Nibelungen Prosody [Reprint 2019 ed.] 9783110811667, 9789027930668

Table of contents :
TABLE OF CONTENTS
PREFACE
I. THEORETICAL BACKGROUND
II. THE NIBELUNGEN STANZA: GENERAL FACTS
III. LEXICALITY AND PROSODIC DISTINCTIVENESS
IV. THE STRESS FEATURE
V. THE BESCHWERTE HEBUNG
APPENDIX I
APPENDIX II
APPENDIX III
APPENDIX IV
APPENDIX V
BIBLIOGRAPHY
INDEX

Citation preview

DE

PROPRI ETATI BUS

LITTERARUM

edenda curat

C. H. VAN SCHOONEVELD

Indiana University Series Practica,

112

NIBELUNGEN PROSODY

by RAY M.WAKEFIELD

1976 MOUTON THE HAGUE - PARIS

© Copyright 1976 Mouton & Co. B.V., Publishers, The Hague No part of this book may be translated or reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publishers.

ISBN 9 0 279 3 0 6 6 x

Printed in the Netherlands

TABLE OF CONTENTS

PREFACE 1 I. Theoretical Background 5 1. Pre-Heuslerian Theories of MHG Prosody 6 2. The Nibelungen "Feud" 8 3. Heusler and Lachmann Compared 9 4. The Domination of Heusler's Theory by the Viertakter . . 1 4 5. Metrical Editing in de Boor's Nibelungen Edition . . . . 1 7 II. The Nibelungen Stanza: General Facts 23 1. The Nibelungen Manuscripts 23 2. The Formal Divisions of Manuscript B 24 3. Length of Metrical Units in Terms of Syllables 27 4. Length of Words in Terms of Syllables 33 III. Lexicality and Prosodic Distinctiveness 43 1. The Group One/Group Two Distinction 43 2. The Lexical Feature 46 3. The Lexical Feature and Manuscript B 51 4. Dipody and DHL 4 56 5. The Hiatus Rule 59 IV. The Stress Feature 67 1. Linguistic Stress and Lexicality 67 2. Syllabic Variants and Postvocalic Liquids 71 V. The Beschwerte Hebung 77 1. Heusler's Theory and the de Boor Edition 77 2. A New Theory of the beschwerte Hebung 82 3. Summary and Concluding Statement 85 Appendix I 89 1. Questionable Consonant Clusters in Manuscript B . . . . 8 9 2. Metrical Editing in the de Boor Edition 92 Appendix II 95 1. Distribution of Line Length in Terms of Syllables . . . . 95 2. Distribution of Line Length for Lines 1-3 and Line 4 . . . 95 3. Distribution of Halflines according to Length 96 4. Distribution of Words in Terms of Length across the Halfline 96 5. Distribution of Monosyllables in the Sample Stanzas . . . 97 6. Distribution of Disyllables in the Sample Stanzas . . . . 97

vi Appendix III 1. Distribution of Group I Words across the Halftone . . . . 2. Distribution of [+LEX] across the Halfline 3. Distribution of [+LEX] for DHL's 4 with [+LEX] in slot 5 and DHL's 4 with [-LEX] in slot 5 4. Distribution of [+LEX] after Realignment according to the Hiatus Rule 4.1 Distribution of [+LEX] for AHL's 1 4 and DHL's 1-3 4.2 Distribution of [+LEX] for DHL's 4 with [+LEXJ in slot 5 and DHL's 4 with [-LEX] in slot 5 5. Halflines affected by the Hiatus Rule Appendix IV 1. Distribution of [+ST] across the Halfline 2. Monosyllabic G2 Verbs in slot 3 before [+ST] in slot 2 . 3. [+ST] Monosyllables in slot 3 3.1 Nouns of Address her and vrou 3.2 Monosyllables with Disyllabic Variants 3.3 Monosyllables with Postvocalic Liquids 4. Iambic Words with [+ST] in slot 3 4.1 With Trisyllabic Variants 4.2 With Postvocalic Liquids Appendix V 1. Non-Alternating DHL's 4 with [-ST] in slot 5 Bibliography Index

99 99 99 100 100 . 100 101 101 103 103 .103 105 105 106 107 109 109 109 Ill Ill 113 115

PREFACE

In the past few decades, the debate on prosody has assumed the character of a dialogue between two opposing schools. On the one hand, there are those who base their theory on the most probable reading of a line of verse by a speaker of the language and scan the verse accordingly. This approach, by allowing the stress contours of the language to prevail and adjusting the pattern of the meter to accommodate them, is incompatible with a conception of poetic "rhythm" as the tension resulting from a situation in which the meter and the language are at odds. For the proponents of such theories of prosody, the performance of the text and its meter are one and the same thing; that is, if meter and language are equated, then no possibility exists for a discussion of rhythm based on the notion of conflict between the two. For English prosody, a leading proponent of what is commonly called performance-oriented prosody is Northrop Frye (as outlined in his Anatomy of Criticism),! whereas the performanceoriented theories of Andreas Heusler (as presented in the three volumes of his Versgeschichte)2 dominate the discussion of prosody for German verse. Both Frye and Heusler rely heavily on a system of timing and musical terminology to express their performance guidelines, and, as is shown in (1) below, their scansion procedures can be used to establish at least three different readings for a single halfline from the Nibelungenlied, each of which has a legitimate claim to validity in the context of a performance theory: (Frye has never addressed himself to the prosody of the Nibelungenlied. The scansions opposite his name are included only to demonstrate his system of musical notation and its relation to the time values of Heusler's notational system.) (1)

da wir da schouwen miiezen 1511,3 3

a)

mue- zen da wir da schouwen * / / s x x x X x X A ^ f c - J—J- J J

Heusler: Frye: b) Heusler: Frye:

U

zen da wir da schouwen miier r /

A X 2»

4*

J

u

u

«T3

X

X

X

J

J

•J 'C T.

A

2 da wir da schouwen miie- zen u u X X X A

c) Heusler: Frye:

2

J~~3

4 "

J

J J

J

J ^t

The non-performance prosodist would object to the notion that a single line of verse could be scanned in several ways, 4 as in (la-c) above, and would point out as a particular weakness of such an approach the fact that, once Heusler and Frye have completed their scansion, they have little else to say about the verse. Opponents of Heusler (or Frye) will insist that the prosodist make a fundamental and essential distinction between the abstract metrical pattern, on the one hand, and the linguistic characteristics of the verbal material, on the other. Their opposition to performance-oriented theories stems from a belief in the necessity of this distinction, and, as W. K. Wimsatt and Monroe C. Beardsley emphasize, 5 the non-performance prosodist is convinced that the meter is an enduring quality of the verse which can not be changed f r o m performance to performance or from reader to reader. Since the non-performance prosodists consider the meter to be an abstract pattern to which the verbal material can only partly conform, they view rhythm, as elucidated by Victor Zhirmunskij below, as the product of a conflict between the absolute demands of the metrical pattern and the imperfect realization of these demands by the language material: There is thus no pure rhythm in poetry just as there is no pure symmetry in painting. Rhythm exists as the interaction of two things: the natural characteristics of the verbal material, and the compositional law of alternation, which is incompletely realized, owing to the resistance of the verbal material.6

