Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The manuscript relationships of Rudolf von Ems’s Barlaam und Josaphat with special reference to rubrication Zaenker, Karl August 1974-01-29

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata


831-UBC_1974_A1 Z33_9.pdf [ 8.58MB ]
JSON: 831-1.0099942.json
JSON-LD: 831-1.0099942-ld.json
RDF/XML (Pretty): 831-1.0099942-rdf.xml
RDF/JSON: 831-1.0099942-rdf.json
Turtle: 831-1.0099942-turtle.txt
N-Triples: 831-1.0099942-rdf-ntriples.txt
Original Record: 831-1.0099942-source.json
Full Text

Full Text

THE MANUSCRIPT RELATIONSHIPS OF RUDOLF VON EMS1S BARLAAM UND JOSAPHAT WITH SPECIAL REFERENCE TO RUBRICATION, by Karl August Zaenker A thesis submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the Department of German We accept this thesis as conforming to the required standard The University of British Columbia March 19 74 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the Head of my Department or by his representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of The University of British Columbia Vancouver 8, Canada Abstract The medieval legend of the two Saints Barlaam and Josaphat has attracted wide attention since, in the nineteenth cen tury, some of its roots were discovered in ancient Indie Buddha legends and parables. Our study limits itself to the German version by Rudolf von Ems, a free translation from a Latin source, written around 1225. Although this work was edited as early as 1818, it was not until the last decade that some detailed but rather opposite inter pretations were devoted to it. All recent studies of Bar laam und Josaphat have been based on a reprint of Pfeiffer edition of 1843 which, however, has grave shortcomings: it takes only a few manuscripts and fragments into account selects their readings at random, and does not provide a reliable critical apparatus. Therefore, it seemed appro priate to work towards a new, truly critical text edition which would be beneficial to further investigations into meaning and structure of the work. Our first step was to locate all existing manuscripts and fragments and obtain photocopies of them. For the ensuing process of assessing their relationships in order to determine their individual value for a text reconsti-tution, we tried an approach different from traditional practice. Instead of basing a grouping on common readings or mistakes alone, we began by comparing the paragraph -ii-markings (rubricated initials) in the major manuscripts. We believe that they were placed originally to subdivide the narrative and that, generally, they were copied by later scribes and rubricators. During the transmissibn process involuntary or deliberate "misplacements" occurred which would show up in further copies and could thus indi cate group characteristics. Therefore we listed all rubri-cation. marks in a comprehensive chart and calculated the overall agreement between the major manuscripts in percen tage figures. The evaluation showed that the oldest two manuscripts C and D, as well as A and b have a very similar rubrica-tion pattern, probably still close to the original one. This likeness makes it difficult if not impossible to deter mine if another manuscript is related to either of these. On the other hand, a clear affinity could be established between the hitherto overlooked manuscripts G, M, and, to a lesser extent, E, as well as between W and L in the first half of the text and C and L in the second half. We had divided the total number of initials into four even sec tions (covering ca. 4000 verses) to see if the "agreement figure" of one manuscript to another changes markedly. A subsequent look at individual "spurious initials" (mostly faults in rubrication) confirmed the first results and established a third definite grouping, that of DKCKa (Ka is only represented by the text edition of 1818). A comparison of the smaller fragments followed in which the main criterion was their textual agreement with other manuscripts. The rubrication was also taken into account but not overemphasized since conclusive evidence was often lacking due to the shortness of most fragments. We found that in three cases fragments belonged together to one otherwise lost manuscript (dq, mF2, and el). Many of the fragments showed characteristics of the GEM-group, whereas only very few seemed related to either A, b, C, or D. This sampling of common variants and text omissions throughout the work served also to verify the results gained by the "initial method." In general, the above mentioned groupings were confirmed or slightly modified. It became clear that in some cases a straightforward text transmission (as represented in a stemma) cannot be assumed. Especially in the loosely related body of manu scripts A, b, (B), C, L, and W, there is strong evidence of contamination which would make a tentative classifica tion futile. A critical edition should, in our view, follow the old reliable Freiburg codex D as lead manu script and confront its text consistently with the read ings and paragraphs of the other two large groupings, mainly C and G. The closing chapter outlines the spreading of Rudolf's -iv-Barlaam und Josaphat as documented in its manuscript tra dition, from its limited Alpine origin to its popularity within the Teutonic Order of Knights in East Prussia until its last flowering in Southern Germany at the end of the fifteenth century. V Content s page 1. Survey of Barlaam research to the present day 1.1. Understanding of the text 1 1.2. The text tradition 6 2. The Barlaam manuscripts 18 2.1. Major manuscripts 9 2.2. Smaller fragments 2 2 3. Study of rubrication c 3.1. Textual criticism and comparison of initials 2 7 3.2. Chart of initials 3 8 3.3. Evaluation of the initial systems 3.3.1. Frequency in the placing of initials 91 3.3". 2. Manuscripts A, b, C, D 94 3.3.3. Manuscripts L and W 101 3.3.4. Manuscripts G, M, and E 104 3.3.5. The K-manuscripts (Ka, Kb, KC) 107 3.4. Spurious initials; textual variants at the beginning of a paragraph 110 4. The smaller fragments 4.1. Introductory remarks 117 vi page 4.2. Matching fragments 4.2.1. Fragments d and q 119 4.2.2. Fragments e and 1 2 5 4.2.3. Fragments m and F2 130 4.3. Single fragments 4.3.1. Fragment h 139 4.3.2. Fragment i 14 3 4.3.3. Fragment k 146 4.3.4. Fragment n 8 4.3.5. Fragment p 151 4.3.6. Fragment r 3 4.3.7. Fl "Basel fragment" 156 4.3.8. F3 157 4.3.9. F6 "Breslau fragment" 159 4.3.10. F7 "Freiburg fragment" 162 4.3.11. F8 "Fulda fragment" and "Prague fragment" 165 4.3.12. F9 The "Gfittingen fragment" 169 4.3.13. F10. "Hannover fragment" 171 4.3.14. Fll "Oettingen fragment" 17 3 4.3.15. F13 "London fragment" 175 4.3.16. F18 "Berlin fragment" 177 5. Conclusion VX1 5.1. Final grouping of manuscripts based on text and rubrication 5.2. The geographical distribution of Barlaam  und Josaphat List of Works Cited Map: Geographical distribution of Rudolf von Ems' Barlaam und Josaphat 1 1. Survey of Barlaam research to the present day 1.1. Understanding of the text The story of Barlaam and Josaphat, though seemingly foreign to the literary taste of today, is one of the most wide spread themes of world literature. Hiram Peri, in his com prehensive bibliography of the legend, including the few pre-Christian forerunners, cites versions in almost forty different vernaculars.^ The Buddhist core of the subject matter and its complex and disputed tradition will not be considered in this study; the work of Peri gives a good 2 introduction into this field. We shall limit ourselves to the Middle High German version, created by Rudolf von Ems around the year 1225 after a Latin model. This so-called "Vulgata" version from which all the medieval Barlaam texts descend is itself one of two early translations of the Greek Barlaam, composed probably in the eighth century by John of Damascus according to diverse older sources. Although both Barlaam and Josaphat are registered in the authoritative Saints' calendar of the Roman Catholic Church, the Martyrologium Romanum, the text itself does not represent a typical Saint's vita. This is indicated already ^ Hiram Peri, Der Religionsdisput der Barlaam-Legende (Universidad de Salamanca, 1959), pp. 223-272. 2 See also the study by Charlotte Nagler, "Studien zu Barlaam und Josaphat von Rudolf von Ems," Diss. Karlsruhe 19 72, which takes xnto consideration the Stoffgeschichte• 2 by the rather unusual naming of two saints in the title: here the emphasis is not on the self-denial, martyrdom and miracle works of one heaven-inspired man, but rather on the instruction that the neophyte Josaphat receives from his God-appointed teacher Barlaam, his observance and dissemin ation of the Christian teachings and, finally, the reunifi cation of teacher and disciple in a common ascetic life. We accept H. Brackert's thesis, that Rudolf's work should not be regarded as a "miracle legend" (Wunderlegende), but rather as the much rarer type of "conversion story" (Bekeh-rungsgeschichte). The overcoming of heathendom and the victory of Christianity are demonstrated in several repeated instances. The conversion of the heathen prophets Nachor and Theodas (Barlaam 11030-11264 and 13179-13310) ,4 the Christianization of his own kingdom through Josaphat and, as the crowning triumph, the conversion of his own father-all of this constitutes the glory (ruom) of the "elect" (der gotes erwelte reine, 15841) and bestows on him, in martyr dom's stead, the rank of sainthood. To this is also added a fundamental feature of most legends, the personal stead fastness of the neophyte, who must maintain himself against all worldly temptations; in the case of Josaphat, they 3 Helmut Brackert, Rudolf von Ems: Dichtung und Ge-schichte (Heidelberg: Universitatsverlag, 1968), p. 214. 4 Quoted after Franz Pfeiffer, ed., Rudolf von Ems:  Barlaam und Josaphat (1843; rpt. Berlin: de Gruyter, 1965). 3 appear in the form of reason, love, and power.b The popu larity of the Barlaam and Josaphat story in medieval times is due largely to its didactic passages in form of parables, the exempla, with which the lengthy instructions and Bible interpretations were illustrated and in which oriental fairy-tale motifs have found their way into the Western world. The Barlaam material has been extensively studied by diverse disciplines during the past hundred years, which prompted J. Sonet to begin the preface of his book on the Latin and French Barlaam versions with the words: "Le roman de Barlaam et Josaphat a deja fait couler des flots d'encre." Surprisingly, on the other hand, the Germanists had treated 7 Rudolf's work "stiefmutterlich," although such an author ity as de Boor termed it "von der Form her . . . das reinste, 8 klassischste Werk Rudolfs." In fact this negligence is 9 inexplicable, the more so as Rudolf's Barlaam was one of 5 see Johannes Erben, "Zu Rudolfs Barlaam und Josaphat," in Germanistische Studien, ed. J. Erben and E. Thurnherr (Innsbruck, 1969), pp. 34-35. g Jean Sonet, Le_ Roman de Barlaam et Josaphat: Recherches sur la tradition manuscrite latine et francaise (Namur, 19 49T7 7 Heinz Rupp, "Rudolfs von Ems Barlaam and Josaphat," in Dienendes Wort: Festgabe Bender (Karlsruhe, 19 59) , p. 11. g H. de Boor, Die httfische Literatur (Munich, 1953), p. 187. 9 see Roy Wisbey, "Zum Barlaam und Josaphat Rudolfs von Ems," ZfdA 86 (1955/56), 294. 4 the earliest reprinted texts in the history of German medi eval philology. It was only in the last decade that Barlaam began to receive greater consideration, primarily due to the efforts of H. Rupp's two essays and his reprint of Pfeiffer's edition. Since then there has evolved some discussion as to the literary evaluation of this work: does it manifest an inherent religious crisis of its author; is it purely contemptus mundi poetry (according to de Boor, Die httfische  Literatur, pp. 177 and 181); or is the accent more on the work and effect of the Saint within the world, and does the author actually disassociate himself from the idea of asce ticism presented in the Latin source, as R. Schnell postu lates in accordance with Rupp's interpretation?"^ Does the 12 "httfische Form" stamp the character of this religious work, or Is it to be understood almost as didactic literature (lere) and "Exemplum eines dieser lere entsprechenden Welt-verhaltens," according to H. Brackert (pp. 214-2 20)? And how does this relate to the verdict of "Epigonentum" which has been commonly applied to Rudolf's works? Does his Heinz Rupp, "Rudolf von Ems und Konrad von Wurz-burg," Der Deutschunterricht 17, No. 2 (1965), 5-17. See also footnote 7. RUdiger Schnell, Rudolf von Ems: Studien zur inne-ren Einheit seines Gesamtwerkes (Bern, 1969) , pp. 84-115. 12 Xenja von Ertzdorff, Rudolf von Ems_: Untersuchungen  zum httfischen Roman im 1_3. Jahrhundert TMunich, 1967), pp. 216 and 349. 5 reshaping of the Barlaam legend point into the future of this genre, and is Rudolf therefore not a mere imitator of the great medieval epic authors, but rather a forerunner of 13 later developments, as H. Rupp sees it? Or is there in Barlaam an ambivalent mixture of idealizing and problematici-zing tendencies, an unsolved conflict between legend and courtly romance, which would characterize the work as epi-gonal"?14 These questions must remain open in this context. In order to answer them, it would be necessary to make an in tensive analysis of the text in respect to its Latin and 15 Greek precursors, as well as other similar literary works of the time (Saints' vitae, courtly legends and romances) to determine the specific position of Barlaam in terms of its genre. We have merely alluded here to the var ious facets of this work of Rudolf von Ems and to different approaches adopted by literary critics during the past dec ade in order to achieve a better understanding of the text -s c and reassess the rank of its author. 13 H. Rupp, "Rudolf von Ems und Konrad von Wurzburg," 13. 14 Ulrich Wyss, "Rudolfs von Ems Barlaam und Josaphat zwischen Legende und Roman," in Probleme mittelhochdeutscher  Erzahlformen, ed. P. F. Ganz and W. Schroder (Berlin, 1972), pp. 214-238. x^ The study by Hannah Czizek, "Rudolfs von Ems Barlaam  und Josaphat und seine lateinische Vorlage," diss. Vienna 19 31 applies questionable categories and is of little use. 6 1.2. The text tradition The aforementioned contributions, diversified as they are in their methods and interpretations, concur in one respect: they do not take into account the manuscript tradition of Barlaam. The only available edition of this work, 130 years old, seems to be tacitly accepted as presenting the "right" text and is not expressly questioned. An evaluation of the existing Barlaam manuscripts and their text versions has not been attempted so far, although it would seem important to come to positive conclusions regarding the text tradition —which also includes the text reception by later scribes— before drawing any inferences as to the intention of the author and the genre of the work. As long as the text it self does not stand on firmer ground, interpretations of it cannot be well-founded. Let us take, for example, the discussion over the mean ing of the author's digressions (the courtly "Damenpreis," 11735-870 ,. and the "Schimpf rede," 12259-289) in the context of the entire work.. None of the modern interpreters has remarked on the fact that these passages appear complete in only four of the twelve major manuscripts that we know of (in A, B, b, and E, as opposed to C, D, G, Ka, Kb, Kc, L, and the Vienna manuscript W), and none has examined the con sequences which could be drawn from this fact with regard to the reception or possibly even the conception of Barlaam 7 und Josaphat. Or, to cite a lesser example, H. Rupp bases his thesis that Josaphat is summoned to be active within the world, to fulfill his position as a Christian ruler of a country partly on verses 6571-75, in which Barlaam admonishes his disciple not to follow him to his hermitage. wis ein bredigaere gotes unde ein lerer sins gebotes, wan dus gar gewaltic bist: alhie so kreftic niemen ist, der wider dir getlirre sin ..." H. Brackert, on the other hand, contradicts Rupp's argument as follows: "Doch vergleichen wir den Text. Barlaam sagt: 'wan dus gar gewaltec bist.' Der Genitiv bezieht sich auf die beiden vorhergehenden Substantive bredigaere, lerer. Es ist also keineswegs vom Ftirstenamt schlechthin die Rede" (Brackert, p. 217). On examining the manuscripts, we con clude that Pfeiffer's reading in 6 57 3 is based only on man uscripts E and D, strictly speaking. Manuscripts A and G come very close ("wan du des gar gewaltic bist," and "wan du is gar gewaltic bist"), and also manuscripts C, E, L, and W read basically the same ("wan du sein gar ..." referring thus to gebote). Rupp's explanation is supported, however, by the reading of K ("wan du gar gewaltic bist") and equally by the Berlin codex K which omits verses 6571-74 entirely and continues by reversing lines 6575 and 8 6576 ("wan niemen geturre wider dir sin"), which would allude to the "FUrstenamt" rather than the "Predigeramt" which Brackert emphasizes primarily. To be sure Brackert's interpretation is supported in this case by more manuscripts than that of Rupp. However, the "correct," or rather the most probable, reading cannot be taken from a purely numeri cal "majority decision" from the manuscripts, as long as their relationships one to another have not been examined. The rediscovery of Rudolf's Barlaam is connected with the names of the great literary critics of the Enlighten ment, Gottsched and Bodmer. The seventh volume of the lit erary magazine Beytrage zur critischen Historie der deutschen, Sprache (Leipzig, 1741, 406-4-14) , edited by Johann Christoph Gottsched, contained a short description by Conrad Arnold Schmid and an extract of almost 200 verses of an incomplete manuscript found near Llineburg (later bought by the British Museum). Shortly after this, Johann Jacob Bodmer printed several fragmentary passages of Barlaam und Josaphat in addition to 17 his Nibelungenlied text. Bodmer had received both manu scripts, the Nibelungen codex C and the Barlaam manuscript A, from the library at Hohenems Castle, and followed them 16 See H. L. D. Ward, Catalogue of Romances in the  Department of Manuscripts in the Britis~h~Tluseum, II (London, 1893), 142. 