A , L X V I (Berlin, 1929), 200-202. ---------- . “ Studien zu den Skalden des 9. und 10. Jahrhunderts,” P a la e str a , C L I X (Leipzig, 1928). R e u s c h e l , H . “ Wie ein Fuss dem andern. Zum A ufbau der H am Sism ál,”

P B B , L X I I I (Halle [Saale], 1939), 237-249. R i e g e r , M . “ Dietrich und Theoderich,” Z f d M , I (Göttingen, 1853), 229-235.

---------- . “ Ingävonen Istävonen H erminonen,” Z f d A , X I (Berlin, 1856-1859), 177-20 5. R o e d i g e r , M a x . “ D ie Sage von Ermenrich und Schwanhild,” Z d V e r f V o lk s .,

I (Berlin, 1891), 241-250. S c h n e i d e r , H e r m a n n . G erm a n isch e H eld en sa g e, G ru n d riss d er g erm a n isch en

P h ilo lo g ie unter M itw irkung zahlreicher Fachgelehrter, begründet von H er­ mann Paul, X / 1 - 3 . Berlin und Leipzig, 1928-1934. ---------- . “ Studien zur H eldensage,” Z f d A , L I V (Berlin, 1913), 339-369. S c h r ö d e r , E d w a r d . “ D ie Heldensage in den Jahrbüchern von Quedlinburg,”

Z f d A , X L I (Berlin, 1897), 24-32. Sch ütte,

G udmund.

“ A n gan ty-k vad ets

geografi,”

A r k iv ,

XXI

(Lund,

1904-1905), 30-44. ---------- . “ D ansk Kongetals tusendaarige T ra ad ,” H is to r is k T id s s k r ift, 9 R ., 5 B. (K øbenhavn, 1926-1927), 23-76. ---------- . “ T h e G eats of Beowulf,” J E G P h , X I (Urbana, 1912), 574-602. ---------- . “ D ie N ationalität der Baninger,” P B B , L V I I (Halle [Saale], 1933), 23 ^ 2 3 9 . ---------- . “ Tendensdigtning i H eltesagn et,” D a n s k e S tu d ie r , 1934 (K øben­ havn), 14 5 -16 5. ---------- . “ Saxes norske Skjoldungrække og V idsid,” (Norsk) H is to r is k T id s ­ sk r ift, 5 R ., 6 B. (Oslo, 1927), 353-364. ---------- . “ Vidsid og slægtssagnene om H engest og A n gan tyr,” A r k iv , X X X V I (Lund, 1 9 1 9 - 1920), 1-32. S ij m o n s , B. H elden sa ge, G ru n d riss der germ a n isch en P h ilo lo g ie , herausgegeben

von H ermann Paul, I I I , 606-734. 2. verbesserte und vermehrte Auflage. Strassburg, 1900. ---------- . R eview , Jiriczek, D e u tsch e H eld en sa g en , I , 33I - 352*

G G A , 1900 (Berlin),

SM

T H E LEGENDS OF E R M A N A R I C

S i j m o n s , B . "‘D as niederdeutsche Lied von K ön ig Ermenrichs T o d und die

eddischen Hamþésm 53> *37> l88> 26o> 285; in Saxo, 124, 133 f., 137, 141 f., 166, 177, 187, 188; in p id r ik s s a g a } 166, 174, 186 f., 187 f., 188 f., 189 f., 198, 199; in W id std y 1 4 9 ,1 6 6 ,1 6 7 ,1 7 7 ,1 8 7 ,1 8 8 Erm anaric: conquests of, 1 f., 2 f., 137, 165, 166, 184 f., 262, 296, 297; death of, T, 2, 3 f., 5, 15 ff., 23, 25 f., 51 f., 77, 103 f., 175, 179, 180, 199 f., 220, 234 ff., 246, 251, 254, 260, 261, 271 f., 274, 286, 292 f., 296; dual personality of, 166 f., 168, 186 ff., 198, 260; as foe o f the H uns, 1, 2, 3, 100 f., 149, 162, 163 f., 167, 192, 258, 262, 285, 286, 291, 295, 296; as generous ring giver, 137, J49> ^ 2 , 164, 165, 188, 189, 260, 285, 286, 291, 297; as king o f the Danes, 123, 127 ff., 138 ff., 142, 144, 183; as king o f the HreitSgoths, 149, 162, 163, 183; as king o f Spain, 183, 263; as kinsman slayer, 53, 102, 134, 141, l 9S> l 99 f-» 2 1 8, 249, 255 f., 259, 264, 295; as leader o f a group o f renowned heroes, 149, 162, 163, 164, 167; as persecutor o f D ietrich von Bern, 178, 180, 189, 190 ff., 218, 228, 246, 249, 258, 259, 260, 262, 295 f.; as Rom an Emperor, 183 f.; as slayer o f the H arlungs, 136, 137, 140, 178, 188, 189, 205 ff., 218, 258, 259, 260, 295; as slayer o f his son, 23, 52 f., 54, 140, 178, 180, 188, 190, 195 ff., 218, 250 ff., 253 f., 255 f., 258, 260, 261 f., 295; as slayer o f his wife, 23 f., 52 f., 253, 287; as tyrant, 3, 103, 134, 137, 149, 161, 162, 164 f., 165 f., 167, 177, 184 ff., 188, 189, 192, 218, 220, 249, 252, 258, 259, 262, 295, 297; as wrap wcerlogai 149, 162, 164, 165 f., 167; see a lso Ermanricus, Erminrikr, Erm rich, Jarmericus, jQrmunrekkr Ermanricus, 103, 146 f., 190, 195, 199, 205, 220, 226, 234; character ,of, 186, 1 89; see also Erm anaric, A n n a le s Ermelricus, 183 Ermenrich, 105, 210, 223

IN D EX

S3!

Ermenricus, 269 E r m in r ik e s D o ty 104, 179, 180, 219, 234, 236 ff., 240, 24T, 246, 249, 279, 293; relation to H am 'Sism ä l, 104, 237 ff., 254 Erminrikr, 60, 179, 191, 195 f., 197, 198, 205 f., 216, 219, 221, 222, 225, 227, 228, 230, 233, 240, 253; character of, 186 ff., 190, 198; s e e a lso Erm anaric, p tö r ik s s a g a Erm rich, 19 1, 196, 199, 201, 202, 203, 204, 205, 219, 220, 221, 222, 223, 228, 229, 250 f., 252; character of, 180, 184; see also Ermanaric, D ie tr ic h s F lu c h t Erntrich, 206, 251 Erpam ara, 215 E rp f, I0 4 f. Erpr, son o f Gu