My investigation, as the title indicates, will deal with the prosody of the Nibelungenlied. Specifically, for reasons which will be discussed at a later point, the study will concentrate on Manuscript B of the text, 7 for which a computer-generated concordance has been prepared t o aid in the scrutiny of the data. The investigation will adopt a stance opposite to that of Heusler. The goals will be descriptive, drawing, in a general way, on the theoretical assumptions and methodology of the Russian Formalists, that is, their view of meter as an abstract pattern and their sense that the task of the prosodist lies in a description of the relation of the language material to that abstract pattern. In typical Formalist fashion, 8 I shall align the Nibelungen data against a grid and attempt to analyze them in terms of the distribution of various linguistic phenomena or categories across this positional frame of reference. My special concern will be the cadence configurations of the Nibelungen halfline. In this respect my study, in its methodological orientation, is consistent with that of all previous researchers. The results will be largely of a statistical nature, and the presentation will proceed f r o m cruder to more refined levels

3 of distinction, that is, from lower to higher orders of approximation to the verse structure. I wish to express my gratitude to the Office of Research and Advanced Studies at Indiana University for its financial support and, especially, to Ron Tschudi for his assistance with the computational work involved. I am indebted to Professor Frank Banta, Professor William Shetter and Associate Professor James Poag for their careful reading and useful comments. In particular, I am grateful to Associate Professor Karl Magnuson. Without his perspicacity and encouragement, this study would not have been possible.

NOTES 1 Northrop Frye, Anatomy of Criticism: Four Essays (Princeton, 1957), pp. 251-55. 2 Andreas Heusler, Deutsche Versgeschichte (Berlin, 1925-29). Subsequent references to this work will be included in the text with volume and page numbers in parentheses. 3 Helmut de Boor, ed. , Das Nibelungenlied, 20. Auflage (Wiesbaden, 1972). Subsequent references to stanza and line numbers will be based on this edition. 4 This statement is not intended as a rejection of the concept of metrical ambiguity, but rather as a rejection of the performance-oriented prosodist's notion that almost all lines of verse are, in some detail of the performance, metrically ambigous. My understanding of metrical ambiguity can be demonstrated by the construct below in which the verbal material readily accepts the two different scansions given, that is, one can scan the line either as iambic pentameter or as dactylic tetrameter: * since hfs own sister he would not defend * sfnce his own sister he would not defend Lines of this type (i. e. , where the verbal material of the line, in isolation, supports a scansion according to more than one abstract metrical pattern) are, in a fundamental sense, metrically ambigous, but this concept of metrical ambiguity bears little resemblance to the potential ambiguity which the performance-oriented prosodist sees in almost every line of verse. 5 W. K. Wimsatt, Jr. and Monroe C. Beardsley, "The Concept of Meter: An Exercise in Abstraction," PMLA, 74 (1959), 587-88. 6 Victor Zhirmunskij, Introduction to Metrics, ed. E. Stankiewicz and W. N. Vickery (The Hague, 1966), p. 21. 7 Das Nibelungenlied und die Klage: Handschrift B, ed. K. Bischoff and H. M. Heinrichs and W. Schroder, Deutsche Texte in Handschriften, 1 (Koln, 1962). 8 Zhirmunskij refers to several studies which use this methodology. See in particular p. 48ff. and p. 74ff.

I. THEORETICAL BACKGROUND

For most of the past half century the theories of Andreas Heusler have dominated research in the field of MHG prosody. New handbooks present the high points of his theories to students; new editions put his theories into practice. With reference generally to performance-based prosody, Wimsatt and Beardsley have found the primary difficulty to be the strong element of subjectivity. 1 In the case of Heusler's work, however, the problem of this subjectivity is greatly intensified by the fact that the language to which he addressed himself is no longer spoken. Moreover, there is the further difficulty that performance-oriented theories of prosody tend to obliterate the distinction between meter and linguistic actualization. The contours of the language and any inherent prosodic ambiguities serve as the basis for the scansion and yield, predictably, a variety of patterns for any single segment of verse. For the performance-oriented prosodist the meter of a verse text is in a state of continuous flux. A single line invites as many scansions as there are possible performances of the line and, if carried out rigorously, a performance-based determination of the meter can result in a view in which no two lines of the text have precisely the same metrical pattern. The above is not intended to suggest that constants play no role in Heuslerian theory. Central to Heusler's understanding of verse structure is the notion of the recurrence of linguistic stresses in equal time intervals, a concept most commonly (though not by Heusler) referred to as isochronism: "Belauschen wir die Ordnung, das Gleichmass... ein wenig genauer, so hören wir: von Hebung zu Hebung sind gleiche Abstände: wir erwarten die nächste Hebung in einem bestimmten Zeitpunkt" (1,17). Given Heusler's dependence upon the concept of isochronism in his basic definition of meter, it is perhaps not surprising that he turned subsequently to the time values of traditional musical notation to illustrate his theory. He even provided a table with his own signs to represent the time values of musical notes, from the whole note all the way to the sixteenth (I, 33-34). To make certain that the reader takes these time values literally, he drew the analogy between the reading of a poem and the playing of a violin with a metronome (1,46). Heusler was convinced additionally that the four-stress line constituted the natural Germanic verse unit (II, 74), a conviction, as we shall see, which had far-reaching consequences in other parts of his theory of MHG prosody. In this introductory chapter, I shall discuss what I consider to be the major shortcomings of Heusler's theory of MHG prosody and his widely accepted view specifically of the Nibelungen stanza. Heusler's theory has had particularly unfortunate consequences whenever it has been "applied" in the pre-

6 paration of editions of MHG texts. For this reason, I shall devote some space in a later section of this chapter to an investigation of the amount and kind of metrical editing present in the edition of the Nibelungenlied prepared by Helmut de Boor.

1. Pre-Heuslerian Theories of MHG Prosody In order to place Heusler's views in their proper perspective, it is necessary to discuss briefly the work of two important scholars before him, Karl Lachmann and Karl Bartsch. Lachmann was the first to propose an integrated theory of MHG verse,2 and most scholars of the 19th century adhered strictly to his precepts. In his theory Lachmann subscribed, with certain exceptions which we shall discuss later, to a principle of strict alternation, that is, the idea that metrical stress and unstress (in Lachmann's terms Hebung and Senkung respectively) must alternate and that their linguistic actualization must be monosyllabic. (In this study, positions of metrical prominence will be called metrical stress and metrically weak positions of the pattern will be called metrical unstress.) As an exception to this principle, Lachmann provided for two consecutive positions of stress (with no intervening actualization of unstress) under certain well-defined conditions. He described the collision of metrically stressed syllables simply as the omission of a Senkung between two Hebungen (only later was this configuration called beschwerte Hebung) and proposed, in general, that contiguous stresses may occur on any two consecutive syllables so long as the first has primary word stress and is a long syllable. There are several qualifications to this generalization, but by far the most important is the one dealing with disyllables: if the second syllable of a disyllabic word contains , then contiguous stresses can be accepted only on the condition that the following syllable also contain . (Although it may seem strange, the use here of the grapheme is in accord with Lachmann's concept. He did not specify sounds which could be represented by a single phone or phoneme, but rather all sounds represented by the grapheme .) For the Nibelungenlied, this rule would allow a beschwerte Hebung in halfline 2,4: vil verLIESEN DEN lip, but prohibit a similar scansion in halfline 17,3: wie LIEBEMITleide, because in this case the monosyllable mit following potential contiguous stresses does not contain . Lachmann, although he advocated in general the monosyllabicity of stress and unstress, provided one rule which allows two syllables in unstress and a second rule which allows two syllables in stress. For metrical unstress he stipulated that both syllables must contain < e > and have only a single consonant between them (e. g., leiDETE, luTE GEnuoc)\ for stress he required that the first syllable be a short syllable with primary word-accent and that the second contain an < e > (e.g., UBER ir, UBELgenuoc). Lachmann did not consider these rules