17 Chriemhilden Rache und die Klage, zwei Heldengedichte  aus dem schwabischen Zeitpuncte (Zurich, 17 5 7), pp. 251-286. 9 in his edition. His Barlaam excerpts are headed by titles and present the following sections: "Anfang des Gedichtes" (1-62), "Eingang" (125-164), "Liturgica" (6673-6956), "Vor-trefflichkeit der christlichen Religion" (10825-10950 and 12747-12894), "Hymnus" (139 0 7-14049), "Traum" (12 32 5-125 32), "Ablegung der Krone" (14751-14904), and, finally, "Ende des Gedichtes" (16022-16164). Bodmer seemed to be interested merely in offering his public a characteristic selection of the work, and not in preparing a complete edition of Bar  laam, which, after all, was only of minor interest to him compared to his NibeTungenlied studies. He did, however, take a first small step in the direction of a critical edi tion, in that he annotated his first passage with several variants of a Strassburg manuscript (probably the large fragment, Pfeiffer's sigle: a). This Hohenems manuscript used by Bodmer came into the possession of Freiherr Joseph von Lassberg at the beginning of the nineteenth century, who claimed it to be Rudolf's own handwritten work. Bodmer had already shown a much more enlightened judgment on the 18 value of his two Barlaam manuscripts, and Karl Lachmann in his letter to Jacob Grimm of March 27, 1821 dismisses 19 Lassberg's assertion rather ironically. 18 "Es scheint, dass jeder Schreiber sich eine eigene Buchstabierart erfunden, und grosse Freyheiten genommen habe." (Chriemhi1den Rache . . ., p. 253). 19 Briefwechsel der Briider Jacob und' Wilhelm Grimm mit  Karl Lachmann, ed. A. Leitzmann (Jena, 1927) , pp. 289-290 . 10 The first complete edition of Barlaam und Josaphat was published in 1818 by Friedrich Karl Ktipke, with an appendixed list of corrections and commentaries contributed by the 20 young Karl Lachmann. Kfipke, who was Gymnasialprofessor in Ktinigsberg at that time, based his edition on two manuscripts found in the Kttnigliche Btichersammlung at Kfinigsberg (Pfeif-fer's sigles Ka and Kb), as well as on a manuscript preserved in Berlin (Pfeiffer's sigle Kc), of which he obtained a copy written for him by J. G. Btisching. In addition K5pke men tioned some of the variant readings of the Bodmer selections. Kttpke defended the method of his edition in his preface as follows: "Es sind in den neuesten Zeiten uber die Art, wie alte deutsche Gedichte herausgegeben werden sollen, ver-schiedene Ansichten bekannt geworden, so lange aber aus diesen noch nicht ein bestimmtes Ergebniss gezogen werden kann, schien es am gerathensten, die alteste von den Hand-schriften, welche zu Gebote standen [this would mean Ka], zum Grunde zu legen und von dieser nur dann abzuweichen und die Lesart einer andern aufzunehmen, wenn die erste einen entschieden verderbten Text bietet; alsdann muss aber frei-lich in den Lesarten Nachricht davon gegeben werden. Dieses Verfahren ist bei nachfolgendem Abdrucke beobachtet worden" (KSpke, pp. VII-VIII). 20 F. K. Kttpke, Barlaam und Josaphat von Rudolf von  Montfort (Kttnigsberg, 1818) . We quote from its second edi-tion (Leipzig, 1838), henceforth referred to as "Kfipke." 11 This principle of editing—following a lead manuscript —appears nowadays perfectly legitimate. It is however a long way from the method of textual criticism which became standard for medieval editions by the second half of the nineteenth century. It is understandable that Lachmann, who is regarded as having initiated this method in the field of German philology, expressed some cautious reservations in the appendix of Kttpke1s edition: "Ubrigens ist Ihr Streben sowohl wie meines nur auf einen lesbaren Abdruck gegangen: zu einer kritischen Ausgabe fehlte es an Hulfsmitteln" (Kopke, p. 436). Since the two Kfinigsberg manuscripts in all probabil ity were destroyed at the end of World War II, we remain dependent on Kttpke's edition as far as our investigations into Ka and K*3 are concerned. For this reason, it appears necessary to respect Lachmann's opinion, since he had com pared Ka and with Kflpke's version. First, he makes sev eral negative comments on the value of Ka, remarking on ". . . die ungeheure Menge von Schreibfehlern, die schlechte Orthographie, und die nur seiten schfine, aber sehr ungleiche Schrift . . ." (KSpke, p. 428). While this judgment may primarily reflect Lachmann's own idealistic concept of a uniform Middle High German poetic language, his verdict on Kttpke's reliability as an editor cannot be overlooked. In his correspondence with Jacob Grimm, Lachmann gives free rein to his displeasure. He not only reproaches Ktipke for 12 being "borniert, unwissend, trage und lacherlich eitel," he also accuses him of dishonesty and deliberate deception (in another edition, Ktipke had tacitly omitted several ver ses). He recalls ". . . wie 1816 bei meiner Ankunft sein druckfertiger Barlaam, mit dem Glossarium von 6-8 Quartblat-tern, aussah (es fehlten ganze Verse, von der schlechten Orthographie war eben das fehlerhafteste beibehalten, sammt alien Schreibfehlern, in der ersten Halfte stand daz, in der zweiten das; im Glossarium kein Citat, aber enwizzen 21 und andre Ungeheuer)." The only information given by Kttpke which we can still verify concerns his notes to the Berlin manuscript K . Many of its pages are not at all annotated, but at least KSpke mentions the many omissions in K , albeit partially incorrectly: instead of 157,27 - 158,25 it should be 157,7 - 158,35, and instead of 356,31-32 it should rather be 356, 23-24 and 356,27-28 in KSpke1s edition (this corresponds to 6229-6297 as well as 14307-308 and 14311-312 in Pfeiffer's edition). Neither Kfipke nor Lachmann have given any infor mation as to the placement of initials in the Kfinigsberg manuscripts. Therefore, we can merely suppose that a para graph in Kttpke's edition compares with an initial in the text of Ka. In spite of these objections, we must naturally use Kttpke's text as representative of Ka, but we are entitled 21 Briefwechsel, p. 223. Letter of November 5, 1820. 13 to some scepticism as to its accuracy. Moreover, the mere b c lack of a variant for K or K in Kfipke's apparatus does not prove eo ipso that they share a particular reading with Ka. Five years after the second printing of Kfipke's Barlaam  und Josaphat Franz Pfeiffer published his critical edition of this work. He named the author no longer after his over lord, the Count of Montfort, but rather according to his place of origin, Hohenems (thus following the continuator of Rudolf's Weltchronik). In the preface of his edition Pfeiffer sets himself apart from Ktipke with self-confidence: "Seinen Zweck bloss einen lesbaren Abdruck zu geben hat er ohne Zweifel erreicht, und wenn der Abdruck auch Manches zu wunschen tibrig liess, so waere es doch unbillig, den Mass-stab unserer Zeit daran legen zu wollen. Mein Streben gieng dahin, eine Ausgabe zu liefern, wie sie der gegenwartige Standpunkt der Wissenschaft verlangt" (Barlaam, p. XIV). This recent stage in literary criticism to which Pfeiffer refers is embodied at its best in Lachmann's second edition 2 2 of Hartmann von Aue's Iwein, which also appeared in 1813. We can allude to this method briefly as that of classical philology, introduced by Lachmann into the field of medieval text editions. According to the usual description it con sists of three steps: of recerisio (critical examination of 22 We used the sixth edition (Berlin: de Gruyter, 1962) which contains the original "Anmerkungen und Lesarten zum Iwein" by Benecke and Lachmann. 14 all manuscripts, investigation of their degree of relation ship) , emendatio (elimination of errors in the text tradi tion, securing the best text version), and conjectio (hypo thetical reconstruction of the original reading in face of 2 3 a corrupt passage. Measured against these standards, Pfeiffer certainly did not reach his goal—but the same holds true for Lach-24 mann's own editorial work. To begin with, Pfeiffer did not base his edition on "den gesamten damals bekannten 2 5 Handschriften-Bestand," as J. Klapper maintains, but only on the following six manuscripts: A, B, C, D, E, and b (the incomplete Heidelberg manuscript which Pfeiffer lists under his fragments). In addition, he draws on fragment a and three lesser fragments c, d, and e, as well as (for the Parable of the Unicorn only) on fragments f and g. Pfeif fer' s apparatus contains also different readings of the Kttpke edition (Pfeiffer's sigle K), without however passing on its variants for K and K . Pfeiffer's choice of 2 3 See Friedrich Neumann, Studien zur Geschichte der  deutschen Philologie (Berlin, 1971), pp. 17-18. 24 The discrepancy between Lachmann1s rigorous theore tical demands and his own practice in the field of textual criticism has been pointed out convincingly by Rudolf A. Hofmeister, "Lachmann1s Role in the Transmission of Parzi  val," Seminar X, 2 (1974) f 87-100. 2 5 J. Klapper, "Barlaam und Josaphat," in Verfasser-lexikon, ed. W. Stammler, I (Berlin, 1933), p. 170. 2 6 Except for two cases, the omission of the author's digressions in K^ and Kc (see Barlaam, pp. 449 and 451). Even here Pfeiffer's information is partically faulty in its details. 15 sigles is generally unfortunate since it does not differ-tiate between vellum and paper manuscripts by using capital and small letters, it does not rank b among his major manu scripts, and it confuses sigle K and Ka. There were five further manuscripts which Pfeiffer should have been acquainted with, as they were already mentioned in von der Hagen's 2 7 L11erarischer Grundriss. These were G, the Gotha manuscript H, a manuscript in Hamburg M, the incomplete London manuscript mentioned above (p. 8) N, a manuscript formerly owned by Raimund Krafft at Ulm W, a former Ambras codex in Vienna. Pfeiffer mentions G and W at one place in his apparatus (Barlaam, p. 449, concerning the author's digression), but he had apparently not consulted them himself. Furthermore, Pfeiffer did not undertake to study in detail the possible relationships of his manuscripts; he contents himself with a categorical remark on the affinity of D and "K" as well as of B and b. On the other hand, he did not attribute an outstanding value to any of his Bar laam manuscripts which would have allowed him to follow it as a "Leithandschrift." Therefore he feels entitled to select the appropriate reading from any one of the manu scripts, depending on his own judgment alone (see Barlaam, 27 Literarischer Grundriss zur Geschichte der Deutschen Poesie von der altesten Zeit bis in das 16. Jahrhundert, ed. F. H. von der Hagen and J. G. Busching (Berlin, 1812), pp. 282-94. 16 p. 409). It cannot be expected that Pfeiffer's apparatus contains the divergent readings of all manuscripts for each verse; Lachmann himself emphasizes the need for withholding such 2 8 superfluous information in his Iwein commentary. But we cannot overlook the fact that the variants provided by Pfeiffer are frequently misleading or faulty. As proof, let us rectify here merely some of Pfeiffer's information regarding the omission of verses. 91,21-22 are not missing in E; 120,11-12 and 120,33-34 are missing in b; 155,17-18 are omitted in E (and KbG); 230,36-38 are omitted in E; 275,32 and 277,10 are not omitted in d; three leaves are missing in C between 276,23-389,30; 402,3-4 reversed in Ka and Kc, but not in A. These are only a few out of many more incorrect indications, but the amount of suppressed information of that kind is even far greater in Pfeiffer's apparatus. Moreover, Pfeiffer neglected to mention that C and E insert Latin Bible quotations in certain passages. In one regard Pfeiffer went beyond Lachmann's practice, in that he included paragraph markings in his critical apparatus. But here, too, his information is so sporadic and inaccurate that it can only be considered a step in the right direction. This lack is particularly regrettable in the case of manu scripts B and a which have meanwhile been destroyed. 28 Iwein, 6th ed. (Berlin, 1962), pp. 362-63. 17 During the past 130 years a great number of additional text witnesses have been discovered. However, a small por tion of these has yet been evaluated in the context of the entire text transmission. The first and so far only published attempt in this regard was made in the dissertation by F. 2 9 Stthns. S6hns based his study mainly on Pfeiffer's appara tus which explains many of its flaws. In addition to it, he took into account one further, manuscript (L) and six frag ments (h, i, k, 1, m, n). Regarding fragments m and n, Stthns drew his conclusions from second hand, according to the manuscript descriptions and collations of Diemer and Minzloff (see chapter 2.2). Of fragments i and k SGhns had obtained a handwritten copy, whereas he himself examined only L, h, and 1. Stthns's own collations of L, h, i, k, and 1 are added by H. Rupp to his reprinting of the Pfeiffer edition (Barlaam, pp. 464-505). This appendix is of rela tively little value however, as L shows a very corrupt text 3 0 version, and two of the fragments, l and k, rank low com pared with others. H. Rupp is well aware that his reprint can only be a temporary solution (see his "Nachwort," Barlaam, p. 512). It has become obvious that future research into the struc ture and meaning of Rudolf von Ems1s Barlaam und Josaphat can no longer be based merely on Pfeiffer's and Stthns's 29 Franz Sflhns , Das Handschriftenverhaltnissin Rudolfs  von Ems Barlaam, Diss. Erlangen 18 78 (Erlangen, 18T8'J~ 30 See F. J. Worstbrock's review, ZfdA 77 (1966), 114. <3 18 readings in the last instance. For the time being, the manuscripts themselves have to be consulted in each case. This impractical and time consuming procedure would make a 31 truly critical edition highly desirable, but this goal cannot be reached without extensive preliminary studies. Our purpose, therefore, is to contribute to this end by making an inventory of all existing manuscripts, assessing their possible relationships, and comparing their rubrica-tion. 2. The Barlaam manuscripts The following survey proceeds from the information provided by Pfeiffer and S films—updated, corrected or completed whenever deemed necessary—and is based for most of the remaining fragments on the listing established by Worstbrock (see above). In order to avoid confusion, the sigles intro duced by Pfeiffer and Sfihns will be kept here; furthermore we follow Worstbrock's numbering of fragments (our sigles "Fl" to "F18"). Deviations from this procedure will be accounted for. Information on the age and material condi-31 Apparently, a new edition is planned by Siegmund ^W^A ^ Prillwitz in Hamburg (according to his notei(of June 2 , 1973). His unprinted thesis "Rudolfs von Ems BuJ. Uberlieferung und lateinische Vorlage" (list of dissertations in progress in Jahrbuch fur Infernafionale Germanistik, II, 2, No. 1828) has not been available. 19 Irion of manuscripts is usually taken from the pertinent library catalogues or manuscript descriptions which are only explicitly mentioned when a point is under discussion (see 32 chapter 4). 2.1. Major manuscripts Pfeiffer's text consists of 16164 verses, but none of the existing manuscripts contains this text in its entirety. Due to text omissions or physical damage, they are more or less reduced in size and could all be called "fragmentary". Therefore it does not seem logical to list manuscript b under "Bruchstucke," as Pfeiffer does, while it has pre served nearly as much text as C. Likewise, we rank the al ready mentioned "London fragment" among the major manu scripts, since it presents three extensive text sections from the beginning, the middle, and the end of the work, altogether more than half of all the verses. The smaller fragments, on the other hand, consist only of very few leaves, the largest of them does not even contain one tenth of the entire text.. The major Barlaam manuscripts which we used are: A formerly at Hohenems, now Fttrstlich FUrstenbergische Hofbibliothek, Donaueschingen. Vellum, thirteenth to \ fourteenth century; 16122 verses. Microfilm. sJLuU^ 32 ' We would like to thank all the libraries mentioned in this section for their assistance in providing micro films. 20 b Universitatsbibliothek Heidelberg, Cod. Pal. Germ. 811. Paper, fourteenth century; twelve leaves are missing, 14052 verses. Microfilm. C Bayerische Staatsbibliothek MUnchen, Cgm. 16. Vellum, 1284; eight leaves are missing, 14292 verses (including 58 verses of the beginning which were later added). Microfilm. D Universitatsbibliothek Freiburg i. Br., Hs. 480. Vellum, thirteenth to fourteenth century; five leaves are mis sing, 15234 verses. Microfilm. E Bayerische Staatsbibliothek Munchen, Cgm. 273. Paper, 1459; 16118 verses. Microfilm. G formerly at Gotha, now Niedersachsische Staats- und Universitatsbibliothek, Gttttingen, 2° Philol. 188/10. Vellum, fourteenth century; 15966 verses. Microfilm. Ka formerly Universitatsbibliothek KHnigsberg, Hs. 89 8, missing since 1945. Vellum, fourteenth century; ca. 15660 verses. Text taken from Kflpke's edition (see above pp. 10-12). Quoted as K, unless Kfipke's or Lachmann's commentary explicitly assign a reading to Ka. K Staatsbibliothek Preussischer Kulturbesitz, Berlin, Germ. Fol. 20. Paper, fifteenth century; 15410 verses. Microfilm. L Universitatbibliothek Bonn, S 502. Vellum, fourteenth century; 15 59 0 verses. Microfilm (loan). 21 M British Museum, Additional MS 10,288. Vellum, fourteenth century. Three large sections with missing leaves in between, 9142 verses. Listed by Worstbrock as fragment No. 12. Our sigle: M. Microfilm. W Osterreichische Nationalbibliothek Wien, Cod. Vind. 2884. Paper, fourteenth century; 16028 verses. Our sigle: W. Microfilm. The following manuscripts are either not preserved or unavailable: B formerly Johanniter Bibliothek Strassburg, A 144. Vellum, fourteenth century. Destroyed in 1870. Some variants and paragraph indications in Pfeiffer's apparatus. Kb formerly Universitatsbibliothek Kfinigsberg, Hs. 89 0b. Vellum, fourteenth or fifteenth century. Missing since 1945. Some variants in KSpke's apparatus. H formerly Staats- und Universitatbibliothek Hamburg, Cod. Germ. 19 (acquired from the Uffenback collection in Frankfurt). Paper, fifteenth century. Missing since 1945 (see Worstbrock, 112). Manuscript M contains a few in serted pages on which J. J. Eschenburg copied short pas sages from H in order to integrate the three sections of 3 3 M into the context. Our sigle: H. N until 17 39 owned by Dr. Raymund Krafft at Ulm (see Worstbrock, 113), missing since then. Vellum. Bernhard 3 3 • John Koch, "Fragmente von Rudolfs von Ems BuJ in einer Hs. des Britischen Museums in London," ZfdPh 13 (1881), 78-89, gives short samples of Eschenburg's copy. 22 Docen in his review of Kttpke's edition mentions "... die Kraftische Handschrift zu Ulm (alt und gut, aber wo jetzt?)."34 Our sigle: N. P privately owned by H. P. Kraus, New York, formerly Bib-liotheca Bodmeriana, Cologny-sur-Geneve. Paper, 14-69. Only illustrated Barlaam manuscript with 13 8 full-page pen and water-color drawings from the atelier of Diebolt Lauber at Hagenau. Ca. thirty leaves (out of 3 79) are 3 5 missing. Our sigle: P. 2.2. Smaller fragments d Zentralbibliothek ZUrich C 79c I. Vellum, thirteenth century, two leaves. Photocopy. e Bayerische Staatsbibliothek Munchen, Cgm. 5249. Vellum, thirteenth century, one double leaf. Microfilm. h Universitatsbibliothek WUrzburg. Vellum, thirteenth century. Eight leaves, slightly damaged. Microfilm. i Staatsbibliothek Preussischer Kulturbesitz, Berlin, Germ. Fol. 720a. Vellum, thirteenth to fourteenth cen tury, two leaves. Microfilm. k Staatsbibliothek Preussischer Kulturbesitz, Berlin, Germ. Fol. 720b. Vellum, fourteenth century. Three 34 Docen, Wiener Jahrbucher der Literatur, XI (1820), 113. 35 ... According to the detailed manuscript description which Mr. H. P. Kraus kindly provided. 23 leaves, greatly damaged. Microfilm. 1 Bibliothek des Germanischen Nationalmuseums Ntirnberg. Vellum, thirteenth to fourteenth century. One double leaf, damaged. Microfilm. m "Gttttweig fragment," its location could not be esta blished. Two vellum leaves, thirteenth century. Studied 3 6 from Joseph Diemer's description and list of variants, n M. E. Saltykov-Shchedrin State Library Leningrad, No. 2568. Two and a half leaves. Photocopy. p privately owned by Gerhard Eis, his signature: 153. Vellum, fourteenth century. One leaf, badly damaged. 3 7 Studied from G. Eis's transcription. Our sigle: p. q Staatsarchiv, Schaffhausen. Vellum, thirteenth century. One leaf, damaged. Studied from Peter Ochsenbein's 3 8 transcription. Sigle q introduced by Ochsenbein. r Staatsarchiv, Schaffhausen. Vellum, fourteenth century. One leaf, damaged. Studied from Ochsenbein's transcrip tion. Sigle r introduced by Ochsenbein. Fl Universitatsbibliothek Basel, N.I.4 BI. S. Vellum, thir teenth century, one leaf. Photocopy. 3 6 J. Diemer in Sitzungsberichte der Akad. der Wissen  schaft en. Phil. Hist. Klasse. Vienna, xi-(1853), 640-53. 37 G. Eis, "Ein neues Fragment von Rudolfs von Ems BuJ," GRM 49 (1968), 448-50. 3 8 We are much obliged to Dr. Ochsenbein (Basel) for sending a copy of his transcription. F2 Staatsbibliothek Preussischer Kulturbesitz, Berlin, Germ. Fol. 737, 16-18. Vellum, thirteenth century. Two leaves, damaged. Microfilm. F3 Staatsbibliothek Preussischer Kulturbesitz, Berlin, Germ. Fol. 737, 20-21. Vellum, thirteenth century. Two leaves, damaged. Microfilm. F6 University Library, Wroclaw, No. R3259. Vellum, four teenth century. One double leaf, slightly damaged. Photocopy. F7 Universitatsbibliothek, Freiburg, Hs. 529. Vellum, thirteenth to fourteenth century, two leaves. Microfilm. F8 Hessische Landesbibliothek, Fulda, Hs. C4a. Vellum, four teenth century. One badly damaged double leaf. Photo copy. F9 Niedersachsische Staats- und Universitatsbibliothek, GBttingen, Philol. 189b. Vellum, thirteenth century. Two leaves. Microfilm. F10 Kestner-Museum, Hannover, Inv. Nr. 3979a/b. Vellum, thirteenth century. One and a half leaves, greatly dam aged. Photocopy. Fll Furstlich Oettingen-Wallerstein1sche Bibliothek und Kunstsammlung, Schloss Harburg, 1,3,4°, I. Vellum, four teenth century. One badly damaged double leaf. Microfilm F13 British Museum, Additional MS 10,288, ff. 157, 158. Vellum, thirteenth century. One leaf, slightly damaged. Microfilm. 25 F16 National Museum Library, Prague, IE a 7. Vellum, four teenth century. One double leaf and another greatly damaged leaf. Microfilm. F18 Deutsche Staatsbibliothek, Berlin (DDR), Fgt. 93b. Vellum, fourteenth century. One double leaf, greatly damaged. Microfilm. The following fragments were unavailable: a formerly Johanniterbibliothek Strassburg, A 94. Des troyed in 1870. Vellum, fourteenth century, ca. twenty leaves. Some variants in Pfeiffer's apparatus. c formerly owned by Gymnasialprofessor Heinrich Schreiber at Freiburg, but not to be found among his papers in 3 9 the City Archives at Freiburg. Vellum, fourteenth century, four leaves. Some variants in Pfeiffer's appa ratus . F17 formerly K5niglich-ttffentliche Bibliothek, Stuttgart,40 but today not registered in the Wurttembergische Lan-41 desbibliothek at Stuttgart. Vellum, fourteenth to fifteenth century, one damaged leaf. Worstbrock lists incorrectly the damaged double leaf Germ. Fol. 923 Nr. 2 of the Staatsbibliothek Preussischer Kultur-besitz which, however, represents another independent Bar-39 We are obliged to Dr. W. Hagemaier of the Univer-sitStsbibliothek Freiburg for this information. 40 Hermann Fischer, "Fragment aus BuJ," Germania 30 (1885), 102-103. 41 Letter of June 12, 197 3. Worstbrock's information is erroneous. 26 laam version together with fragment C 79c II of the Zentral-bibliothek Zurich.42 Fragments which consist exclusively of one or several of the Barlaam exempla have not been considered. They were taken out of the context of the narrative and revised by various authors, among them Strieker or lesser poets in his 4 3 manner (the so-called "Strickerschule"). Thus they have a tradition of their own and can hardly shed any light on the Barlaam text transmission. This is true for f Osterreichische Nationalbibliothek, Wien, Cod. Vind. 2705, Nr. 87-92 ("Wiener Strickerhandschrift") g Universitatsbibliothek, Heidelberg, Cod. Pal. Germ. 341, ff. 188b and 202d ("Heidelberger Strickerhandr-schrift"). F5 Staatsbibliothek Preussischer Kulturbesitz, Berlin, Germ. Oct. 137, ff. 148v-150r F14 British Museum, Additional MS 24,946, f. 65r F15 Bibliotheca Bodmeriana ("Nikolsburger Bispelhand-schrift"), see Ute Schwab, Die Barlaamparabeln, p. 175. 42 This Zurich fragment is also wrongly attributed to Rudolf in Katalog der Handschriften der Zentralbibliothek  ZUrich. I. Mittelalterlxche Handschriften by L. C. Mohlberg (ZUrich, 1951), p. 45. 43 See Ute Schwab, Die Barlaamparabeln im Cod. Vind. 2705 (Naples, 1966). Unfortunately, this study also goes astray occasionally since it relies on Pfeiffer's apparatus. Example: Barlaam 118,37 (4697) reads "ein lutzel honicseimes" not only in E and f,g (as U. Schwab claims, p. 199), but also in M and i. 27 3. Study of rubrication 3.1. Textual criticism and comparison of initials "Recent studies in textual criticism mark the end of an age long tradition. The ingenious technique of editing evolved by the great masters of the nineteenth century has become as obsolete as Newton's physics, and the work of generations of critics has lost a good deal of its value. It is no longer possible to classify manuscripts on the basis of "common errors," genealogical "stemmata" have fallen into discredit, and with them has vanished our faith in compos ite critical texts."44 Before we undertake to study the relationship of the Barlaam manuscripts, we must question the validity of our project in the light of Vinaver's negative judgment. Vin-aver is influenced to a large degree by the arguments of Joseph Bedier. Both articles put forward three main objections to the traditional method of textual criticism (for the sake of convenience called "the Lachmann method," although Lachmann never propounded his ideas in a theoreti cal treatise), which could be summed up as follows: 44 \ Eugene Vinaver, "Principles of Textual Emendation," in Studies in French Language and Mediaeval Literature presented to M.. K. .Pope (Freeport, N.Y., 1939), p. 351. 45 J. Bedier, "La tradition manuscrite du Lai de 1'Ombre. Reflexions sur l'art d'editer les anciens textes," Romania LIV (1928), 161-196 and 321-356. 28 a) The criterion of "common errors" is declared misleading in the grouping of manuscripts. In its stead Vinaver gives a thorough demonstration of how deviations from the right version can frequently be the result of some mental slip on the part of the scribe during the copying process. b) The further back a genealogical stemma is traced, the more hypothetical and less meaningful it becomes. Most of these stemmas end, or rather begin, with an archetype and two major branches of text transmission descending from it, mainly because- the text researcher has been carried away by the "force dichotomique." "Le systeme lachmannien l'a lance dans la chasse aux fautes communes, mais sans lui donner aucun moyen de savoir a quel moment il a le devoir de s'arreter" (Bedier, 176). However, it is highly impro bably that no more than two copies were made from most of the originals. c) A possible interaction of several manuscripts is not suf ficiently recognized by the "genealogical method," which generally assumes a straightforward transmission from a 4-6 single source to a copy. Many of the medieval manuscripts however show traces of interference (contamination) from other sources. Bedier concludes from these and other objections that, 46 The traditional textual criticism declares itself powerless, indeed, against the occurrence of contamination* see Paul Maas: "Gegen die Kontamination ist kein Kraut gewachsen." Textkritik (Leipzig, 419 60) , p. 30. 29 in editing a medieval text, the editor should follow the best-transmitted manuscript, which should be emended only in the case of obvious flaws in the text. For such emenda tions there is no mechanical procedure based on a stemmatic grouping of the manuscripts, the editor must let his own taste ("gout") be the ultimate judge. As indicated earlier, Pfeiffer published his Barlaam edition before the genealogical method was at its apogee. Thus his readings are selected entirely on the basis of his personal preference and not on a systematic scheme. It was not until more than thirty years later that F. Stthns undertook to establish a stemma of Barlaam manuscripts, and this work seems to confirm Bedier's negative-ironic opinion. S5hns divided all the manuscripts into two major branches, BCLE on the one side, and ADKabc on the other, according to the criterion of common errors. Parallel readings such as waere - was, wirt - ist, ersehen - versehen are considered as proof of a genealogical relationship (SShns, pp. 4-5). Subsequently Sfihns attempts to determine the "correct" rea ding of one manuscript and one manuscript branch over another on the basis of "innere Grunde" and by comparing them with the Latin version. His conclusion in this res pect is negative: "Legen wir diesen Masstab der Kritik an die einzelnen Handschriften, so ergiebt sich im Ganzen dasselbe Resultat, das wir von den beiden Reinen behaup-teten. Es ragt keine an Qualitat entschieden vor den 30 andern hervor, es hat bald diese bald jene einmal die richtige mit der Quelle [i.e., the Latin source] congruente Lesart erhalten" (Sflhns, p. 24). The doubtfulness of Stihns's stemma was pointed out already in a contemporary review: "Der Herr Verfasser hat bei bei der Auswahl der zum Beweise angeftlhrten Lesarten hier wie anderswo viel zu wenig erwogen, inwieweit Handschriften auch zufallig und unabhangig von einander oder von einer gemeinsamen Quelle in einer Lesart zusammentreffen kfinnen, und daher ... in der Mehrzahl solche Stellen vorgefuhrt, 4 7 die nichts beweisen kttnnen." Certainly, Lambel's critique is not an attack on the method involved, but rather on its careless application in Stthns's study. However shaky this stemma may be, it has nonetheless 48 remained attractive enough to be reprinted and quoted as a standard of reference, even up to this day, as shown in an article by Gerhard Eis: "Die Einordnung des neuen Bruchstuckes ist mit Hilfe der Dissertation von Franz Stthns mflglich. Es gehttrt zur Gruppe BCLE, die sich deutlich von abc 49 der Gruppe ADK abhebt. . ." It is not surprising that Eis's classification of his fragment fails, since Sdhns's information was incorrect and incomplete (see below 4.3.5). 47 H. Lambel, Germania 25 (1880), 377. 48 See Paul Piper,' Hafische' Epik, Dt. Nat.-Lit., ed. J. Kurschner, 4. Bd., 1. Abtlg. (Stuttgart, n.d.), Ill, 561. 49 G. Eis, "Ein neues Fragment von Rudolfs von Ems Barlaam und Josaphat," GRM 49 (1968), 448-450. 31 To return to our opening question: in working towards a new Barlaam edition, is it possible to renounce completely an investigation into possible manuscript groupings and to content ourselves with printing one complete good manuscript? On first glance, this appears already problematic, consider ing the various lengths of the transmitted versions. The most extensive manuscript, the Hohenems codex A, contains numerous obvious mistakes as well as signs of a later revi sion, so that even Pfeiffer made use of it "nur mit grosser Vorsicht" (Barlaam, p. 408). The other nearly complete manuscript E is of a very late date (mid-fifteenth century) with unsuitable dialectal forms, and thus not applicable. The remaining manuscripts would require the insertion of missing passages, in which case again we would have to decide to which manuscripts to refer. To some extent a complete Barlaam edition would therefore be a "composite text" in any event. In our view, a diversified text trans mission such as that of Barlaam makes a preliminary study of manuscript relationships imperative. This does not mean that we believe in setting up a complete stemma, but rather that a comparison of the manu scripts would facilitate the choice of a lead manuscript pr of lead manuscripts). Based on it (or them), one would have to consult the main representatives of the other groupings in dubious cases and weigh their divergent read ings. This would not lead to a mechanical principle of 32 selection, but, it is hoped, towards achieving greater con sistency and accuracy in establishing a text version which would be more reliable than the one resulting from Pfeiffer's purely subjective approach. F. Whitehead's remarks con cerning a new edition of the Chanson de Roland could well be applied here to Barlaam: "It seems ... as though what is needed is less a new critical doctrine than a return to old and well-tried principles of textual criticism, which seem to have been strangely neglected . . . from the days 5 0 of the early eclectic editors down to our time." A cau tious return to the more traditional ways of textual crit icism, without the dogmatic pretension.of the late nine teenth century scholarship, has also been observed and 51 endorsed by Karl Stackmann. A study of the manuscript tradition of a medieval text remains an indispensable pre paration for a critical edition. The criterion of "common error" has until now been mainly applied to textual variants. It is rare that a cri tic draws attention to noticeable concurrences in the placement of initials or that such parallels are used as an argument for a possible relationship of manuscripts. Edi tors have almost traditionally neglected the exterior struc-^° F. Whitehead, "The Textual Criticism of the Chanson de Roland: An Historical Review," Studies in Medieval  French presented to Alfred Ewert (Oxford, 1961), p. 86. 51 K. Stackmann, "Mittelalterliche Texte als Aufgabe," Festschrift fur Jost Trier (Cologne, 1964), pp. 240-267. 33 tural marks in manuscripts, and where they have been regis tered (as in Lachmann's preliminary studies to his Parzival and Willehalm editions), they usually do not appear in the critical apparatus. Only recently has structural research paid special attention to these characteristics: "To some extent indications of the structure may be found in manu scripts. It seems therefore necessary to pay greater heed not only to formal principles themselves but also to those technical details of manuscript production, which, though frequently passed over by editors, may . . . assist in the 5 2 determination of this structure." Linke's study in par ticular pursues this aspect and takes the paragraph markings of the manuscripts as basis for determining the "authentic" 5 3 textual divisions of the works of Hartmann von Aue. B. Schirock's dissertation follows a somewhat similar course for Parzival. Schirock, however, stands in opposition to Linke when he considers the paragraph markings in the var ious manuscripts depending on their place and importance in a pre-established manuscript grouping. "Erst wenn wir die Uberlieferung der Gliederungszeichen auf dem Hintergrund der Handschriftenverhaltnisse, der Gruppenbildungen und Konta-minationen beurteilen, lassen sich gultige Ergebnisse ab-52 M. S. Batts, "Poetic Form and Medieval German Scribal Practice," JEGP LXII (1963), 702. 