7

exceptions to his fundamental principle of monosyllabic stress and unstress, but understood them rather as situations in which two syllables had the metrical value of a single syllable. In practice, Lachmann's rules accept the scansion in (la) below with two syllables in stress and also the scansion in ( l b ) with two syllables in unstress, while rejecting, in spite of the similar language material, the scansion in (lc) which would put two syllables either in stress or unstress: (la) er BADETsich in dem bluote (lb) sus enDETE sich diu hohzft (lc) von wannen ir, edel Sfvrit

100,3 689,4 106,2

Lachmann's rule for disyllabic unstress is violated in (lc) by the presence of a grapheme other than < e > in one of the potential syllables of unstress (i. e . , Ir), and his rule for disyllabic stress is violated in (lc) by the presence of a long syllable initially (i.e., WANneri). The frequency of "unacceptable" halflines such as (lc) above suggests that Lachmann's rules for disyllabic stress and unstress are much too restrictive, and indeed, they forced him into frequent metrical editing. In his volume of notes on the Nibelungenlied,3 for example, Lachmann used these two rules to suggest numerous metrical emendations and on this basis even declared whole stanzas to be corrupt. Lachmann's rules for the beschwerte Hebung cadence and disyllabic stress or unstress were accepted for a long time as absolutely valid. Karl Bartsch was the first to propose a change which later found general acceptance. Although Bartsch followed Lachmann's general guidelines on primary word-stress and long initial syllable for the beschwerte Hebung, he took issue with Lachmann's qualification of this rule for disyllables,4 that is, Lachmann's contention that disyllables with the grapheme < e > in the second syllable can only qualify as beschwerte Hebungen if the following unstressed syllable also contains . Bartsch could not understand why Lachmann focused attention on the first unstressed syllable following the beschwerte Hebung. He contended that the presence or absence of < e > in the unstressed syllable following the beschwerte Hebungcould have little effect on the relationship of the contiguous stressed syllables to one another. For Bartsch it was this relationship of the first stressed syllable to the second stressed syllable in the beschwerte Hebung cadence which was most important, and thus, he departed from Lachmann's theories on this point and formulated a less constraining rule for the beschwerte Hebung: contiguous stresses are possible whenever the linguistic stress of the first syllable is equal to or greater than that of the second syllable. Bartsch demonstrated this rule with a scansion of halfline 17,3 which Lachmann's rule does not accept: wie HEBE MIT leide. (Actually, Lachmann's emphasis on a weak unstressed syllable following a beschwerte Hebung seems to indicate that he intuited a fundamental principle of prosody. Karl Magnuson and Frank G. Ryder, in their recent article "The Study of English Prosody: An alternative Proposal,"

8 discuss the principle of loss and recovery of equilibrium, in which the disturbance of the dominant pattern demands immediate reaffirmation of this pattern. 5 It is not improbable that Lachmann perceived the occurrence of a beschwerte Hebung as a disturbance of the predominant pattern of alternating stress and unstress and that his insistence on an extremely weak unstressed syllable following the beschwerte Hebung stemmed from a desire to see the predominant alternating pattern immediately reestablished.)

2. The Nibelungen "Feud" « Given the fact that the manuscripts of the Nibelungenlied gave no direct information on authorship and that the three major early manuscripts (i. e . , A, B and C) showed much greater differences than were apparent in the manuscripts of the medieval courtly romances, the major areas of investigation in the 19th century tended to center on questions of origin, authorship, the relationship of the manuscripts to one another and the choice of the best manuscript for an edition. Lachmann dominated the early phase of the discussion with his famous Liedertheorie.6 According to this theory, the Nibelungenlied existed originally as a series of short heroic lays and was only later collected into its present form as a single epic. In addition, Lachmann was convinced that only Manuscript A was appropriate for an edition, the shortest of the three major early manuscripts and the one best suited to his Liedertheorie. Shortly after Lachmann's death, one of his followers, Karl August Hahn, put out an edition numbering only the stanzas which Lachmann had attributed to the "original" heroic lays.7 Jakob Grimm, who had never really accepted Lachmann's Liedertheorie, was astounded to discover from Hahn's edition that the number of stanzas in each of Lachmann's "original" lays was almost always divisible by seven. Grimm added this to his other reservations and came out strongly against the Liedertheorie.8 At about the same time, Friedrich Zarncke 9 and Adolf Holtzmann 10i were formulating their arguments against Lachmann's choice of Manuscript A. In 1854, they both came out in favor of Manuscript C. Lachmann's followers, led by Karl Mullenhoff, H attacked the new theories vigorously and defended their deceased master as though their personal honor were at stake. Holtzmann counterattacked viciously,12 and the battle continued unabated for the remainder of the 19th century, complicated in 1865 by Bartsch's support of Manuscript B as the best for an edition. New theories on the origin of the epic were also hotly contested. Bartsch, 13 for example, followed Franz Pfeiffer's theory which made a poet known as von Kiirenberg the author of the Nibelungenlied,14 but less than a decade later this theory was disproved by Karl V o l l m o l l e r . 1 5 It was not until the 20th century with Wilhelm Braune's work on the manuscripts 16 and Heusler's work on origin 1 7 and prosody that a degree of stability entered the discussion

9 of topics related to the Nibelungenlied. Any review of 19th century secondary literature on this epic—whether on origin, manuscripts or prosody—would not be complete without indicating that the exchange of ideas took place in an atmosphere of heated emotions. Tensions were so great in 1877 that Hermann Paul implored his contemporaries to replace their polemics with objectivity or share responsibility for the destruction of their own profession: "Nur unter dieser Voraussetzung gibt es einen Ausweg aus der Rechthaberei, dem Cliquenwesen, den unseligen Zerwürfnissen, die seit Decenien der Krebsschaden der deutschen Philolgie sind." 1 8