5 3 Hansjttrgen Linke, Epische Strukturen in der Dich tung Hartmanns von Aue (Munich, 1968) . 34 lesen. For this purpose, Schirock chiefly makes use of the work of Gesa Bonath who herself is one of the few schol ars to recognize the importance of initials "... als wichtiges Hilfsmittel zur Feststellung der Abhangigkeitsver-hSltnisse. . .1,55 Our study will take up this idea and investigate whether a systematic comparison of initials in Barlaam manuscripts can serve as a guide through the maze of seemingly contradictory readings. We shall set out from the following hypotheses which have to be differentiated later: a) Initials and other paragraph markings are to be regarded primarily as structural signs and not as ornaments. b) Initials are generally taken over from the source manu script by the scribe (and the rubricator) of a copy and not placed at random. c) A marked agreement between the initial patterns of two or more manuscripts clearly different from the practice of others indicates a "genealogical" relationship. The extent of coincidence in rubrication can be calculated in percen tage figures. Some obvious objections could be raised regarding 54 Bernd Schirock, "Der Aufbau von Wolframs Parzival," diss. Freiburg 1972, p. 63. ^ Gesa Bonath, Untersuchungen zur Uberlieferung des  Parzival Wolframs von Eschenbach. Germanische Studien, No. 238 (Llibeck and Hamburg, 1970), p. 53. 35 these three statements which should be discussed at this point. With regard to a): Initials also fulfill an aesthe tic function by creating colourful patterns on a manuscript page, sometimes they are of decidedly artistic character (illuminated initials). Their prime purpose is nevertheless to organize the narrative into smaller units. The rare case that a scribe places initials for purely aesthetic reasons without concern for the intrinsic structure of the 5 6 text is described for the Parzival codex G, and for the 57 Munich Tristan codex (Cgm 51). Such manuscripts would naturally be of no value for our purpose. The same would hold true for manuscripts in which initials are set accor ding to a mechanical principle, e.g., every thirty verses. In Barlaam manuscripts, however, there is no such regular ity nor any striking visual principle in the placement of the initials; one needs only to look at the first thousand verses, which have been uniformly handed down. With regard to b): In the field of textual criticism it is commonly acknowledged that there were wide differences in the reliability .of scribes. We distinguish the "good" scribe who preserved the version of his source manuscript without wilful alterations from the "thinking" scribe who tried to improve on the original. The same distinction can be applied regarding rubrication. It is true that the 56 B. Schirock, p. 97. 57 M. S. Batts, 699. 36 execution of the initial by the rubricator can be an addi tional source of errors, but such flaws can usually be iden tified rather easily. That a scrupulous scribe overlooks, misreads, or accidentally adds an initial occurs now and then. We suggest, however, that such errors are less likely at rubricated paragraphs than in the middle of a text pas sage. The scribe's concentration is more challenged at these points than anywhere else, since he has to leave out the initial capital letter, indent one line or even several lines, and write the required letter minutely on the margin ("Repr&sentant") so that the rubricator can execute the correct initial afterwards. Thus, the position and the reading of a paragraph beginning are more likely to be handed down through generations relatively undisturbed and can generally be considered more reliable than most text variants. Naturally, in the case of a "thinking" scribe who might introduce paragraph divisions of his own, it would be very difficult to determine his source just by looking at the rubrication. But again, other copies depen dent on his manuscript would be even more easily recognized, and this would result at least in a partial grouping. With regard to c): A coincidental agreement can occur with initials just as with readings. The scribes could set paragraph markings independently from one another at the same prominent places in the narrative (e.g., at the begin ning of a speech, change of location or time), where their 37 sources did not have an initial. Obviously this might lead to wrong conclusions with respect to a manuscript grouping, but this coincidence factor diminishes when we compare texts of greater length. Therefore it seems reasonable to express the agreement of rubrication in percentage figures among the larger manuscripts only: in comparing the smaller fragments we shall pay special heed to the placement of ini tials, but only in conjunction with a comparison of textual variants. The common variants at the beginning of a paragraph could also be of great importance, as they are usually pro-5 8 duced by a rubricator who misinterpreted a "ReprSsentant." We concur with A. Dain's remark: "Les fautes dues aux erreurs de rubrication font la joie des philologues et sont souvent d'un secours precieux pour le classement des manu-scrits."59 In spite of our statistical approach, we must not forget that we are dealing with literary products from which we cannot expect mathematical regularity. Bedier's words: ". . . le critique litteraire ne devrait jamais consentir ^ 6 0 . . . a s'effacer devant le statisticien" should prevail 5 8 Heinz Schanze provides some examples for the Willehalm manuscripts in "Dreissigerinitialen in der Willehalm-Handschrift 6," in Wolfram-Studien, ed. W. Schrtider (Berlin, 1970), pp. 174-176. 59 A Dain, Les manuscrits (Paris, 1964), p. 37. 60 J. Bedier, Romania LIV (1928), 329. 38 as a warning. Nonetheless, a statistically evaluated com parison of rubrication could provide textual criticism with information which, properly interpreted, could shed more light on the text tradition of Barlaam und Josaphat. 3.2. Chart of initials The following chart of initials sets out from the paragraphs in Pfeiffer's text (in the verse count of the reprinted edi tion) , in order to show on which manuscripts Pfeiffer based his subdivision. The manuscripts are subsequently listed in alphabetical order with the exception of M which follows after G for the sake of a better perspective. In the few cases where Pfeiffer's apparatus notes initials for B, these will be mentioned on the right hand margin, together with the initials of all fragments which will be dealt with in chapter four. Initials which match a paragraph in Pfeiffer's edition are marked as x on the same level, or as X in the case of a large initial. Initials at other places are indicated by the verse number (last three digits) for the first manuscript, underlined if it is a large initial. A straight vertical line symbolizes a loss of text in a manuscript due to physical damage, irregular vertical line indicates an omitted text passage. The numerous paragraph titles which can be found in some manuscripts (mainly in E, titles such as "hie chumbt 39 barlaam zue Josaphat") are not mentioned in the chart. They are usually added on the margin of the manuscript and do not constitute genuine structural marks. However, we indicated where there is a capitulum sign instead of an initial (in our chart as *)• Probably the scribe had over looked an initial in his source and afterwards marked the paragraph by a- capitulum sign on the margin. In a few cases it is doubtful whether there was meant to be an initial or not. Places where the rubricator clearly forgot to draw an initial—where a "ReprSsentant" or an indentation give evidence of the scribe's intention— are represented in the chart in parentheses. This is more difficult to decide in the case of the initial J (occurring very often with the name Josaphat). This letter is mostly drawn out on the margin and by its particular shape does not require any indentation at the beginning of the lines. It is not surprising to see that all large initials in the Barlaam manuscripts (with the exception of the very first initial "Alpha" and 3045 in L) are capital J's. For this reason alone, as well as for their very haphazard occur rence, we do not believe that these large initials con stitute the remains of an authentic major structure of the work ("Grossgliederung"), comparable to the one that B. Schirock tried to establish for Parzival on this basis. Nevertheless, we will keep the distinction between regular and large initials since it might be another aid for esta-40 blishing relationships. At places where a marginal initial J was most likely forgotten (e.g., "Osaphat"), we set an (X) . The size of "regular" initials is sometimes not con sistently maintained by the scribe (who allotted the free space) and the rubricator (who executed the design). In the beginning of manuscript D, initials vary between one and two lines height. In L the standard size of initials is increased after verse 1287 from two to three lines. The scribe of W, on the other hand, seems to have been fond of drawing out the shafts of capitals at the beginning of a line, mainly at predominant places where one might other wise expect an initial (e.g., 11603 and 12435). Here only a comparison with the usual practice of rubrication in W can tell which one is a true initial and should be taken into the chart. It would be of great help in some doubtful cases to check with the original manuscripts themselves as coloration and variations in ink do not show sufficiently on microfilms. But these few exceptions, even if misinter preted, could not seriously distort the overall statistical results. 41 A b C D E G M K K L V/ 1 X X x X X X x X X X 33 x x xxx 63 x x X x x x 75 89 X X X X X X X X xxx 125 x x X X X X 123 X 165 xxxXxxIxxxx 197 X X X X X X X XX X 227 xxx xxx X X 253 265 X X X X X X 295 X X XX X X 313 X XX 273 X X 42 A b C D E G M K Kc L M 335 x x x Xxx 3^1 369 x 401 x X X X X X X 431 X XX X X X 475 X X X X X 509 X X X X X X 5^5 X X X X X X 577 x x x x x x 617 XX X X xxx X X X X X xx x x F2 x x F2 x x x x F2 xx x F2 5^7 x x x x F2 x F2 659 x (x) XX X X x x x F2 43 A b C D E a M K Kc L W 691 XX XXX x x F2 737 xxxxxxxxxx*F2 769 xxxxxxxxxxx 799 xxxxxxxxx x 801 829 XX X X X X X X XXX 859 X XX X X X X X X X 887 x x XXXX - XXX 909 933 xxxxxxxxxxx 971 X X XX X X X XX X * 44 AbCDEGMKKcLW  1009 X X XX X X X X X X X 1039 XX XXXXXXX X 1071 XXXXXXX XXX 1103 x x x x x (x) x x x x F7 113 1135 xxxxxxxxxxx F7 II65 x x xxxxxxxxx F7 1191 x X x x x X 1227 xxxxxxxxxxx 1253 XXXXXXX X ' 255 1287 xxxxxxxxxxx 45 A b C D E G M K Kc L W 1317 xxx xxxxxxx.x 13^7 X X X XXXX X X X X 1379 xxxxxxx xxx 392 1423 xxxxxxx xxx 1455 xxxxxxx xxx 1491 xxxxxxxxxxx 1513 X X X XX X xxx 1545 XXX XX X X X X 1585 X X X X X X 1619 x XXXX XXXX 46 A b C D E G M K Kc L W 1651 X X X X X X 1691 X X X X X X 1729 X X X X X X 1765 XXXX X 1795 XXXX 815 X 1837 xx x x x x I863 X X X X X X 1897 x X 1937 X X X. X X X 1981 X X X X X X XXXX xxx XXXX X X X X X X X x x p x P 47 AbCDEGMKKL W 2019 x x x x 2051 x x x x x x x F? 2089 x x x x x F? 2131 X X X X X X x x x F7 2161 x 2196 X X X X 195 x X (175) x F7 B:2195 22*H x xx x x B 2281 x x x xxx xxx 2311 x XX X X X 2351 X X X X ' 3^3 x X X X X 48 A bCDEGMKKLW 238? X X X X 2^29 X X x x x 2463 x x x x x x 2493 X X X X X X 2529 X X X X X X 2569 x X X 2615 X X X X 26^9 xxx xxx X B X xxx 567 x B xxx x B 2693 xxx xx jxxxxn 2729 xxx xxx ixxxxn 49 A b C D E 0 M K Kc L W 276I XXX XX xxx 2797 X X X X X X X X XX 2837 x x x x x x 2881 x x x x x x 291*4- $t5 x x 933 29^7 x X X 937 x • B-.933 xxx 2985 X x X x 3019 x x x x X 009 x x x B 039 30^5 X X X X XX x X x 3083 X X X X X X 101 x X X 50 A b C D E G M K Kc L W 3125 x x x x x x x x x x m 3155 xx x x x x x x x x m 3177 X X 3187 X XXX X m 3225 XXXX 237 x x x x x m 3257 x x x x x x x x x x m 3287 xxxx x x m,B 3319 x x x x xx 339 x 3351 x xx x x x m x x x m 3389 xxxx x x m 399 51 A b C D E G M K Kc L W 3425 x x x x x x x x x m 3^59 x x X X X x 3511 xxx xx x B 35^1 x x x x X x xxxx 3573 x x x x B 3603 X X X X 3643 X X XX xxx 3675 XX X 679 3701 xxxx X X X 3733 x x x x x x 52 A b C D E M K Kc L W 3759 x x x x x x 3805 x x X x x X 3839 x x x x 3883 x 3915 x X xx X X 39^-1 x x x x xx 3969 x 991 3993 X X X X X X X X x x X X 819 x B X X Fl xxx F1.F8.B X X 992 4-013 x X X XX X X X XX Fl 4C43 x x x x x xx x Fl 53 A b C D E G M K Kc L W 406? x.xxxxxxx xx F8 4105 xxxx xxxx 4137 xxxx xxxx 147 4175 xxxx xxxxxxx F8 4207 x x 211 4233 xxx x X X XX X X X 4269 x x X 4301 xx x x x 4341 X 311 X X 325 X X X X 2£5 4375 xxx. X X x B X X 54 A b C D E G M K Kc L W 4391 x S x (x) x / X X 4413 xxxxxxxxxxxl 4^3 xxx Bs443 4457 x xx xxxxl 4491 X X X X XX X X 1 4527 x x x x xxx 1,B 5^3 4565 x xx x (x), x X X X X 1 4597 X X X X XX 603 X X 4629 x x x x 4663 xxxxxxxxxx 4705 x x • X X X 55 A b C D E G M K Kc L W 4737 xx x x 4767 xxxxxxxxxxi 775 x 4797 XX X X B 4829 x x xxxx 486l x xx x x x 4895 x 897 4931 x x 4963 x x 4991 X X 5027 X X xxxx X X X xxxxxxx xxxxxxxx 987 i»987 X X 5009 X X x X 041 x xxi x x x x x i 56 M K K L W 5059 x x x x x B 5091 xxxxxxx xxxi 5121 x x xx x x xx xx xi 143 xx i,k 5151 xxxx xxx 5189 xxxx XX X x F13 5223 x^xxx xxxxxx k,F13 5265 x 5307 x 5337 x 5369 x xxxxxxxxx k,F2,F13 273 XxxXxxxXX k,(F2) 321 XX XX xx x x x x F2 375 381 x x k ) ( 57 A b C D E 0 M K Kc L W 5396 x x x X X 5^3? x 546? X 5497 x xxxxxxxxx k,F2 447 X XX X X XX X X 479 X x x (X) x x x X X F2 5525 x X X X X X xxx 5555 x X XX X X 5585 x 5615 x 5649 5679 653 X X .599 X X X X X 631 X X X X 663 X X X 689 695 X X X x k, n k,n,B X X xxxx k,B 58 A bCDEGMKKclW 5713 x 57^5 x 5785 x 5817 x 5855 x 5893 x 5933 x 5965 x 5995 x 6031 x 703 x xx x x x x k xxxxxxxxxk x x x x x x 825 X X X x X X 973 x x X X X X X XX X X X X X X X B xxxxxxx F18 x x (X) X X X X X B.F18 x x B x x x x n,q 59 A bCDEGMKKcLW 6061 x x x x x x x x x x n, q 6099 X xxxxxxxxx 6131 X X X X X X x x q 6163 x X X x xx B 6197 189 XX X X X X 6229 x 6263 X XX xxx 2k7 x XX X x x x ra x X X m 6297 x 291 xx xxxxxx m.Fll 6329 x xxxxxxxxx m,Fll 636I xxxxxxxxx m.Fll 60 AbCDEGMKKcLW 639? x xxxxxxxx m,q,F11 6427 x X X x x x x m 6459 x XX X 467 xxx x x m 6491 x X X X XX xxx m,q,Fl6 6521 xxxXxxxxxm 6551 553 6587 x XXXXXXX X X X X X X 567 X X 6613 X XX XX X X 6645 x X X X X X X 6677 x X X 697 xxx Fll 61 A b G D E G M K Kc L W 6705 x 6733 x 676I x 6793 6825 x 6857 x 6891 x 6921 x •6957 x 6983 xxx x X X X B x x x x x x Fll B xxxxxxxxx XX XX xxxx 897 863 919 X X X X XXXXX xxx B 959 x xxxx xxxB 62 A bCDEGMKKcLW 7019 x X X x x B 029 X X 7057 x xxxxxxxxx 7087 x X x B 7123 x 7157 x xxxx xxxx 141 XX XXXXXX XX 7189 xxxxxxxxx 7223 XX X x 7255 x 7287 x 237 X X X X x x x X X X X X x x x B X ( 7321 xxxxxxxxxxx 63 A bCDEGMKKcLW 7351 xxxx xxxx 365 X X 7383 XXXXXXXX. XX 7415 x x X X ) (X) X 7447 xxxxxxxxxxx 467 x x 7473 XXX XX 7503 x X X X X X X X X 7537 xxxx x 541 X x X X 7569 xxxx B 7599 xxxxxxxxxxx 621 625 7637 xxxxxxxxxxx 64 A bCDEGMKKcLW 7667 X XX 7699 X X X X 7729 xxxx )xxxxxx 7759 x x x x x x xx x xx 7789 XXXX. XXX XX 7817 X XXX X X X 7847 xxxxxxx xxx 7877 xxXX xx XX 887 x x 7907 XXXX XXX XX 7939 xxxxxxx xxx 65 A bCDEGMKKcLW 7969 XXXXXXXX XX 8003 X xx X 8033 XXXXXXX XX 035 8063 X X XX X X X XX O85 X X 8101 xxx xxx xx 8137 xx x x x x xx x y 8165 X X X X 8195 XXXXXXXX X J 8223 XXX 8255 x x x x  : 66 A b CD E G M K Kc L W 8289 xxxxxxxxxxx 295 8321 x x x x 327 x x 8351 X X X X xxx 355 8379 X X X X 8409 x x x x 417 xxx xx 8447 x x xxXxxxXX 84?7 x x X X X X XXX 8509 x X X X X X X F6 8539 x X X. x (X) X x F6 8571 x x 561 XXXXXXXX F6 67 AbCDEGMK Kc L W 590 8602 601 xxx B«601 8631 X X x x B 8659 xxxx 8693 x x X X x X X x x X X F6 703 707 x -8729 xxxxxxx xxxF6 8761 xxxxxxx F6 8791 xxxx XX 8819 XXX X 8851 xxxxxxxxxxx 869 869 8881 xxxx xxx 895 XXX X X X X X xxxx XX X X XXXX x x x x xx 6816 x xx x 6£l6 xxxx 6ET6 x xx x x x 6606 xxxx $906 xx xx ££06 x x x xxx £006 6L6 X X xxxxxxxxxxx xxxx (X) xx Ti68 89 69 A bCDEGMKKcLW 9220 219 xxx 9247 x x x x 9277 x XXX X 9313 X X X X X X 93^3 x x 9373 9^03 x x 9433 9^63 x * 9^95 x X XX X xxx X X X X X X X XX X 331 X X X X 353 367 379 x B X X X X B xxx 70 A b C D E G M K Kc L 9527 x x 9557 x x 9591 x x 9629 x 9665 X X 9701 X X 9735 x 9765 x x 9799 xxxx 567 x x 599 x xxx 647 X X 679 X X ... 707 x XXX x x X 777 9829 X X XXX X X X X X X X X X X X X X 713 721 X 741 747 757 X X XX 71 A b C D E G M K Kc L W 9859 XXXXXX XX 9891 xxxxxx xxxx 9923 * X X X 99^5 xx xx x 957 x N x / 2^1 9987 X * j X X X ) x ( j 10017 X x I X / x / 10047 x x ' x x (X) xx X 071 X x IOO83 X X . X X X X X \ \ / ' / 1011? X * / X X X XX I X ; 143 x 10153 X X X X X X f 1 i j X 72 A b C D E G M K Kc L W IOI85 X X ) X X X X \ X / 198 ! 