Heusler and Lachmann Compared Heusler, still caught up in the polemics of the 19th century scholars, used the introduction to MHG verse in his Versgeschichte as an opportunity to deliver a blistering attack against the theories of Lachmann (II, 103-109). Heusler did not, however, criticize Lachmann's system for a lack of internal consistency. Rather, his objections were based on the fact that his own theory, which prescribed four stresses for each halfline, was not compatible with Lachmann's rather strict scansion rules. Heusler subscribed, instead, to Bartsch's more accomodating rules for the beschwerte Hebung and proceeded to discard, in addition, Lachmann's special rules limiting the conditions under which two syllables are allowed into metrical stress or unstress. Heusler continued by introducing his own special concept of the paused stress {pausierte Hebung), i. e . , the proposition that metrical stress can occur even in the absence of language material (II, 113-115). The notion of the paused stress appears to be the direct result of Heusler's hypothesis that the Viertakter is the basic line for German verse from Otfried all the way to Opitz (II, 74). Thus, while admitting that the first three Abverse of the Nibelungen stanza have only three stresses realized (e. g., klagen herzen nöt), Heusler demanded a final paused stress (pausierte Endhebung) in that part of the stanza to make these halflines conform to his hypothesis of the Viertakter (e. g., A\kldgen\herzen\ndt\AA\\). Heusler allowed, moreover, for verse pauses to be inserted, internally (e. g., von Stade A er schieben [II, 114]) or finally, in any halfline where the language material was not sufficient to support a scansion with the required four stresses. The hypothesis of the four-stress line unit was the constant in Heusler's theory. The scansion devices were the variables, the tools with which it became possible, essentially, to make a Viertakter of every line of German verse over a period of seven centuries. The implications of this procedure are disturbing. Given the necessity of a Viertakter and Heusler's liberal scansion techniques, it is possible to scan as Viertakter lines of verse varying in length from three to twelve syllables. Heusler's theory could accept, for example, the constructs in (2a) and (2b) below as metrically equivalent Viertakter:

10 (2a) * AI Kriem-1 hi'l-1 d£ A I A A || (2b) * Do en-1 spränc von dem | stürmküenen | rösse der | helt || Heusler continued his discussion of MHG verse by criticizing Lachmann's terminology (II, 123-126). He found Lachmann's use of only two terms for final cadences, klingend and stumpf, to be inadequate and proposed, instead, his own expanded set of terms with which one could identify eight different halfline-types as opposed to Lachmann's two. One way Heusler expanded the inventory of terms was to mark the number of syllables in each final cadence (e. g., 2-silbig klingend as distinguished from 3-silbig klingend). The main departure from Lachmann's terminology, however, was an application of the individual terms to a different domain. Lachmann was concerned with the character of the final cadence and used the term klingend for final stress + unstress (e.g., die ich muoz tougenÜCHE) and stumpf for stress + juncture (e. g., wart do bleich unde ROT). Heusler's terms, by contrast, were created to indicate the length of each halfline as compared to the theoretical ideal of the Viertakter, for example: his term voll described a halfline in which all Takte were filled and no pauses were necessary (e. g., A | wärt dö \ bleich \ Cinde I /"dill); his term klingend had secondary stress on a final [a] syllable and required paused unstress in the final Takt (e. g. , die\ ich muoz\ töugen-\ Ii-\ che A||); stumpf was applied to halflines where the entire final Takt was paused (e. g., A | klagen \ herzen\ not\ A A||); and überstumpf was used for halflines where Heusler's system of timing had to account for a final unstressed pause followed by a paused final Takt (e. g., mit\ minem\ schil-\ de A | A A ||). Since Lachmann's prosody did not posit an ideal halfline-length such as the Viertakter, his terminology, as indicated above, did not compare each line of verse to a theoretical ideal, but simply described the final cadence. Both Lachmann and Heusler agreed on a Nibelungen stanza composed of four long lines, and both also considered the basic unit of the Nibelungen verse to be the halfline segment. Thus, in the view of Lachmann and Heusler, each Nibelungen stanza had eight cadences, four caesural and four final. The stanza in (3a) giving Heusler's scansion and terminology and (3b) with Lachmann's scansion and terminology could be considered a "norm", since it contains the statistically predominant cadence-types:

(3a) Heusler A| Näch den| hérge-| sél-| lèn A | | A| wärt ein| böte ge-| santi A A I I A| ób sii wóldenl scóu-| wèn A | | Al nfuwezl fr ge-| wäntl A A | | obi éz den| héldenl wae-| rè A II ze| kürz undl óuch ze| lanci A A II ez| was in| réhterl mä-| zè A I | desi sagtenl si den| fróuwenl däncl | 1.4 Hebungen, 2-silbig klingend 4 Hebungen, 1-silbig stumpf

11 2. 4 Hebungen, 2-silbig klingend 4 Hebungen, 1-silbig stumpf 4. (3b) Lachmann Nach den hergesellen ob si wolden scouwen ob ez den helden waere ez was in rehter mäze

4 Hebungen, 1-silbig voll

wart ein böte gesaht niuwez ir gewant ze kurz und ouch ze lanc des sagten si den frouwen danc

1 . 3 Hebungen, klingend

3 Hebungen, stumpf

4.

4 Hebungen, stumpf

A comparison of (3a) and (3b) reveals that the differences in the scansion procedure and terminology reflect two basically different approaches to prosody. It seems unreasonable of Heusler to complain of a lack of variety in Lachmann's terminology when the "variety" in his own terminology is the result of concepts such as timing and the Viertakter with which Lachmann was unfamiliar. Lachmann, in defense of his terminology, might have responded, for example, that the variety in Heusler's terminology was based on unnecessary and artificial distinctions. In any case, Heusler's criticism of Lachmann obscured the fact that he borrowed a great deal from Lachmann's theories and was in basic agreement with Lachmann on at least three major points with respect to the structure of the Nibelungen stanza: (1) he agreed with Lachmann that the final cadence of the four ascending halflines was most frequently disyllabic (I use the term ascending halfline for the German Anvers); (2) he also agreed that the final cadence of the four descending halflines was most frequently monosyllabic (I use the term descending halfline for the German Abvers); and (3) his concept of the fourth descending halfline following Lachmann's description of this halfline as the longest in terms of the number of syllables. In addition to the most frequent halfline-final cadences, as shown in (3) above, the theories of Heusler and Lachmann provide for variations. One such variation, as given in (4a) and (4b), is the occurence in ascending halflines (abbreviated hereafter as AHL's) of a trisyllabic cadence with a short initial syllable: (4a) Heusler er| hiez ge-| wihnenl Hage-| nen A | | 4 Hebungen, 3-silbig klingend (4b) Lachmann er hiez gewinnen Hagenen 3 Hebungen, klingend

12 The occurrence in descending halftones (henceforth abbreviated DHL's) of a disyllabic final cadence with a short initial syllable, as shown in (5a) and (5b), is another frequent varation of the most frequent halfline-flnal cadences: (5 a) Heusler sprachl SÌ-1 vri't der| dégenl A A II 4 Hebungen, 2-silbig stumpf (5b) Lachmann sprach Sfvrit der degen 3 Hebungen, stumpf In Heusler's terminology, the cadence shown in (5) was given a different label if it occurred in the halfline-final position of a DHL 4. (6a) and (6b) illustrate this difference between Lachmann and Heusler: (6a) Heusler den| kiinecl Günt-| hér ver-| jehenl | 4 Hebungen, 2-silbig voll (6b) Lachmann den künec Gunther verjehen 4 Hebungen, stumpf An infrequent deviation from the "norm" was described by Heusler as an instance where a halfline of the type DHL 4 with its full four stresses was allowed to occur as an AHL. Heusler's description of these deviant AHL's remained somewhat vague: "Zäher dauerte der Kadenzantausch 4 v[oll] für 4 k[lingendj im Anvers, besonders bei Eigennamen (Wólfhart; Hildebränt). Das NL [Nibelungenlied \ hat noch 650mal, =7%, vollen Anvers" (II, 262). Apparently Heusler assumed that words of the type réckèn constitued primary stress + secondary stress (i. e. , klingend) in the caesura, whereas words of the type Wólfhart were to be interpreted as primary stress + primary stress (i. e. , voll) in the caesura. The problem at this point is to ascertain which disyllables in the caesura Heusler wished to handle like recken and which he wished to handle like Wolfhart. Heusler provided no specific morphological or phonological information, but only a mention of proper names, that is, the examples Wólfhart and Hildebränt and a total for voile Anverse of 650 occurrences. Of course, it is possible that Heusler intended the two examples above to contain all the necessary morphological and phonological information, but an examination of Manuscript B, the primary manuscript for the de Boor edition, yields only 427 caesural occurrences of disyllabic and trisyllabic proper nouns like Wolfhart and Hildebränt. If the morphological criteria implied by Heusler's exam-

13 pies are relaxed somewhat and disyllables and trisyllables with a derivational suffix in the final syllable are accepted (e. g . , wdrheit, meisterschaft), then it is possible to locate 550 AHL's which Heusler would label voll. These halflines are distributed in the following way: (a) 70% with disyllables (i. e . , root morpheme + root morpheme or derivational suffix) in the caesura; (b) 25% with trisyllables (i. e. , root morpheme + reduced vowel + root morpheme or derivational suffix) in the caesura; and (c) 5% with other wordtypes such as monosyllables or iambic words (e. g. ,gesach) in the caesura. The additional 100 occurrences in Heusler's total of 650 can not be accounted for. (7) gives examples of AHL's to which Heusler would probably have assigned the label voll: (7a) Heusler Nu| lón iu| got her| Sf-| vritl I 4 Hebungen, 1-silbig voll A I fliihel méisterl Hi'lde-I bräntl | 4 Hebungen, 1-silbig voll (7b) Lachmann Nu Ion iu got her Sfvrit 3 Hebungen, klingend fliihe meister Hildebrant 4 Hebungen, stumpf There are two very infrequent deviations from the " n o r m " in DHL's 1 -2 of the Nibelungen stanza which Heusler labeled überstumpf. One of these cadences, as shown in (8), was disyllabic with a long initial syllable: (8a) Heusler miti mineml schi'l-| dè A I A A I I 4 Hebungen, 2-silbig überstumpf (8b) Lachmann mit minem Schilde 2 Hebungen, klingend The second of the two very infrequent cadences in DHL's 1-2,. as illustrated in (9), was trisyllabic with a short initial syllable: (9a) Heusler A I sprach dò I Häge-| nè A I A A | I 4 Hebungen, 3-silbig überstumpf (9b) Lachmann

14 sprach dö Hagene 2 Hebungen, klingend Heusler summarized his view of the Nibelungen stanza with a composite model of the primary cadences, putting the secondary cadences beside them in parentheses (II, 262): (10) 1.4 2. 4 3.4 4.4

AHL's klingend (voll) klingend (voll) klingend (voll) klingend (voll)

4 4 4 4

DHL's stumpf (überstumpf) stumpf (überstumpf) stumpf voll

Rhyme a a b b

Heusler claimed that Lachmann's two terms klingend and stumpf were not sufficient to describe the verse. He redefined Lachmann's terms, added the terms voll and uberstumpf and identified the number of syllables in each final cadence. Heusler's polemics notwithstanding, Lachamnn's theory allowed one potentially important distinction which Heusler's theory did not allow: Heusler required four-stress halflines, while Lachmann distinguished between two-, three- and four-stress halflines. Heusler further claimed that Lachmann's restrictions on the use of the beschwerte Hebung and on the types of syllables in disyllabic unstress led to an unnecessary inflexibility in scansion. He berated Lachmann's lack of insight in failing to grasp the necessity of the final paused Takt for DHL's 1 -3 of the stanza and the necessity of secondary stress on the final [a] syllable of the cadences klingend (i. e . , tougen-\ If-\ che A | |) and uberstumpf (i. e . , schil-\ de A | A A | I). Heusler's scansions in these instances, however, were necessitated only by his own theory of the Viertakter. There is no linguistic justification either for the paused stress or for secondary stress on the final J a] syllable of the cadences klingend and uberstumpf. Since Lachmann did not base his theories on the necessity of the Viertakter, he saw no reason to use paused stresses and furthermore, acknowledged the lack of linguistic prominence in the final [s] syllable of Heusler's cadences klingend and uberstumpf by putting it into metrical unstress (i. e . , tougenlfche, schilde). Ursula Hennig's view is clearly valid that Lachmann's scansion and terminology are truer to the language material than Heusler's.I 9

The Domination of Heusler's Theory by the Viertakter It is crucial at this point to examine the implications of Heusler's concepts in terms of general metrical theory. Heusler's theory has certain basic organizational features in common with most theories o r prosody, as, for example, the distinction he makes between abstraction and realization. 20 That is, he postu-

15 lates an abstract scheme for the stanza having to do with four long lines, each of which contains two halflines which are characterized in turn by four Takte, the latter being not necessarily linguistically realized. In other words, he deals (as do most prosodists) with the rules of meter on the one hand (i. e. , the rules which give the abstract metrical scheme, the number of positions in the line, lines in the stanza and so forth) and the rules of prosody on the other, that is, those rules which govern the relation of the linguistic material to the abstract pattern. These rules, the rules of prosody, fall into two kinds: 2 i sequential rules which govern the realization of the basic metrical relations (for example, the relation of stress to following unstress or the other way around) and positional rules which govern the realization of the larger metrical units, namely the halfline, line or stanza. Heusler posits three main positional rules, as given in (11), all of which have to do both with the realization of the Viertakter and the stanza: (11a) The final Takt in the AHL is never completely paused, that is, all AHL's are, in Heusler's terms, always klingend or voll and never stumpf or uberstumpf. ( l i b ) The final Takt in DHL's 1-3 is always completely paused (i. e. , always stumpf or uberstumpf). (11c) In DHL 4 the final Takt is always completely filled (i. e. , always voll). Heusler supplements the rules in (11) with a number of positional guidelines which indicate the most frquent realizations of the larger metrical units. (12) gives examples: (12a) AHL's are most often klingend. (12b) DHL's 1-3 are most often stumpf. (12c) Upbeat ( A u f t a k t ) may be missing or contain up to two syllables, but it rarely has more than two syllables. (12d) The second stress of DHL 4 is the verse position where the beschwerte Hebung cadence is most often initiated. The distinction I note above between rules and guidelines closely parallels the distinction Roman Jakobson made in his examination of Slavic epic verse, where rules are called "constants" and guidelines are called "tendencies." 2 2 Rules (or "constants") constitute constraints which are not violated by the verse in question, whereas guidelines (or "tendencies") give statistically favored metrical phenomena. Heusler's examination of sequential relationships did not, however, yield the distinction between rules and guidelines which resulted from his analysis of positional relationships. In an effort to capture every last bit of data without compromising the principle of the