207 X X \ 10217 X X j X X X XX / X / \ 229 X X N 10251 X XXX XX XX XX 267 10289 x x x x x x x 307 x IO325 XXXXX X xx 10359 x x X x x x X X 373 x x x B 10395 xxxxxx xxxx 415 10419 xx x 425 XXX XX 10457 xxxxxx xxxx 10493 XXXX XX 501 x B.501 73 A b CDEGMKKCLW 10525 xxxx X X IO563 xxxx X X IO593 xxxx 10629 xxxx X X X IO657 X XX X X X ' 681 x IO685 X X X X xxxx 678 10717 x X X X X 107^7 xxxx X X 10791 X X X X X X xxxx 10825 X X X X 74 A bC DEGMKKCLW IO855 x x 10885 10921 x x 10951 x x 10983 x x X 11013 xxx 11043 x x X xxx xxx 969 x x (x) 999 X X 029 X XXX 11073 xxxx 11103 x x X x 11133 X X X X X X X X X X xxx 929 x x X x d,h,F10 x F10 x x x d x x X X d x x d,h x x h 75 A b C D E G M K Kc L W  11161 XXX 11191 x x X x x (X) x x X X F3 11223 x xx xx x x xx x h,F3 11257 x x x x x x x x 265 11287 xxxxxx xxxx 11317 x j x x (X) xx X FJ 11345 xxx xxxx 11375 XXXX X 387 x d 11407 x x X x x X d 11441 x x x x x x x x d 76 A b C D E G M K Kc L W 11475 x x x x x x 11507 xxxx x 521 11549 X X X XXX 11581 XX X II613 xxxx 11643 x x x x (621) 11671 X X X 11701 X XXX X X 11735 . x X II769 x X X X X X xxxx 603 /XX 645 663 X X / / / i ) X / \ I 77 A b C D EG M K Kc L W 1180? x x 11841 x 11871 x x 1 x \ X 11901 xxxx 11931 xxxxxx 11961 (x) x x x x 11991 X XX 12023 X X XXX X 12053 xxxxxx 12087 xxxx \ x x x x h x x x h x x x x h xxxx x x x h x x x h x x x x h x x h 111 Billl 78 A b C D E G M K Kc L W 12119 xxxx x x h 12149 xxxxxx xxx 1218? x x x x 12215 x 12247 x xxx 191 x 223 (X) x (X) ( 12289 x x x x xx 12325 x xx x 12357 192 164 x x (X) X 307 XXX I ) x 291 290 x x x h x x h 12391 xx x xx x x x x x xh 12421 xxXx xx xh 435 X X x (X) 79 A b C D E G M K Kc L W 12453 x x X X (X) x X X h 12487 xxxxxxxxx xh 12517 xx x h 523 533 x x 541 5^1 12551 xxxx x x h 573 x N B 12581 x x X x x X h 12613 xxxxxxxxxxx 12643 x x x x xx x x xx 12673 xxxxxxxx xx F10 679 12701 X XX X 12731 XXXX XXX XX 80 A b C D E G M K Kc L W 12765 xx x 12797 x xx 12829 x x x 12859 x x 12895 xx x 12933 xxx 12963 xxx 12993 x x X 13023 xxx 13053 xxx X XXX X xxxx X X X X X X x (X) X x X X X 079 X X 799 X X X x x x F3 x x x h x x x h 981 x (X) X (Fl6) x h xx h 81 A b C D E S M K K L W  13087 xxxx x x h 13119 xxxx xxxxxxh 13147 xxxx xx 175 13179 xxxxxxxxxxx 13213 xxxx xxxh B:221 231 XX X 13247 xxxx x 13281 XXXXXXXX XX 13311 xxxx xxxxxx 13343 X XX B«353 371 13375 XXXXXXX XXX 82 A b C D E G M K Kc I, W 13^07 xxxxxxxxxxx 1343? xxxx xxxxxx 13467 xxxxxxxxxxx 13499 xxxx xxxx 13529 xxxxxxxxxxx 13561 X X X XX XX X X X X 13591 xxxx xxxxxx 13621 x X X X XX 13651 xxxxxxxxxxx 1368I xxxx xxxx 83 A b C D E G M K Kc L W 13711 XXXX X X 731 13745 xxxx xxxx 13775 XXXX XX 781 X X 13811 xxxx xxxxxxx Fl6 13841 (x) * x x^ x x 13873 XXXX XX 13907 XX XXXXXXXX F9 929 13937 XXX 9^5 13969 X * X X xxx 14001 xxxxxx xxxx 84 " A ' b G D E G M K K L W_ 14033 xxxx xxx xx 055 14065 xxxx x x 079 1409? xxxxxxxxxxx 14127 xxxx xxx x 14157 xxx xxxx Bi14162 14187 xxxxxxxx xx 14219 x xx x x x x x x x xr (241) X (X) X 14251 xxx x r 14283 xxxx (x) x x x x x r 14315 xxxx xxxxxxx 327 x 85 A b C D E G M K Kc L W 143^5 XXXX X X 353 * 359 359 14375 xxx xx 14403 xxxx xx 14433 xxxxxxxxxxx 14465 xxxxxxxxxxx 14495 xxxx x I 14527 XXX XXXX X 14561 xxxxxxx xxr 589 14597 xxxx xx601 609 x X 613 B«613 14627 xxx 64^ 86 A b C D E C- M K Kc L V/ 14661 xxxxxxxxxxxr 14691 x xxxxxx 14721 xxxx 14751 X X X i 14783 xxx 14813 xxx 14843 xxx xxx x x 735 x x xxxxxxx 831 X X x (x) 861 14873 xxxxxxxxxxx 14905 xxxxxxxxxxx F9 923 x F9 14935 xxxx xx 950 87 A b C D E M K Kc L W 14969 XXXX X 973 x x X X F9 14999 xxxx xxx xxF9 15035 15061 x x 15091 x x 15117 x x 15147 x x 15183 x x 15211 15241 x x 029 xxx X 039 045 x XX xxxx X X X X XXXXXXXX X X X X X X X xxxx xxx XX xxx 235 XXX XX 88 A b C D E G M K Kc L J 15275 x x 15305 X X xxxxxxxx 287 X X 317 x 15337 x X 15367 x X 15397 x X 15425 X X 15459 15497 x x 15527 x X X xxxx X X X X X X XX X xxxxxxxx XX XXX xxxxxx 519 547 x (x) 553 x x 15557 xxxx x x (x) 569 581 89 A b C D E M K Kc L W 15593 xxx X X X X 15623 XXXXXXXXXXX 645 15653 XXX XXX XX 15687 xxxx xxxxxx 691 15717 157^7 x X X 15779 x x X 15793 15809 xxx 15839 xxx 769 761 X X X X X X x X x (X) X x x x B X X 90 A b C D E G M K KC L W 15869 xxx 15899 16029 16057 x I6O89 x 16129 x 16151 x I6158 x X X 15927 xxx 15959 xxx 15993 xxx X X X X X 001 (x) 105 X X X X xxxx X X X X xxxx x (X) x x xxxx X X X X 91 3.3. Evaluation of the initial systems 3.3.1. Frequency in the placing of initials Before we attempt to calculate the agreement of the manu scripts with each other on the basis of their initials, we should bear in mind that some scribes place initials more frequently than others and that this factor might lead to wrong conclusions. The first four listed manuscripts, not ably D and C, have paragraph markings at fairly regular intervals, here and there omitting one initial or inserting an additional one. But they never come close to those clus ters of initials such as occasionally shown by L and others (e.g., chart p. 69/70) or to long passages free of initials common to K and others (e.g., chart p. 65). To indicate the frequency of paragraph divisions in the manuscripts we will simply divide the number of verses by their number of paragraph signs. Thus, there is a new paragraph in C every 33 verses (14292 verses, 434 paragraph signs) D every 33 verses (15234 verses, 460 paragraph signs) b every 35 verses (14052 verses, 403 paragraph signs) A every 35 verses (16122 verses, 457 paragraph signs) L every 39 verses (15590 verses, 402 paragraph signs) W every 39 verses (16028 verses, 411 paragraph signs) G every 43 verses (15966 verses, 369 paragraph signs) M every 44 verses ( 9142 verses, 206 paragraph signs) E every 46 verses (16118 verses, 347 paragraph signs) 92 K°~ every 48 verses (15712 verses, 327 paragraph signs)bX c 6 2 K every 50 verses (15410 verses, 310 paragraph signs) Pfeiffer's text with its 16164 verses has a paragraph approx imately every 32 verses (501 paragraphs) and thus comes close to the average frequency in C and D. Although Pfeif fer' s critical remarks concerning the various manuscripts (Barlaam, p. 408) might suggest that he followed E more than any other source, his selection of paragraph divisions dis proves his own words. His reference in this matter is the practice of D, C and A (probably B as well, more so than b). It is only in a few cases that he abandons this procedure and follows the paragraph markings of other manuscripts or his own conception of what is appropriate for the narrative. Thus he begins a new paragraph on the second verse of a couplet in 2196, 2914, 8602 and 9220, and inserts a para graph at 3177 (with EG), 3993 (with E), 10419 (with EGK), 15035 (with bK) and 15793 (with EGKL). Pfeiffer was certainly well advised in following mainly A, D, and C. A comparison with the age of the Bar laam manuscripts (see pp. 19-21) shows that the younger ones tend to place their initials less frequently and, more 61 Kttpke's edition (K) consists of 16060 verses and 332 paragraphs. If these paragraphs correspond to initials in Ka is doubtful, however, and will be discussed further. 6 2 The figures concerning the number of verses are mostly approximated. Even with slight adjustments the over all picture would hardly change. 93 important, less regularly. We can assume that the older texts represent a more authentic system of initials which has undergone considerable changes and omissions at a younge stage. Consequently, the absolute number of initials in which a younger manuscript (Y) agrees with an older one (X) might not be very indicative of their actual relationship since Y might have omitted a great number of its source X. However, if we calculate the percentage that these common initials constitute in relation to the total number of ini tials contained in Y (as well as X), then the relationship would become more obvious. To clarify this further, let us look at the agreement c of K , a relatively young manuscript (fifteenth century), with D (thirteenth century). K has 284 initials in common with D. This constitutes only 61.5% of all the D-initials, a rather low percentage. But we have to keep in mind that K has only two thirds as many initials as D. This fact is accounted for by the second percentage figure which indi cates that these 2 84 common initials constitute 9 6% of all • • • • c c the initials existing in K , very few of K 's initials do not have a counterpart in D. This would suggest that K either derives from D or that they have a common source. For the sake of brevity we write down the agreement of K 's initials with those of D simply as Kc: 96% - D - 61.5% Furthermore, we can obviously compare two manuscripts only 94 in those passages extant and legible in both. Wherever one of them has a gap in its text transmission due to extrinsic causes (straight vertical lines in our chart) we exclude the corresponding initials of the other manuscript from the count, so the percentage figure will not change. On the other hand, when one manuscript has deliberately left out certain passages or inherited such cuts from its source (waved lines in our chart) whereas the other manuscript shows the full.text we include the latter's corresponding initials into the count; thus, the dissimilarity of the two manuscripts is reflected in the percentage figure as well. 3.3.2. Manuscripts A, b, C, D. As mentioned before, A. b, C and D show a high degree of congruence in their initials. The statistical figures indi cate this fact very clearly. A b C D 95.5%-D-90% 95.5%-D-92% 94.5%-D-94% 94% -C-94.5% 91.5%-b-90% 90% . -A-91.5% 88% -A-91.5% 90% -A-95.5% 91.5%-C-88% 92.5%-C-89% 89% -b-92.5% 92% -b-95.5% 77.5%-W-86.5% 79% -W-87.5% 76% -W-89.5% 77% -W-91.5% 62.5%-KC-92% 60% -KC-91% 61% KC-94.5% 61.5%-KC-96% 71% -L-80.5% 72% -L-83.5% 72.5%-L-90% 72% -L-87% (59% -Ka-81.5% 59% -Ka-82.5% 58 . 5%-Ka-8 3 . 5% 59% -Ka-86.5%) 61% -G-75.5% 62.5%-G-79% 59.5%-G-78% 61.5%-G-81% 95 A b C D 59% -M-75% 60.5%-M-81% 56% -M-81.5% 60.5%-M-82.5% 52.5%-E-69% 51% -E-70% 51.5%-E-71% 51.5%-E-72.5% The strong affinity that A, b, C and D show in their plac ing of initials sets them apart from the other manuscripts. Their next closest manuscripts W, L and KC have a consider ably lower percentage of agreement with any one of the four, and the last three manuscripts G, M and E even much less so. Does that mean that the first four manuscripts belong to one genealogical branch or even that they copied from one another? The percentage figures do not give a clear answer. In any case, there are no two manuscripts which have a noti ceably higher agreement with each other which would suggest that one is a direct offshoot of the other. C and D have the greatest congruence in this group but it is only slightly higher than that of the other combinations. In our search for a closer grouping within this group we should now look for individual cases of parallel ini tials. We shall disregard all the other manuscripts for that purpose and limit ourselves to those text passages common to A, b, C and D. Are there striking deviations in the placing of initials in two manuscripts against the other two? D has 3 in common with C (3603, 8975, 15717), 3 in common with A (3883, 4207, 11317) and one with M12517). C has one in common with b (4829) and two with A (1897, 9945). The greatest congruence exists between A and b which 96 in contrast to the others contain both Autorreden (11735 -11870 and 12247 - 12288), and it is only there that they have five initials in common against the other two manu scripts. Furthermore, is there evidence for a closer group ing of three manuscripts against one? A, b and C have common initials against D only in 5 cases (2649, 2693, 3511, 4705, 11581). A, b and D stand against C in 4 cases (691, 887, 2569, 8911), but since C omits a lengthy passage (9949 - 10047 and 10083 - 10251) there are an additional 9 cases of common initials in AbD. It is remarkable, though, that the grouping CD plus either A or b appear in many more instances: 23 times ACD without b (396, 401, 1191, 1619, 2161, 2429, 2914, 2947, 3969, 3991, 4043, 4457, 4565, 7287, 7503, 7667, 8631, 9277, 10717, 11991, 14375, 14627, 14691) and 24 times bCD without A (2195, 4767, 7473, 8033, " ? 8101, 8223, 8819, 9923, 10325, 11161, 11345, 11441, 11671, 12357, 12643, 12859, 13147, 13621, 13937, 14157, 14251, 14527, 15595, 15653). Based on this evidence, one would tentatively group CD and Ab together and. assume that C and D have the most au thentic initials but that they are not dependent on each other. Whenever a C-initial contradicts a D-initial we would check with the other branch in order to assess which paragraph sign can be trusted. C and D, moreover, share the omission of the author's excurses. On the other hand, each of the two has"major gaps in its text which are not 97 shared by the other one, C as mentioned above and D from 4213 - 4396 (the scribe probably skipped a leaf of his source). A and b, although in major respects different from CD, cannot be directly dependent on each other, either, Otherwise, it would be surprising that they have no common initials differing from the other two manuscripts (except for those in the author's excurse). Pfeiffer had noticed that b was alternately written by two scribes. He suggests that the first scribe was following B and the second one another source closer to A (Barlaam, p. 408). According to Pfeiffer, manuscript b was written as follows: part 1_ until page 38v (7332), first scribe (between page 37 and 38 a whole quire of 12 leaves is missing) part 2_ page 39 to 64v (10765), second scribe part 3 page 65 to 77V (12944), first scribe part 4 page 78 to 90v (14893), second scribe part 5_ page 91 to 97v (16039), first scribe (final leaf is missing). F. Sohns adopts this theory in principle but modifies it somewhat. According to him, the second source "discovered" by the second scribe was A. A and B together were used as source by both scribes between 7941 and 14608, for the rest, both scribes again copied only from B, "wahrscheinlich aus Lassigkeit."^3 63 F. Sohns, Das Handschriftenverhaltnis, p. 37-38. Stthns claims erroneously that verses 7332 to 7941 are mis sing, instead of 5239 to 7167. Sflhns does not give any reasons for choosing 14608 as a turning point. 98 Pfeiffer's assumption that each of the two scribes fol lowed a different source is highly unlikely and should be rejected on the basis of extrinsic evidence alone. It would make sense only if each scribe within a scriptorium were in charge of one or several quires which were to be written simultaneously in order to speed up the copying process. In manuscript b, however, there are only•two cases (at the transition from part one to two and three to four) where a change in handwriting coincides with a new quire. The sign atures (Kustoden) are perfectly legible on the lower margin of the last page in each quire, so we can tell that a new quire begins with page 13, 25, 37, 39 (the missing leaves, probably more than 12, are not counted in the modern pagi nation; 51, 66, 78 and 90. The change in handwriting is sometimes difficult to recognize but, judging by our micro film reading, Pfeiffer's division is correct. The first scribe writes his cursiva more evenly and within his pre drawn sets of lines, whereas the second one usually exceeds them or completely forgets about them and generally shows more irregularities•in his writing and in the number of lines that he fits into one column (from 29 lines to 4-0 lines). This latter circumstance would make it very diffi cult for a scribe to keep track of the number of written verses and calculate the remaining space so that his part would perfectly match the beginning of the next quire. At one instance, however, at the transition from part 99 one to two, a peculiar gap in the text occurs which could very well be interpreted as "mismatching" of two simultan eously written quires. On page 38va the text continues normally until 7 27 7 on the bottom line but then, with the first line of the right column, the scribe abruptly jumps to 7301 without even completing the rhyme. Did he at this point calculate the remaining space and attempt to fit his verse material into it by conveniently cutting out the nec essary number of verses? If so, he slightly miscalculated it and ran out of his shortened text (7332) two lines above the bottom. Instead of leaving some blank space the scribe filled it with two random verses that he had skipped before: 7293 - 94. Such a mutilation of the text seems to be deli berate, different from simple "human errors," e.g., the omission of a couplet at the beginning of a new page or quire (such as 11037 - 38 at the beginning of page 66r in b where the same scribe continues). While it is likely that in this one mentioned case the second scribe of b began his copying work before the first scribe had finished his part, it does not mean at all that they necessarily followed two different sources. The source-manuscript may have been chopped up into several parts given to the two scribes. If the text handed down by the second scribe was taken from A or a closely related manu script and the rest from B, there should be a noticeable difference in the agreement of initials. We calculated the 100 agreement of b with the other manuscripts for each of these five parts individually and could not ascertain a major change in its relation to A or any other manuscript other than those fluctuations due to the different length of the compared parts. It is unfortunate that we no longer have the testimony of B, all that remains are the notes in Pfeiffer's apparatus. Thus we shall never know with abso lute certainty if b was in fact copied from B by the first scribe; but there is room for doubt. Among the few initials of B that Pfeiffer has passed on to us there are several that do not conform to the pattern of b but agree with other manuscripts (E, G or Ka) or stand alone, e.g., at 7237, 12111 and 12573 in those passages written by scribe I. B and b do have a close affinity (their parallel initials at V2933, 4443 and 14613 being one of many indications), but ? it is not likely that b is a direct copy from B, not even in parts. A later look at common and divergent text omissions and some textual variants should confirm this assertion. It is more likely that B and b were dependent on a common source *Bb, b probably through intermediaries. *Bb must have been related to A as the high percentage of common ini tials and the'preservation of the author's excurses suggest. Thus we assume that "Bb and A on the one side and CD on the other side represent two old branches of the Barlaam text tradition. 