16 Viertakter, Heusler formulated no sequential rules and left the organizational task entirely to very generally stated statistical guidelines. (13) gives examples of some important sequential guidelines in Heusler's theory: (13a) Linguistically strong syllables frequently correspond to metrical stress and linguistically weak syllables to metrical unstress. (13b) The linguistic contours across halflines often indicate an alternation of linguistically weak and strong syllables or, to put it another way, halflines with beschwerte Hebungen or polysyllabic unstress are not predominant in the verse. (13c) Halflines with consecutive beschwerte Hebungen (e. g. , hdhvertige) or trisyllabic unstress are rare in the verse. It is clear from the presentation above that the abstract component and the positional rules (i. e. , the rules which realize the positional pattern) completely dominate Heusler's theory. He demands that the integrity of the Viertakter be upheld, but provides no rules by which the simple metrical relations stress-unstress and unstress-stress may be realized. Nor does Heusler indicate specifically the conditions under which the numerous sequential guidelines may be violated, although it appears that we may discard the sequential guidelines whenever the preservation of the Viertakter demands it. Thus, although we are told that consecutive beschwerte Hebungen or trisyllabic unstress are rare, these scansion possibilities remain, nevertheless, to be used in an emergency, that is, to be used whenever the Viertakter can not be realized without them. The frequent need for such drastic techniques of scansion, however, brings into question the validity of the abstract pattern itself (i. e. , the Viertakter): Indeed, one could most probably operate with a Dreitakter or Fiinftakter as an abstraction, if given the possibility of using consecutive beschwerte Hebungen or trisyllabic unstress or paused stresses whenever such scansions seem necessary. The main problem is that Heusler does not demonstrate that abstraction and realization in his theory fulfill a necessary function with respect to a specific text, but only that they fulfill a necessary function with respect to each other. Without the Viertakter, the permissive scansion techniques would not be necessary; without the permissive scansion techniques, the Viertakter would not always be possible. Priority in a theory of prosody should be given the linguistic givens of the text in question. Heusler, however, reaches outside the text of the Nibelungenlied to grant priority in his theory to an abstract pattern which he apparently considered to be permanently fixed in the collective consciousness of the German people. The result is the virtual absence of rules governing the simplest relationships in this abstract pattern and thus no adequate accounting for the data as they are contained in the various manuscripts.

17 5. Metrical Editing in de Boor's Nibelungen

Edition

The weaknesses, in practice, of Heusler's theory can be demonstrated on the basis of de Boor's edition of the Nibelungenlied, since de Boor follows Heusler's theory.23 A comparison of his edition with its primary manuscript (i. e. , Manuscript B) will yield information on the extent of metrical editing: the presence of any metrical editing at all, however, is puzzling, since the absence of sequential rules in Heusler's theory and the accommodating scansion techniques for "problem" halflines should make it possible to give any manuscript halfline the appearance of a legitimate Viertakter without editing. In other words, one can not establish, given the absence of clearly defined sequential rules in Heusler's theory, that any single metrical emendation in de Boor's edition is necessitated by Heusler's theory. And yet the collective effect of de Boor's editing establishes his reliance on Heusler, since Heusler's sequential guidelines, as given in (13) above, would be grossly in error without the emendations in de Boor's edition. These guidelines posit, for example, that such drastic scansions as consecutive beschwerte Hebungen or trisyllabic unstress are quite rare in the verse, since, as Heusler must have been aware, the cumulative impact of a frequent use of such drastic scansion techniques would have undermined confidence in the validity of the Viertakter and the accompanying positional rules. Thus, as we shall see subsequently, while no single instance of metrical editing by de Boor can be said to result from the violation of a rule in Heusler's theory, all his emendations "improve" halflines according to Heusler's sequential guidelines and thus reenforce, by reducing greatly the frequency of drastic scansions, our general feeling of confidence in the validity of the Viertakter and the positional rules through which it is realized. This does not mean, however, that de Boor has interpreted Heusler's guidelines as inviolable rules and thus has avoided entirely the use of Heusler's more permissive scansion techniques. There are, for example, numerous instances of scansion suggestions in de Boor's edition which make use of such scansion techniques as consecutive beschwerte Hebungen and trisyllabic unstress. Thus, de Boor's dependence on Heusler's theory leads to a confusing inconsistency in his edition: he edits according to Heusler's sequential guidelines in cases where the use of Heusler's permissive scansion devices could preserve the halfline as it stands in the manuscript, and yet he preserves the manuscript version by using the extreme scansion devices in other halflines which could also, for the sake of consistency, be edited according to Heusler's sequential guidelines. Before discussing de Boor's editing in greater detail, it is necessary to consider a certain aspect of scribal convention in Manuscript B where linguistic evidence indicates that de Boor's editing, though it has metrical implications, is the result of linguistic rather than metrical emendation.

18 Namely, scribal convention in Manuscript B allows for the occurrence of words with extremely unusual consonant clusters, apparently because of the omission < e > in the final syllable ( e . g . , reCKN). The appendix contains a complete list of these clusters and the words in which they occur. 2 4 There is linguistic evidence which indicates that these clusters should not be considered a single syllable in blind adherence to the manuscript orthography. Although many linguistic definitions of the syllable demand information which is not retrievable for a dead language, the theory of crest of sonority, as outlined in Heffner's General Phonetics,25 does seem appropriate. The consonants and vowels are ordered according to increasing sonority, i. e. , from voiceless stops and spirants to voiced stops and spirants to resonants (including [s] and finally vowels. The number of syllables is determined by the number of crests of sonority, with or without vowel nucleus. On the basis of this theory, the form reckn has a second crest of sonority on [n] and thus has two syllables. Indeed, all of the questionable clusters listed in the appendix would contain two syllables by this linguistic definition. This definition of the syllable is not, however, intended as the final theoretical answer to the actual syllabic value of the clusters in question, but is included at this point as a justification of my preliminary decision on the handling of the orthography in Manuscript B, that is, the clusters listed in the appendix will be accepted as having disyllabic value and de Boor will not be faulted for metrical editing in adding < e > to them, since there is strong evidence in this instance that his editing is based on linguistic rather than metrical criteria. To facilitate a comparison of the edition with the manuscript, I have selected three 100-stanza sections: (1) from the beginning—stanzas 200-299; (2) from the middle—stanzas 1200-1299; and (3) from the end—stanzas 2200-2299. 146 of the halflines in these selected stanzas (i. e . , 200-299, 1200-1299, 2200-2299) have been edited by the addition of one or more syllables not present in the manuscript, as shown in (14): (14) de Boor: vil wol ervant er ez sint 209,4 Manuscript B: vil wol ervant erz sint Since the halfline in (14) is a DHL 4, Heusler requires four filled Takte, but this requirement could have been met through a scansion with consecutive beschwerte Hebungen, as shown in (15), rather than through emendation: (15) vil |wol er- |vfl«f|erz|sint|| De Boor's decision to edit here is particularly confusing, since some scansion suggestions in his edition, as illustrated in (16), establish that he does not reject scansions with consecutive beschwerte Hebungen: (16) einen|bern|groz|unde|starc|| 957,4