101 3.3.3. Manuscripts L and W The Bonn and -the Vienna codexes L and W both date from the fourteenth century and show a similar frequency in their paragraph divisions, every 39 verses on the average. Other than that, they seem to have little in common, being of different provenance (L from Middle-Franconian territory, W from the famous library at Ambras / Tyrol and of different intrinsic value. Since the oldest manuscripts C and D and, by extension, A and b are relatively close together in their paragraph agreement it is difficult, if not impossible, to assign any younger text to one of them just by looking at the overall statistical figures. Going through the chart for AbCD one discovers a remarkable likeness in the correspondance of the younger manuscripts with any of the four above mentioned. Judging by the statistics alone, W would tend slightly more to D: 91.5% of its initials correspond with parallel ones in D (89.5% with C, 87.5% with b and 86.5% with A)\ L, on the other hand, has. a higher congruence figure with C: 90% (87% with D, 83.5% with b, 80.5% with A). The agreement with the other manuscripts is much lower, for W between 6 8% with GM and 57% with E and for L between 65% for GM and 5 3% for E, therefore, these manuscripts have no importance for L and W in this regard. A comparison of W and L makes it evident that the two 102 have a high degree of affinity. This is especially notice able for L where 84% of its initials are congruent with those in W (L: 84% - W - 82%), thereby nearly equalling the agreement figures with C and D and surpassing those of A and b. L has a somewhat erratic method of setting its initials, frequently they are not concurrent with any other manuscript. However, there are a few cases in which L and W alone show common paragraph markings: 2175, 2567, 5703, 8707, 9379, and together with manuscripts other than AbCD (mainly E) at 4325, 5009, 7237, 8417, 9647. Surprisingly enough, these parallels occur only in the first half of the text, whereas later, L follows more the practice of C. The two parallel initials of C and L in 8869 and 8895 (divergent from the AbCD pattern) seem to mark a turning-point. Indeed, the percentage figures corrobor ate our impression. Up to 8850, L shows an agreement of 92.5% in its initial system with W and 88.5% with C. After that the agreement with C increases up to 91.5%, while that with W drops sharply to 76%. This is a very unusual change which does not affect L's relationship with any other manu script. Moreover, the chart shows parallel text omissions in C as well as in L in the second half which may partly account for the strong statistical concurrence of the two. Thus, we are led to assume that the scribe of L after using a source closely related to W (or even W itself) for more than the first half of his copy changed to another 103 source more closely related to C. It is also possible that he followed one single manuscript based on the two differ ent sources. We believe that he could not have worked con currently from two different sources, relying at first more on the *W manuscript and later more on *C. His many scribal errors and text distortions of which a glance at Sfihns's apparatus gives a sufficient impression make him appear as rather inept. The best one can say of him is that he wrote a very pleasant textura script. The position of W in the text tradition cannot be ascertained by consulting the initial system alone. It could derive from D or a related manuscript since 9 out of 10 initials might have been taken from that source. How ever, unlike D and most other manuscripts (except ABbE), W has preserved the text of one of the author's two digres sions (the "Schimpfrede", 12247 - 289). Yet W does agree with D and most other manuscripts (again except AbBE) in the omission of the first digression (the "Damenpreis", 11735 - 870). Does W represent an ancient branch differ ent from the ones that we have tentatively established so far, or is it dependent through various stages on a common source with D, a source that must have presented the full text? We shall have to leave this question open at this point and refer it to a later study of textual variants and text omissions. 104 3.3.4. Manuscripts G, M, and E Like W, the Gotha codex G and the large London fragment M have been disregarded by Pfeiffer's text edition and by any detailed manuscript study so far. Their inclusion into a new edition seems timely since they offer interesting per-64 spectives. Although M contains only slightly more than half of the entire verse material of Barlaam, its three sec tions cover representative parts of the beginning, the mid dle and the end of the narrative. For this reason we have incorporated it into our statistics. Of all manuscripts, G and M show by far the highest agreement in their paragraph division. There is not one initial in M that does not have its counterpart in G (10 0% congruence), and the common initials with M account for 9 8% of G's number of comparable initials. This nearly total congruence would suggest that one text might have been copied directly from the other, and the question arises as to which one was the source. Both manuscripts date from approximately the same period, and we therefore have to look for internal evidence. G has a total of four initials which M does not share (#4043, 6031, 7123, 14001) but which 64 See Worstbrock's remark in AfdA 77 (1966), 114: "Wenn schon eine Ausstattung des Neudrucks durch zusatzliche Materialien zur Textkritik erwllnscht war, hatten zuerst die Gothaer Hs. und die grossen Londoner Fragmente die Ehre verdient." 105 are in agreement with most of the other manuscripts. There fore it is appropriate to assume that M omitted these para graphs rather than to think that G inserted them on its own. We surmise that M is a copy of G; however, at this point, we cannot rule out the possibility that G and M are two very faithful copies of one manuscript (*G). The Munich codex E offers the youngest complete Barlaam version that we know. Written in 1459, its Bavarian dia lectal forms as well as its Bastarda script give it an appearance very different from that of the much older manu scripts G and M of Middle German provenance. The scribe of E seems to take more liberty in structuring his work: at numerous spots there are paragraph markings different from all the other manuscripts. But in many more cases E shows a basic agreement with the practice of G and M as demon strated by some statistical figures: 84% - G - 79.5% ; 80.5% - M - 79% ; 72% - CD - 51% ; 67% - W - 57%. The affinity of G, M and E is particularly recognizable in our chart since the three differ in many instances from the AbCD pattern. G or rather *G has altered the initial sys tem which we believed to be tantamount to the "authentic" one and has introduced or omitted a great number of para graph divisions. The 5';G-group does not show a different trend of setting its paragraph marks from the very begin ning, but a change takes place only after the first quarter 106 of the text. Before 3000 there are only very few cases of diverging initials in G and E (1815, 2343, 2937), between 3000 and 4000 the number increases and after that it occurs very frequently. This gradual break with the tradition makes it impossible to tell by a mere study of the initial agreement which (if any) of the older manuscripts known to us might have been used as source. In the first 4000 ver ses the congruence with ACD is about equally high, around 90% for G, somewhat less 'for E. After that the agreement figure drops sharply, giving G an overall agreement factor of 81% - D - 61.5%; 78% - C - 59%; 78.5% - b - 62%; 75.5% - A - 61%; 76% - W - 68%; 71% - L - 65%; 56.5% Kc - 6 7%. (The figures for E see above".) The manuscripts of the *G-group do not share any of those initials of D, C or A which deviate from the "main stream" (the consensus of DC and Ab), thus there is no indication here of a closer relationship to the one rather than the other. Among the paragraphs in B given in Pfeif fer's apparatus we find a few coinciding with one or all manuscripts of the *G-group: 4443, 5631, 5825, 5973, 7237, 10373, 10501, 12573. This would suggest either that the scribe of B used a manuscript belonging to the *G-group besides the one he followed mainly (*Bb, see p. 100), or even that B originated the deliberate modification of the old initial system to some extent. At this stage, we can not be certain if the "G-group derives from B or any other 107 manuscript or if it represents an independent, original branch of text transmission. 3.3.5. The K - manuscripts (Ka, Kb, KC) As mentioned above (p. 10-12, 20-21), the only one of the three K-manuscripts still available for scrutiny is the c c Berlin codex K of the fifteenth century. K has relatively few initials but the ones preserved are well in agreement with the oldest manuscripts, foremost with D, as outlined c on p. 92. Here are the percentage figures showing K 1 s agreement: 96% - D - 61.5%; 94% - C - 61%; 92% - A - 62.5%; 90% - b - 59%; 81% - W - 61.5%; 75.5% - L - 58%; 67% - Ka - 62%; 66% - G - 56%; 59.5% - E - 53.5%. This high percentage of congruence with D as well as par-c allel text omissions make it very likely that K stands in the tradition of D. There is no noticeable affinity c .a between K and the "G-group, nor with K . If we assume that the paragraphs in Kopke's edition (K) faithfully correspond to paragraph divisions in the first Kflnigsberg codex Ka, our statistics do not give a clear picture of its relationship to other manuscripts. Its agreement with D is, relatively speaking, the highest, but not at all comparable to the closeness of K and D. In contrast to K , K has a number of initials in common with 108 GM and also WL but not enough to warrant a closer relation ship . 86.5% - D - 59%; 83.5% - C - 58.5%; 81.5% - A - 59%; 82.5% - b - 59%; 78% - W 63%; 73.5% - L - 61%; 73% - G - 66%; 73.5% - M - 66.5%; 63% - E - 59.5%. While it is not impossible that Ka could have shown a rather different initial system, due either to the intention or carelessness of its scribe or to a contaminated source, we feel justified in doubting that Kttpke's paragraphs reflect the initials in K . Neither he nor his collaborator Lach mann ever claimed they did; they do not mention this aspect at all. KSpke, who consulted Kb and K° in addition to Ka, . . c certainly did not use K for his paragraphs, the low agree ment figure speaks against it. Furthermore, in one instance where two leaves are missing in K , KSpke filled in the text according to K° after comparing it with Kb.^ However, KSpke does not follow the initials of K in this particular passage (1910-2132): he leaves out four of K 's nine ini tials, setting paragraphs only at 1815, 1837, 1937, 1981, and 2131. These long paragraph units of up to 150 verses correspond nearly perfectly to the divisions in E and G with their initials at 1815, 1837, 1863, 1937, 1981 and 2131. For an editorial whim, this seems an unlikely coincidence. Probably KSpke followed the paragraphs of finding them KSpke, p. 408: "Der Text ist nach C [i.e., K°] mit Vergleichung von B [i.e.; K*3] gegeben, und die Schreibung der von A [i.e.) Ka] gleich gemacht." 109 more suitable for the narrative. Is there any further evidence that Kb might have been associated with the *G-group? Kb seems to have been a codex recentior in which text omissions and alterations abound. Since we have no information on its initial system, we have to look for textual variants provided by Kflpke's apparatus in order to support our theory. The strongest evidence in our favour is that Kb and manuscripts of the *G-group share in three cases the omission of a couplet and in one instance the addition of a couplet. Missing are 3721-22 (with GE, section missing in M), 6157-58 (with GEM), and 13631-32 (with M; E has changed the order of verses, G has the regu lar text with CD and the other manuscripts). More impor-. tant even is the additional couplet in Kb and GE after 94-00 (section missing in M) in which Kb and E have a blatant mis reading in common. In the disputation between Nachor as in voluntary spokesman for the Christians and the representa tives of other religions, the Chaldean priests end their speech by saying to Nachor: "Nobody else but deceitful \ people would choose your Christian life-style, people such \ ^WJU~v^ as you and your kind. What could be more guileful?" 9399 an diz leben sich ergebent 9400 die mit valscher trtlge lebent GEKb: als du und die genozen din G: waz mochte gouclicher sin ("more guileful") EKb: waz mochte gotlich' sin ("more godly") 110 Thus, we may conclude that Ku must have belonged to the *G-group, probably closer to E or M than to G. Going one step further now, we suggest that in many cases where the KSpke edition has paragraphs in agreement with those of the "G-manuscripts, they are most likely to have been taken over from . This means that our approach to establish the position of Ka on the basis of its paragraph division is not viable. We have to study the text variants"of KSpke1s edition and hope that KSpke has shown more accuracy in transcribing the wording than he has done in preserving the textual divisions of K . 3.4. Spurious initials; textual variants at the beginning of a paragraph So far we have made it a principle to study only the posi tion of paragraph signs and disregard all textual readings (with the exception of K^) in order to achieve a first grouping of the Barlaam manuscripts. When we now look at textual variants at the very beginning of a paragraph and see if they support or contradict our provisional grouping we seem to fall back on the traditional textual method. And yet, there is a difference due to the fact that mis-readings or rather different readings can be caused by the scribe himself or by the rubricator. It is sometimes clear who is to blame: in cases where there are blatantly nonsen-Ill sical forms of distortions of names, the rubricator, most likely unfamiliar with the text, executed a wrong initial (Initialenfehler). We found several examples for this type of mistake, such as "Alofernus" instead of Olofernus (in b, 2387), "Doboam" instead of Roboam (in b, 2311), "Brachis" instead of Arachis (in A, 7759)66 or "Polus" instead of Eolus (in E, 10229). It is obvious that such striking mis takes would not likely be taken over into a dependent copy and therefore these cases hardly ever shed light on the relationship of manuscripts, In some instances, however, the change of an initial can modify the sense of the phrase without making it mean ingless. It must be remembered that, because of their ornamentation, some initials look very much alike although they are different capital letters. Thus the change could have been caused by the misreading of the scribe or by the carelessness of the rubricator. In such cases it is impro per to label them as "false initials" (Initiale rifehler) since the original'reading is undetermined. The beginning of the paragraph at 19 81 provides a good example. In the preceding passage, Barlaam praises the Christian God and the trinity and contrasts them with the pagan idols. The new section continues the praise. 66 There is an occasional confusion between the names of the two counsellors "Barachias" and "Arachis" in other manuscripts, especially in L. 112 c • Pfeiffer reads with manuscript K : "Einen got vxl lobesa-men/geloube ich, einen in drin namen ..." The Bonn codex L shows the same variant without an initial. The other manuscripts read "Disen got" : C, "Dinen got" : DW, "Minen got" : AbGp, "Ainen got" : E. The only version that does not make sense is that of DW, since Josaphat is still a heathen at that point of the dialogue. The most satis fying form is indeed "Einen got", underlining the contrast between the one almighty God and the many powerless idols; c however, only the more recent manuscripts E, K and L sup port this version. It seems as if the initial in,D had actually been altered, but it is impossible to discover which one might have been the original capital letter. This gives room for speculation. W either followed D—our statistical study showed that W could probably depend on D or a common source—or the rubricator of W could have made the same mistake independently from D. C could have found this form in its source and have tried to correct it by altering the "Dinen" into "Disen." On the other hand, "Disen" might represent the original reading distorted into "Dinen" by DW. The third version "Minen got" is perfectly possible and could be the original reading also. The four manuscripts presenting it are related in pairs, fragment p belonging to the *G-group and b coming forth from a source close to A (see above). Was there any interaction between these two groups or do we have independent attempts to 113 correct an obviously faulty variant "Dinen?" These considerations aptly demonstrate the difficul ties in evaluating such divergent order to deter mine which one might be the authentic version, if we assume there was only one. The editor, in such a case, will most likely resort to his taste, just as Bedier had suggested, unless the Latin text gives strong support for one theory. Here, in this instance, the Latin version, as presented by Migne's PatroTogia Tatina, t.73, p. 464-, fully justifies Pfeiffer's choice: "Neque enim ex eorum numero sum qui multos hos et petulantes deos colunt . . . verum unum 6 7 Deum agnosco et confiteor qui in tribus personis . . ." Fortunately, such cases are rare. One single common vari ant of that kind cannot be regarded as evidence for a clo ser relationship between manuscripts: it needs a more fre quent occurrence or cases where there are altogether differ ent words or phrases at the beginning of a new paragraph which could not be ascribed to an inattentive rubricator. This third type, the proper textual variant, is probably the final result of a copying process in which errors of rubrication were tentatively amended at a later stage. Thus they stand as a different version of their own and are 6 7 For a more detailed comparison Migne's text—a later translation from the Greek—would not suffice and the preserved Latin manuscripts of the "vulgate" version of Barlaam should be consulted. Sonet lists 62 such manu scripts (Le Roman de BuJ, I, 74-88); whether Rudolf's source is among them is unknown. 