19 56 halftones in the three 100-stanza sections contain emendations, as in the example in (17), where de Boor deletes one or more syllables present in the manuscript: (17) de Boor: manec frouwe und manec meit 264,2 Manuscript B: vil manec frouwe un manech meit The halfline in (17) is from DHL's 1-3 of the stanza and thus Heusler requires either stumpf or iiberstumpf. Yet here again the emendation could have been avoided, since Heusler allows in difficult halftones for trisyllabic upbeat, as illustrated in the scansion in (18): (18) vil manec|frouwe un|manech|meit| A A | | In this case, confusion is created by the fact that de Boor's own scansion suggestions show that, as in (19), he does not reject scansions with trisyllabic upbeat: (19) daz imez|bluot|under|fue-|zen A | | 2294,4 9 other halftones from the selected stanzas have been emended through a rearrangment of the language material. (20) gives an example: (20) de Boor: des bin ich, vrouwe, bereit 1257,4 Manuscript B: vrouwe des bin ich bereit Since this halftone is from DHL's 4 of the stanza, Heusler's requirement of four filled Takte is met by the manuscript version as well as by the edited version. Although Heusler's guideline concerning the beschwerte Hebung in DHL's 4, as in (12d) above, is not met by the Manuscript B version of the halftone in (20), de Boor accepts other DHL's 4 with language material almost identical to that in (20) and without emendation to yield a beschwerte Hebung. (21) is an example of such a halftone:

(21) gerne het ich daz bekant 419,4 There is a complete list of the discrepancies between edition and manuscript in the appendix (pp. 92-93), and the total establishes beyond any doubt that there is a great deal of metrical editing in the de Boor edition. Fully 8.8% of the halftones in the selected stanzas have been metrically edited. This approaches a rate of one "improved" halftone per stanza; it is simply not acceptable. Yet Heusler himself speaks out strongly against the metrical editing necessitated by Lachmann's theories (I, 7), and de Boor also takes a firm stand against editing of this kind in the introduction to his edition.26 Indeed, Heusler's theory with its absence of sequential rules could accept the manuscript without metrical editing, but only if one applies his permissive scansion techniques more frequently than is indicated as acceptable by

20 Heusler's own vague guidelines. Thus, Heusler's theory places an editor in the impossible situation of either refraining entirely from metrical emendation or attempting to edit apparently corrupt halflines without the help of precisely stated sequential rules. In the case of Manuscript B, Heusler's general guidelines would indicate that metrical editing is necessary to prevent a frquent use of drastic scansion techniques, but his theory never specifies for de Boor or other editors the kinds of linguistic strings which are to be considered violations of his theory. The result in de Boor's edition has been, as was shown above, a striking arbitrariness in editing: there is no indication as to why he has edited certain halflines and has not edited others with almost identical language material. The weaknesses in Heusler's theory as well as the unfortunate consequences of these weaknesses in practice (i. e. , de Boor's Nibelungen edition) point out clearly the need to provide a fresh approach to the problems of Nibelungen prosody specifically and MHG prosody in general. I shall, in what follows, avoid Heusler's preliminary assumptions concerning the absolute validity of the Viertakter and the positional rules and shall accept as valid, after selecting one of the three major Nibelungen manuscripts as the basis for my claims, only those structural principles which are supported by the linguistic facts of the manuscript in question. In other words, my investigation will have the descriptive orientation which was characteristic, in particular, of much of the work of the Russian Formalists. Initially, I shall deal with gross distributional "tendencies" such as line length, halfline length and word length (in terms of syllables), turning subsequently to more refined distributional properties such as the distribution of linguistic data with potential metrical significance (e. g. , word-stress and lexicality) across the halfline. Finally, an attempt will be made to identify certain "constants" with respect to cadence structure, an aspect of Nibelungen and MHG prosody which has been particularly problematic.

NOTES 1 Wimsatt and Beardsley, 585-98. 2 Friedrich Zarncke, ed. , Das Nibelungenlied, 2nd ed. (Leipzig, 1866), xli-lxxvi (Introduction). In referring to Zarncke's concise presentation of Lachmann's theories, I follow the example of Ursula Hennig, Untersuchungen zur frühmittelhochdeutschen Metrik am Beispiel der Wiener Genesis (Tübingen, 1968), p. 47. Lachmann himself never presented his views on MHG prosody in a single essay or book, but rather left them dispersed in sections of lectures, book reviews, letters and Anmerkungen. 3 Karl Lachmann, Anmerkungen zu den Nibelungen und zur Klage (Berlin, 1836). 4 Karl Bartsch, Untersuchungen über das Nibelungenlied (Wien, 1865), p. 138. 5 Karl Magnuson and Frank G. Ryder, "The Study of English Prosody: An Alternative Proposal," CE, 21 (1970), 815. 6 Karl Lachmann, "Über die ursprüngliche Gestalt des Gedichts von der Nibelungen Not" (Berlin, 1816), rpt. in Kleinere Schriften, ed. Karl Müllenhoff (Berlin, 1876).

21 7 Karl August Hahn, Die echten Lieder von den Nibelungen nach Lachmann's Kritik (Prag, 1851). 8 Jakob Grimm, "Die echten Lieder von den Nibelungen nach Lachmann's Kritik von K. A. Hahn (1851)," in Kleinere Schriften (Berlin, 1871), V, 476-79. 9 Friedrich Zarncke, Zur Nibelungenfrage (Leipzig, 1854). 10 Adolf Holtzmann, Untersuchungen über das Nibelungenlied (Stuttgart, 1854). 11 "Untersuchungen über das Nibelungenlied von Adolf Holtzmann," Die allgemeine Monatsschrift (December, 1854). 12 Adolf Holtzmann, Kampf um der Nibelungen Hort gegen Lachmanns Nachtreter (Stuttgart, 1855). 13 Bartsch, p. 369. 14 Franz Pfeiffer, "Der Dichter des Nibelungenliedes: Vortrag (1862)," rpt. in Freie Forschung (Wien, 1867), pp. 3-52. 15 Karl Vollmöller, Kürenberg und Nibelungen (Stuttgart, 1874). Wilhelm Braune, "Die Handschriftenverhältnisse des Nibelungenliedes," BGDSL, 25 (1900), 1-222. 17 Andreas Heusler, Nibelungensage und Nibelungenlied (Dortmund, 1921). 18 Hermann Paul, "Nibelungenfrage und philologische Methode," BGDSL, 5 (1877-78), 447. 19 Hennig, pp. 319-20. 20 Among those prosodists making the basic distinction between abstraction and realization in their theories are such important figures as Lachmann, Zhirmunskij and the Russian Formalists, and Wimsatt and Beardsley and their followers. 21 The rules I refer to as sequential rules are called Base Rules by Magnuson-Ryder, and although they do not formulate positional rules in their article, they point to the presence of such rules as a necessary component in a complete theory of prosody. Magnuson and Ryder, 801-802 (Footnote 18). 22 Roman Jakobson, "Slavic Epic Verse: Studies in Comparative Metrics," in Roman Jakobson Selected Writings (The Hague, 1966), IV, 418-20. 23 De Boor, pp. LII-LVI (Introduction). 24 Appendix I, pp. 89-92. Hereafter, references to the appendix will be included in the text with page numbers in parentheses. 25 R-M. S. Heffner, General Phonetics (Madison, 1950), pp. 74-75. 26 De Boor, p. LIV (Introduction).