114 usually conducive to recognizing closer groupings of manu scripts . The comparison of the various paragraph beginnings in the Barlaam-manuscripts yields no surprises; it confirms mainly two clear groupings which were shown already by our statistical survey: a) the *G-group (GEM) and b) the close relationship between W and L. 6 8 a) 1455 Pf. Er hate daz wol ervarn GEM Barlaam hatte wol ervarn 4325 Pf. Noch wil ich dir ktinden; Doch . . . L GEM Aber wil ich . . . (initials in GEMLW), text missing in D 712 3 Pf. Nach dirre lere wart getan GEM Nu diz wart alsus getan (no initial in M) 7287 ACKK° Noch la dich des durch got gezemen GEM Doch la . . .; So la . . . D BWL Nu la . . . (no paragraph); no text in b 750 3 Pf. Als er des boten rede vernam GM Da Zardan die botschaft vernam; Zardan . . . ] 14315 Pf. Alsus-nahet im der tot GEM Hiemit nahete im'. . . b) 165 Pf. Hie vor in der gnaden zit WL Die in der gnaden zit 6 8 Pfeiffer's text (Pf.) represents the variants of all the other unmentioned manuscripts. The spelling is unified. 115 431 Pf. Do vuogte sich . . . WL Nu fugete ez sich . . . 3541 Pf. Ouch hat uns bispel gegeben; Doch hat . . . b WL Nu hat uns . . .; Ich han uns geleuchniz . . . E 39 91 Pf. Dar an raerke minen rat WL Heran merke . . . 84 7 7 GEMAb Noch waer min rat also getan; Doch . . . D WL Nu wer . . .; Joch . . . K; Ouch . . . K In the second half of the text, there are no further com mon variants in the paragraph beginnings of W and L. This would confirm our findings in the study of initials, namely that L changes its source after approximately the first 8850 verses and follows C or a common source. Since C is usually reliable in its initials and paragraph beginnings, L and C do not show any striking variants from thereon. At 160 29, the scribe of L must have misread the line "In kriechisch man diz maere schreip" and changed it into "ZQ krichin man." An even greater misinterpretation occurs in W: "Ein kriechschen man dis schreip" and in E: "Ain krieche ditz maere schraip." However there is no further evidence for the above mentioned possibility that W and D might belong to one group. On the other hand, the theory that K and D are close ly related to one another is supported by the paragraph at 1009 where D, K° and K read "Do was bi im . . ." instead 116 of "Nu was . . ."as all the others do. Two further examples are 8477 as mentioned above (D, K and K have slightly dif ferent forms from the rest, but not quite congruous with each other), and 10359 where D and K read "An alsus getanes leben" instead of "In alsus . . ." like the other manu-c c a scripts including K . Thus, it seems that both K and K (which could not be assessed on the basis of its initials) are derived from a manuscript very close to D, but probably not D itself. D shows a few particular misreadings which are no found in K or K: 10 39 "Der sprach" instead of "Er sprach"; 7287 "So la dich des durch got gezemen" instead of "Noch la . . ." in K and Kc, "Doch" in GEM and "Nu" in BWL; or 9129 "Sit vride wart gevestent dort" (shared by b) instead of "Mit vride wart ..." The relationship of the two major manuscripts C and D is as yet the least clarified. The study of their initials showed great likeness, however the comparison of their para graph beginnings did not provide us with any further clues. When we now proceed to analyse the Barlaam-fragments in their relationship to the complete manuscripts, we hope to gain more material whereby we might be able to assess more accurately their rank within the framework of the text tra dition . 117 4. The smaller fragments 4.1. Introductory remarks The following chapter takes into account all available smaller Barlaam fragments and attempts to assess their im portance for a reliable new text edition. A considerable amount of confusion is noticeable in this matter. As men tioned before, Pfeiffer used only very few fragments for his edition and Sflhns's thesis added information on six fur ther fragments, albeit some of it from second hand. Since then, a number of further Barlaam fragments have been dis covered and some have disappeared; even since the last and so far only list of fragments after Pfeiffer and S6hns, that of Worstbrock in 1966 (see chapter 1.2), three further fragments p, q, and r have been made known. Worstbrock justly corrects the errors contained in H. Rupp's Nachwort to the reprinted Pfeiffer edition, but his own indications are not completely free of mistakes. Worstbrock does not update Pfeiffer's and Sflhns's information and he does not deal with the intrinsic value of the fragments he lists. Various figures for the overall number of manuscripts and fragments have been suggested; Worstbrock claims: "In der Tat lassen sich mindestens 45 Textzeugen nachweisen." For our purpose, such figures are irrelevant. What we are concerned about is to learn which fragments have preserved an independent version of the Barlaam text and could be con-118 suited for a re-edition. That means we have to find out first which fragments are "matched," are part of the same, otherwise lost manuscript and have to be considered as one text witness. The next step is to assess how the fragments relate to each other, to the major manuscripts and manu script groupings and which of them deserves special atten tion in view of a text reconstruction. During this pro cess we gather material concerning not only the fragment under study but also having a bearing on other manuscripts. This evidence is needed for a critical look back upon the tentative results gained in chapter three and hence for a final evaluation of the initial method applied there. An attempt to group manuscripts on the basis of tex tual variants is constantly confronted with the problem of selection and credibility and thus open to the reproach of subjectivity. This is sometimes unavoidable, a detailed justification for the citation of common variants or the omission of others cannot be expected in this framework. For convenience's sake,, the slightly different spelling of various manuscripts showing parallel readings has been dis regarded, abbreviations are spelled out and obsolete letters (the two different forms of -s-, -r-, and -z-) or diacriti cal marks modernized or given up. Whenever Pfeiffer's text represents the consensus of all manuscripts the sigle Pf can stand alone in their place. 119 4.2. Matching fragments 4.2.1. Fragments d and q The Zurich fragment d is among the few that Pfeiffer used for his edition. He ranked it above the others ("Von den Bruchstucken wusste ich keines auszuzeichnen ausser d, das gut und alt ist." Bar1aam, p. 409), without for that rea son always following its reading. These two vellum leaves of the thirteenth century contain 242 verses (10967 - 11084 and 11331 - 11454), neatly written in two columns of 29 to 31 lines with initials alternating in red and blue. The seven initials in d agree with those in CDb with two excep tions: ddoes not have a paragraph at 11345 as all the others do and at 1140 7 the rubricator set the marginal ini tial I one line too low: ICh bin durch daz her zuo dir komin / Ich han uon dir ain tail uirnomin. The scribe had clearly meant it to be drawn in front of the first ICh as the capital follow-up letter indicates. Therefore, the rubricator's slip could have easily been recognized and corrected by scribes copying from this manuscript. The text in d is reliable, indeed, there are only a few cases in which d does not show a meaningful variant, such as 11009: Do gibot der kunic sa / Daz sich schiede des k'nappin strit, instead of "des' kamphes strit," or at 11033 the singular form instead of the correct plural: Sinen maister nam er do / Die schiedin dan mit im unvro. 120 Pfeiffer claims erroneously in his apparatus that verses 10972 (275,32) and 11030 (277,10) are missing; however, there are no text omissions at all in d. SShns, who dealt with the position of d in relation to the other manuscripts, associated it correctly with the DKaKc-group (obviously giving up his initial claim that cL C this Reihe consists of A as well as DK K ). He supported his statement by four textual variants, the last of which is invalid (11408 "von dir ein teil" is read not only in d and DKaKc, but also CGWLb. Pfeiffer follows A and E which read "ein teil von dir."). It seems adequate, therefore, to offer our own brief selection of common readings in order to establish the affinity of d with the DK K -group. 109 76 Pf: (ein riche daz niemer zergat/) unde an ende vreude hat d: daz anegande vroude hat; und anegende . . . DKaKC 10992 Pf: got lobte solher guete Kcd: got lobter siner glieti; got lobte siner guete DKaE 11001 Pf: Swenne er gen im kerte Ld: Swenne er sich gein im k.; S. e. gen im sich kerte DKaK° 11336 Pf: so disiu veste wart erkant; disiu hoch-gezit W DKaKcd: so diu foste wart irkant 121 11369 DKaKCbd Pf DKaKCd 11371 C Pf DKaKCd 11348 Pf: daz der kttnec belibe alvro daz der kunic belibe vro und waz Nachor in haete getan; u. w. in N. . . . EGb und was N. im hette getan und was N. hete getan ir sic ir saelde ir ere ir saelde ir sige ir ere In conclusion, one could agree with SShns' statement that d as well as DKaKC originated from a common source, although other possibilities could be suggested. It is nota-ble that d shares a common variant alone with K in two instances: . . . daz tuon ouch ich . . , daz tuon ich (und ander hfivesche liute guot//) die den goten machen I ... I die hohgezit 11028 Pf dKC 11431 Pf d: die dem liute machin; den luten machen K It seems that K is- probably even more closely related to d than D and Ka, although it could not have descended from it in a direct line. Their different age and dialect make a a more definitive judgment on the basis of these few verses impossible. In 1972, attention was drawn to two formerly unknown Barlaam fragments kept in the federal archives (Staatsarchiv) 122 Schaffhausen/Switzerland. P. Ochsenbein described them minutely, included two facsimile samples, and attempted to compare their version to that of the other manuscripts (see p. 23, footnote 38). We concern ourselves here with q, the first one of these two fragments, consisting of two partly damaged vellum leaves which contain most of the text between 6017 and 6137 and on the other leaf 6 3 85 to 6 415 and 6475 to 6504, alto gether about 180 verses (some of them cut in half). Con trary to Ochsenbein's assumption that q is an independent witness of the Barlaam text tradition, we could establish that q actually is a part of the same otherwise lost manu-69 script to which the Zurich fragment d belonged. A compar ison of the facsimile sample of q with a photocopy of d proved that the arrangement of the text, the measurements of the columns, the handwriting and the execution of initials are identical. Furthermore, both fragments show the same alemannic dialect features (". . . im ttstlichen hochaleman-nischen Raum entstanden," Ochsenbein, p. 323) and a similar relation to other manuscripts. P. Ochsenbein's statement that a reliable grouping (Eiriordnung) of this fragment is not possible cannot be upheld. His scepticism towards the rigid, yet questionable 69 . . ... Dr. Ochsenbein kindly confirmed our findings in a letter of May 14, 1973. 123 stemma of SfcVhns is commendable, however in this particular case, especially after matching q with d and thus having a larger basis of comparison, we can positively conclude that q belongs to the DKaKC group. The few initials in q are not clearly indicative of this affiliation, but the follow ing three common variants should suffice in support of our theory. 6051 ACLW: DKaKCq: do wart er guotes richer der wart guotis richir GMn: sus wart er g. r.; also wart er . . . E 6112 Pf: dich sihe so wol gehorsam sin DKaq: dich sihe so gar g. s.; dich sihe . . . KC und gar 6125 Pf: sist in siner wunne groz DKaKcq: sist in sinen wunen groz Ochsenbein mentions two striking parallel readings in q and C which in his view make all attempts at grouping q futile. In 6091 C and q use the verb dulten instead of liden, and in 6116 they both avoid the repetition of got through its relative pronoun: 11. . . ze gote willeclichen muot / daz er din ende mache guot." We hold that this latter example could very well be an independent, spontaneous smoothing of style in both man uscripts. Such alterations occur here and there, the oppo site appears e.g. in 6 04-5 where only q has a repeated noun der sweher within two verses and all other manuscripts use 124 the personal pronoun instead. As to the first example, we noticed that the expression arbeit dulten does not occur anywhere else in Barlaam, arbeit liden, however, is used several times in all manuscripts (5318, 8221, 12447). The author uses both terms in very similar phrases, dulden appears in rhymed position at 3486 and 15248 and seems to be treated as synonym to liden in the sense of "to suffer, 70 to bear." The dialectographical study by G. de Smet shows that the two words co-existed in German dialects the thirteenth century and that the preference of the one over the other was fluctuating for a while until finally liden became predominant. We found another example of this fluctuation in A and h at 12035, where muoste dulten rat is replaced by muoste liden rat. Thus, we suggest that at 6091 C and q could have substituted liden by dulten inde pendently from one another due to the scribe's dialectal preference and in accordance with the author's usage in other passages. The affinity of dq with the DK K group is not called in question by these two parallels between q and C. Gilbert de Smet, "Die Ausdrilcke fur leiden im Alt deutschen,"' Wirkendes Wort 5 (1954), 69-79": 1.2 5 4.2.2. Fragments e and 1 At a perfunctory glance it seems unlikely that e and 1 should be part of one otherwise lost manuscript. The two leaves of the Nuremberg fragment 1 contain the text in a continuously written form whereas in the Munich fragment e one verse is assigned to each line. Furthermore, the indi cations as to age and size of the fragments differ some what: e is dated from "Middle of the thirteenth century" to "fourteenth century" (see chapter 2.2.), and it is des cribed as "Grossoktav" as well as "small quarto;" 1, on the other hand, is written at the end of the thirteenth or beginning of the fourteenth century and is in octavo, 71 according to F. Pfeiffer's description. For his Barlaam edition Pfeiffer used fragment e only in form of a tran script, fragment 1 was sent to him after the appearance of his edition. So Pfeiffer never compared both fragments in the original or commented on their possible connection. The same holds true for Sfihns who does not go substantially beyond Pfeiffer. The two fragments were used for bookbinding and so they are both cut down to different sizes as befitting the purpose. However, in both e and 1 the carefully drawn set of lines for the two columns of each page has a height of 71 F. Pfeiffer, "Bruchstdck aus Barlaam und Josaphat von Rudolf von Ems," Anzeiger fur Kunde der deutschen Vorzeit (1854), 108-109. 126 14.5 cm, a width of 4.4 cm, the free space in the middle between the columns is of 0.6 cm and all lines are 0.45 cm 72 apart. Both fragments show the same handwriting with some characteristic features such as the capital S, D and N and have the same abbreviations. The Alemannic dialectal forms are mainly the same as much as one can tell from such a short sample (e has altogether only 138 verses, moreover both scribes vary in their spelling). It is certainly unu sual to find a manuscript where the scribe starts out writing his text continuously and later changes to writing in verse lines, but it does not seem to be impossible. The small format of the vellum makes it difficult for a scribe to fit a whole verse into one line of each column. In frag ment e> which has preserved only the major part of the epi logue, a special effort seems to have been made, and yet in' one instance a verse (16 057) has to be spread over two lines as well. Possibly only the epilogue was written in verses to set it off from the narrative part of the work, but there is no proof for such an assumption. The Nuremberg fragment 1 with its text between 436 8 • c and 4612 agrees m five of its six initials with K. and A (D having a text omission until 4396) and diverges to a stronger degree from those in other manuscripts. Pfeiffer had noticed already that the text of 1 resembles mostly that of D. Sflhns took this theory up and supported it by four 7 2 . . All measurements are taken from the microfilm copy. 127 common variants, one of them incorrect: 45 85 mit vreuden ir slaf inphiengen (instead of Pf: mit vreuden slaf . . .) can be found in GEM as well as DKaKc. He overstrains the argument however by suggesting that DKaKC might have been directly copied from 1; the opposite case ("oder umgekehrt . . .," Sohns, p. 41) does not make any sense at all since KaKC are younger than 1 by at least a hundred years. - The following examples should provide sufficient evidence for the close affinity between 1 and DKaKc: 4465 4556 4560 Pf DKaKGl Pf DKaKGl Pf 4585 DKaK°l Pf DKC1 4595 W Pf DKC1 ir lip ir muot wurden bekleit ir lip ir leben wart becleit und mttgen leider doch niht komen omission of leider mit ir'gezierde trtigelich mit ir geheizen trugelich Swer sich uf si slafen leit Swer sich uf slafen leit Swer sich si s. 1.; Swer sich slafen leit L den lat si ligen in der not den lat ligen in der not In these last two cases, KSpke's text agrees with Pfeiffer's and he does not give any information in his apparatus of a different reading in K . It is possible, though, that KSpke tacitly corrected K 's version in his edition. 