II. THE NIBELUNGEN STANZA: GENERAL FACTS

1. The Nibelungen Manuscripts Since most works of the classical MHG period are represented by more than a single manuscript, it is common editing practice for normalized editions to select the manuscript "best suited" for an edition, that is, a primary manuscript, and to refer to secondary manuscripts whenever the primary manuscript presents material deemed to be "corrupt". A change in the primary manuscript can be the result of the editor's judgment of philological and/or metrical corruption, and in most MHG editions, the criteria for metrical jugdments are synonomous with the rules of either Lachmann's or Heusler's theories. Since metrical editing is a part of common editing practice, editions will contain the particular bias of their editors. Given this situation, one would assume that any thorough investigation of metrical problems would by-pass the editions and deal directly with the manuscripts. And yet, there are conspicuous examples among the studies of MHG prosody where this elementary procedural principle has not been observed (most notably in the dissertation of Leopold Rettinger, "Der Auftakt im Nibelungenof Franz Xaver Meyer, "Die beschwerte Hebung im Nibelungenliedthat lied"2 and Helmut de Boor's article, "Zur Rhythmik des Strophenschlusses im Nibelungenlied"3). In these instances the authors are, of course, merely confirming the metrical bias of the editor upon whose edition the study is based— in de Boor's case, his own. In the present study I shall select a single manuscript as the basis for my claims. With nine complete manuscripts and 23 fragments, the Nibelungenlied belongs to a small group of MHG literary works with more than a score of surviving manuscripts. The earliest complete manuscripts (known as Manuscripts A, B and C) originate in the 13th century, and all editions of the epic have been based on them. Although the number of surviving manuscripts indicates an early popularity, the work appears to have been virtually forgotten by the end of the 15th century. 4 Finally, in 1755 the existence of Manuscript C at the castle library in Hohenems was brought to the attention of Bodmer, who published the final portion of it in 1757 under the title, Chriemhilden Rache. By 1780 the existence of Manuscripts A and B was known, and in 1782 Christoph Heinrich Myller prepared the first complete edition, the first half based on Manuscript A and the second half on Manus-

24 cript C. The first edition based on a single manuscript was prepared by Friedrich von der Hagen in 1816. However, his choice of Manuscript B made only a fleeting impression, since Karl Lachmann's preference for Manuscript A soon enjoyed the support of most scholars. Soon after Lachmann's death in 1851, his preference for Manuscript A was challenged by Friedrich Zarncke and Adolf Holtzmann. They both favored the C manuscript. Karl Bartsch brought the debate on manuscripts and editions full circle in 1866 with a return to von der Hagen's choice of Manuscript B. The disagreement on manuscripts for editions continued as a full scale battle throughout the remainder of the 19th century. Then, in 1900, Wilhelm Braune undertook a thorough study of all the manuscripts and concluded that Bartsch's choice of Manuscript B was correct. 5 From that point on, the B manuscript has gained steadily in favor, until today a reading of the Nibelungenlied has become virtually synonomous with a reading of the Bartsch-de Boor edition of Manuscript B. For this reason, the B manuscript is selected as the basis for this investigation.

2. The Formal Divisions of Manuscript B Differences in the calligraphy of Manuscript B show it to be the work of three different scribes. In terms of the numbering in the de Boor edition (stanzas and lines are not numbered in Manuscript B), scribe I copies from stanza 2 to the middle of stanza 21 6 (stanzas 1 and 3 of the edition are not present in the manuscript); scribe II continues through stanza 392 (p. 19/ 307); and scribe III continues from 393 to the end of the epic (stanza 524 of the edition is not present in the manuscript). There is not sufficient material to characterize scribe I adequately, but scribes II and III offer several minor points of contrast such as the following: (1) _Scrihe_IL

Scribe,HL

sivrit sifrit frouwe frowe vriwent friunt degn, lebn, sehn, degen, leben, sehen, pflegn, etc. in rhyme pflegen, etc. in rhyme Generally speaking, scribe II is more meticulous in his work than scribe III. In stanzas 533-78, for example, scribe III becomes confused and marks stanza, onset for line 3 of each stanza (pp. 26/314-28/316). It is also in his section that one finds several lines with obvious omissions or duplications: (2) do gie der snelle recke da er crimhilde [sach? ] si enpfie in gutliche wi bald er do sprach, (p. 61/349)

25 Guntheres ingesinde daz wart gesundert dan. daz riet diu [ ? ] im vil hazes truch. da von man sit di knehte in der herberge sluch. (p. 89/377) von rittern unt von von vrowen wi man di horte chlagen. (p. 50/338) Although the differences noted above suffice to make a clear distinction between scribes II and III, there is no manuscript evidence to indicate that they are operating with a different concept of the Nibelungen verse. Manuscript B contains 38 major divisions of differing length which editions label Aventiuren. The manuscript marks these divisions neither with a title nor with the label Aventiure, but rather with an illumination of the initial letter. Although space has been provided for it, the illumination for the 18th major division is missing (p. 52/340). The de Boor edition indicates 39 Aventiuren, but the manuscript has no indication of a major break between Aventiuren 26 and 27 as marked in his edition (p. 84/372). Each manuscript page has two columns, and within each column from 8 to 12 stanzas are marked either by an initial majuscule or by an initial letter in the margin or both. (3) presents a transcription of two stanzas with the markings and margins of the manuscript: (I have italicized rhyme syllables.) (3)

Do sprach hagen bruder der chfine danchwwi. mich riwet innechlichen disiu hove vart. nu hiezen wir ie rech ken wi Verliese wir den lip. suln uns in disen landen nu uberwinden diu wip. Mich mfit daz harte sere daz ich chom in diz lant. unt haete min bruder hagene sin waffen an der hant. unt ouch ich daz mine so mohten sampfte gan. mit ir fiber mfite alle prunnhilde man. (p. 22/310)

Syntactic juncture offers additional supporting evidence for the stanzas as marked in the manuscript. Of the first 500 stanzas in the manuscript, 482 are marked by a major syntactic juncture. (Except for the rhyme dot, the manuscript does not contain punctuation. As used here, the term major syntactic juncture refers only to those points in the manuscript where modern orthographic convention places a period, exclamation point or question mark.) The line unit in the manuscript is clearly marked. As is evident in (3) above, each stanza has four rhymes according to the scheme aabb, and each rhyme

26 is marked with a rhyme dot. Syntactic juncture reenforces (i. e. , tends to coincide with) line juncture. In sample stanzas from the beginning, middle and end of the text (stanzas 500-50, 1000-49 and 150049), for example, 381 of 410 full syntactic stops are marked additionally by rhyme dots. (The sample stanzas 500-50, 1000-49 and 1500-49 have been selected as a permanent sample, and thus, all subsequent data based on sampling of the text will be collected from these stanzas.) In numerous instances, as in (4) below, all four rhyme dots of the stanza coincide with full syntactic stops: (The double bar marks major syntactic juncture.) (4)

Man lute da cem mfinster nach gewone heit.|| vrou Criemhilt diu schSne wachte manige meit.11 ein lieht bat si ir bringen und ouch ir %