449 2, only 1 and D share a common reading: Pf: die dirre welte minne bot 128 ID: die dirre wunne bot; die dirre welte wunne bot K K A. These variants would support the view that there might have been a common source for K K on the one hand (possibly-connected with dq) and another one for ID. In any case, they all belong to one branch definitely different from the other manuscripts. A does not share any of the other common variants of this group in the above quoted passage and can therefore not be counted. In the paragraph before 4 4-9 2, both expressions "die welt minnen" (4445) and "der welte wunne" (4451) occur: thus, the agreement of A and KaKc in 449 2 could be accidental or due to contamination. Going back to Sfihns' suggestion that D and 1 might have been copied from one another, the following counter-argu ments must be raised: D has a text omission between 4213 and 4396 in which 1 seems to have contained the text (1 begins at 4368). On the other hand, D has not likely been copied from 1. The gap in D which leaves out the parables of the Prodigal Son and the Good Shepherd and jumps from the middle of Barlaam's speech into Josaphat's response must have been caused by accident. The most likely explanation is that the scribe of D accidentally turned two pages. The missing amount of 184 verses would indicate that the source of D was written in two columns of 46 verses, which is obviously not the case with 1 (its verse 4396 being in the middle of a page). Once the position of 1 has been established we can look 129 at the much shorter fragment e which is , as stated before, part of the same manuscript as 1. Fragment e has been held in high esteem by Pfeiffer and Stihns since it is the only one to mark the acrostic RVODOLF (16151-157) by rubricized capitals. Sunns declares himself unable to group this fragment on grounds of its alleged lack of evidence. In fact, a comparison with D is impossible as the last leaves of of D have been lost, the same holds true for b and M. The initials in e coincide completely wxth those in K (even to the extent of assigning a marginal initial to 16 0 29 and forgetting its execution), but K does not mark the acrostic. Furthermore, there are some common readings and mistakes in e and in KaKc, as follows: 16049 16050 16064 16148 A: die kriechisch kunnen verstan; d.: k. kunden verstan E KaKCe: die criecsche kunnen wol verstan CGWL: die kriechisch kunnen sich verstan CGW: waer ez in kriecheschem gelan KaKce: wer ez criechisch gelan (see Ktfpke1s apparatus) A: ... in kriechscher rede g.; hiet ez der herre nit getan E do geviel diu geschiht do viel d. g. (see Kopke1s apparatus) wunschet mir und iu daz wir; ... . daz ir E wunschit mir und daz wir; w. iu unde mir A Pf KaKCe GWKC K e 130 1616 0 AEGWK : hilf uns daz wir von schame rot; daz wir niht von . . . C K e: hilf uns daz von schame rot The last two scribal errors could possibly be found in Ka as well and have been tacitly corrected in Kttpke's edition. Fragment e shows two or three minor variants of its own (the most important: 16071 "an suozer rede" instead of a c c lere), just as K or K do (K leaves out 16055-56), but the strong overall agreement between e and K K is beyond any doubt. It must be remembered that Kc is a very late manuscript with numerous cuts, thus the affinity is the more striking. In conclusion we maintain that the fragment consisting of the two parts e and 1 belonged to a reliable manuscript of the DKaK° -branch, similar to dq, and should be consulted whenever a reading in D is in doubt or non-existent. 4-. 2 . 3 . Fragments m and F2 The so-called 'Gdttweig fragment' (m) owes its name appar ently to a misunderstanding. It was described for the first time by the nineteenth century Austrian germanist Joseph Diemer, director of the University Library in Vienna, who began his article with the somewhat ambiguous words: "Zwei Pergamentblatter in Folio aus dem Ende des 13. Jahr-hunderts, deren Mittheilung ich der zuvorkommenden Gute \ 131 des hochwlirdigen Herrn Bibliothekars and Subpriors des Benedictinerstiftes zu Gttttweig, P. Gottfried Reichardt 7 3 verdanke." Stihns who based his study of m on Diemer's description concluded from the above quoted word "Mit-theilung" that this fragment was actually located in the monastery of GBttweig (Sfihns, p. 4-). Since then fragment m is known and referred to as the "Gttttweig fragment." However, within the past two centuries there have never been any Barlaam fragments in the monasterial library at Gttttweig and it might be assumed that P. Reichardt acquired these two leaves from a private owner and forwarded them ("mitteilen") 74 . to J. Diemer. - Unfortunately, their whereabouts are unknown since then. It would seem a fruitless endeavour to analyse a lost fragment on the mere basis of a collation, if it were not for two reasons. a) m must have been one of the oldest known manuscripts and Diemer's description and catalogue of variants is very extensive b) we discovered a matching fragment to m which broadens the basis for our analysis and allows us to take outward criteria such as handwriting, execution of initials into 7 3 J. Diemer, "Kleine Beitrage . . .," p. 6 50. 74 We are much obliged to P. Petrus van Aalst of, the Stiftsbibliothek Gflttweig and Dr. Otto Mazal of the Oster-reichische Nationalbibliothek in Vienna for their kind . information. 132 consideration as well. This other fragment containing parts of the same otherwise lost manuscript is presently kept at the Staatsbibliothek Preussischer Kulturbesitz, Worstbrock lists it as No. 2. The four features proving that m and F2 belong to the same manuscript are a) both fragments consist of vellum folio leaves and date of the late thirteenth century, b) their pages contain three columns of 5 8 lines each, altogether 174- verses per page, a very high number, c) their marked dialectal forms are the same (Tyrolian, according to Diemer who lists the particularities of m), d) m and F2 show an identical relationship to other manu scripts, foremost to C. Taking m and F2 together now, a total amount of 1210 verses is at our disposal: more than half only in Diemer's vari ants (3107-3448 and 6203-6548) but the remaining 522 verses extant in F2 (411-758 on the first complete leaf, and on 'the second one with its upper half cut off 5193-5220, 5251-5278, 5309-5336; 5367-94, 5425-52, 5483-5510). A glance at the paragraphs of m and F2 in comparison to the other manuscripts gives the first clue as to their grouping: m and F2 show the same pattern as AbCD, yet, more precisely, F2 shows the same large initials in length and in shape as C (we cannot judge for m since Diemer did not differentiate). The I initial at 5497 ("Ich horte sagen . . .") is nine lines high both in F2 and C and at 133 5 30 7 ("Josaphat sprach . . .") it is five lines high in both (the upper two lines are cut off in F2 but one can easily calculate the initial's full height). The three other manuscripts which place large initials at these same spots are L and W and to a much lesser extent G where the rubricator forgot the execution of the marginal I at 549 7. Therefore, one has to study the relationship of mF2 with C in the first place before widening the scope. The closeness of m and C was pointed out already by S5hns (Das Handschriftenverhaltniss, pp. 41-43). Of the eight examples given by him to prove mC's common stand against the other manuscripts two are incorrect, however: 82,27 ( not 82,17 as Sfihns mistakenly writes) reads hohist in A as well as in mC and 164,14 reads dir volgen und diner  lere in all manuscripts except DKaK°. The following common variants of m and C are a few out of many, the sigle m will be applied also for F2 from hereon. 439/440 the two verses are interchanged in Cm (WL read completely different, see Stthns' Anhang) 590 Pf ob er so reinet sinen sin mC ob er so raine wil hie sin 711 Pf nu hilf ich dir alrgernest mC nu hilf ich alrgernest 742 Pf (daz er der genaden gotes/) gar vergaz.. mC. omission of gar 134 3448 Pf: nu mac niht geschehen daz mC: nu kan niht . . . 5392 Pf: man vert die witen straze gar / diu gen des - 94 todes porte gat / der enge stic verwahsen stat mC: plural forms gant / starrt, die engen stige 6312 Pf: zeinem huse hat gegeben mC: ze ainem huze ist gegaeben 6366 Pf: guote bruoder dar ir brot mC: guote. bruoder ander brot; gute bredigaere ir b. WL 65 39 Pf: doch nach dem tage . . . ; da nach tage . . . mC In some instances a common reading of mC is shared by WL, in other cases it is similar but not quite identical: 477 682 Pf mCWL Pf mC WL 5272 DKAG MmC L 6494 Pf mCWL er dachte helfen im durch got er gedahte hilf ime dur got ich wil dir alsam ein kneht I ... I iemer dienen ich wil dir helfen als ain knaeht ich wil dir dienen . . . den iemer mere staeten hort; d. i. streten hort E den iemer maere waerenden hort den ymer werndin stedin hort; W omits stedin von der gesihte er sere erschrac von der geschiht . . . The great likeness of m and C makes it quite possible 135 that one-is the direct copy of the other, but since they both date from the same period, the internal evidence alone will have to decide. In the text covered by the "Gttttweig fragment" two bible quotations occur after 3204 and 6206 which the scribe has left first in their Latin original before translating them (the Latin words are included by Diemer in his list of variants). Based on this evidence, S5hns concludes that C must have been copied from m as it allegedly does.not have the Latin version. He even suggests a reason for this omission: 11. . .viel wahrscheinlicher wird bei der Fluchtigkeit und dem Leichtsinne, mit dem C gemacht ist, dass sein Schreiber diese lateinischen Citate in seiner Vorlage m sehr unntitz gefunden, ja vielleicht nicht einmal verstanden und deshalb weggelassen habe" (Stihns, p. 42). Unknown to Stthns, manuscript C does have both Latin quotations in their entire length, just as m, only Pfeiffer did not bother to remark on it in his appa ratus. Thus, Stthns's argument for a direct dependence of C from m collapses. After sifting through the textual variants of m and comparing them to C and the other manuscripts, two posi tive statements can be upheld: a) m cannot have been copied from C, but b) C could have been copied from m. In regard to a) C has several readings of its own which m does not share, where m agrees with the other manuscripts 136 (represented by PfeifferTs text), such as in manegem hohem stein; in m. holen stain mPf vor der wunden arbeit; von der w. a. mPf alsus hast hie funden; alsus hie hast f. mPf ob ich sein ane waere; obe ich ez ane w. mPf ain reicher man bat christ; a. r. m. der bat c. mPf danne man tuo die enge / durch . . . dringen danne man die enge j dan daz man die e. Pf und im vii chlaine bezzert sein muot und kleine b. s. m. und im sein lieht; und ime daz lieht mPf 6474 and 6482 C: z_e naehst; instead of ze liche: mPf In regard to b) C is so close to m that, in addition to the common variants cited above they share also some obvious mistakes and gaps: 3405/3406 Pf: . . . klagen / . . . getragen mC: ... klagen / . . . begraben (impure rime) 6208 Pf: (gotes gttenlicher gewalt/) von den himeIn ist gezalt mC: in their translation of the preceding Latin quotation "celi enarrant gloriam dei" omit himeln 642 3 Pf: dem neven wart er heinlich 429 CD: 464 C: 488 CE: 5256 C: 5329 C: 5384 C: Wm: 5428 C: mPf: 5430 C: 4 and m: dem wart do er viel h.; dem wart er do v. h. C 137 6537 Pf: und gie mit in vii balde mC: omit gie Furthermore, we pointed out that m and C have strikingly identical initials, with one exception: at 691 m has an initial together with AbDEGLW which C has left out. In this context, this would speak also in favour of C being depend ent on m. Lastly, there are only five cases in which m shows a reading slightly different from C and the other manuscripts: 589 CPf: der vindet ez man lat in drin; 3304 CPf: so der lip mit der sele erstat m: . . . mit sele . . . 5214 CPf: wir miiezen un han wol bewart Gkm: wir miiezen unz han wol bewart 6394 CPf: unde ein vreuden richer ruom m: unde an froide richer r. 6433 CPf: daz ich mit gotes rate m: daz ich mich gottes rate. Do these five divergent readings in m and C rule out the possibility that C was copied from m? The third example (5214) alone is not strong enough to prove or disprove a grouping since the syntactic position of adverbs such as wol is highly fluctuating in all manuscripts of the time. The remaining variants are quite clearly scribal errors, misreadings which any alert scribe copying from that manu-138 script could have easily corrected. Therefore, in our opinion the evidence is convincing enough for the claim that C was indeed copied from m; Sunns1 theory is proven right in spite of his faulty arguments. In conclusion it should be stressed again that the "Gottweig fragment" m and the Berlin fragment F2 taken together deserve the greatest attention for a new critical edition. It is clear that m cannot be considered the archetype from which all manuscripts could be traced, its text omissions speak against it. But m is certainly one of the three oldest, if not the oldest text witness of Rudolf's Barlaam, older and more reliable than C. Moreover, it sheds some light on the relationship between C and LW which will be discussed in chapter five. It seems possible that the two leaves of the "Gttttweig fragment" are still hidden somewhere in Austrian archives and will be unearthed in the future. Then we would gain additional information on on those parts where up to now we had to rely on Diemer's collation. 139 4.3. Single fragments 4.3.1. Fragment h The Wurzburg fragment h, consisting of eight leaves in quarto, is about as long as m and F2 combined. It covers a major part of the text between 10933 and 13214, with seven leaves missing: four after leaf two (leaving out 11237-11852), one after leaf four (leaving out 12148-12307) and two after leaf six (leaving out 12604-12909). Moreover, some leaves are slightly damaged or have been partially cut, so that a total of approximately 1150 verses are pre served. Stthns'slist of variants is included in the Anhang of the reprinted Barlaam edition, pp. 501-502. The first 7 5 mention of fragment h was made by Karl Roth and a few years 7 6 later by Eduard Reuss. Pfeiffer does not seem to have known this fragment although he refers to Roth for fragment e (Barlaam, p. 408). In his thesis Stthns devoted a short paragraph to h claiming that h is a direct copy from A and that, wherever they differ, "... dann hat A gewfihnlich die richtige Les-art, die von h fldchtig und falsch abgeschrieben ist." His conclusion as to the value of h is very negative: "h ist daher bei etwaigen Textfeststellungen ebensowenig Werth 75 K. Roth, Deutsche Predigten des XII. und XIII. Jhdts. (Leipzig, 1839), p. 6. 76 E. Reuss, ZfdA 3 (1843), 446. 140 beizulegen wie b" (Sflhns, p. 40). In the evidence produced by Stthns to support his assertions, he managed not only to misread Pfeiffer's apparatus frequently, but he even contra dicted himself blatantly: in his own variants (see Barlaam, p. 501) he indicated correctly that verse 277,10 (11030) had initially been left out by the scribe of h but after wards been inserted in the bottom line of the same column -probably, we may add, with an omission sign in accordance with the scribal practice of the time (since the margin of the page is cut off, we cannot tell, but 110 30 is put in parentheses). However, in his paragraph on the relation ship of h, S6*hns claims that this particular verse is mis sing in h and since it is extant in A, the latter could not be dependent on hI The Wurzburg fragment gives proof of a careful scribe: its clean and regular handwriting, its initials indented over three or four lines and the meticulously drawn set of lines produce a remarkable contrast to manuscript A. There is reason to believe that h was written before 1300 which would make it one of the oldest text documents we have. Its thirty recognizable initials agree with those in A (28), C (29), b (29), and mostly with D (30). Both h and D have preserved the old paragraph pattern equally faithfully, but, as the textual variants show, they are not closely related. Stihns was right in pointing out the strong affinity between A and h, even though he was led to wrong conclu-141 sions. The following list of major common readings in A and h should be sufficient to prove their relationship: 13081-32 omitted in Ah 10979 Pf: da lebet diu gots essentia Ah: da lebet gots essentia 10984 Pf: ein also vreuderichez leben Ah: ein also vrolichiz leben 11084 Pf: der gar vervluochten diete Ah: der gar virworhten diete 11861 b: daz zeigte ich gerne baz mShtich; das zaigte was mocht ich E Ah: daz zeigit ich gerne moht ich (the "Damenpreis" passage 11735-11871 kept only by AhbE [B^ h carries the text from 11853 on) 11986 Pf Ah 12035 Pf 12073 Pf Ah behielt er luter als ein glas bihielt er als ein luter glas dulten; Ah: liden (see also p. 124) twanc si nach dem geheize han; . . . wan Cb twanc si nach dem geheize an. It cannot be concluded from these common readings that h was copied from A, as Stthns proclaimed. In a number of places A presents a version of its own where h agrees with the other manuscripts and Pfeiffer's edition. 11003 hPf 11198 hPf A daz er belibe sigelos ; d. er wurde s. A swenn er din reinez herze siht swenn er din herze reine siht •1.4 2. 11230 hPf A 12054 hPf A 12080 hPf A 12452 hPf A des got an dir begunnen hat d. g. an dir gegangen hat des ktineges kint von Syria; . . . assyria C d. k. k. von persia din liehtiu jugent d. 1. varwe behalten gote und siinden bar behalten und s. b.; behalten reine und s. b DKaKG These examples being only a few out of many, there can be no doubt that A offers a much less reliable text than h and that it presents a later and possibly contaminated stage. Fragment h contains hardly any scribal errors, apart from 12090: daz muoz iemer mich; instead of: muet, , and 12538: the omission of wolte. In conclusion, we believe that A and h represent an individual branch or subgrouping of the Barlaam text tra dition, clearly different from the DKaKc group or any other version. The Wurzburg fragment would fully deserve the attention of a new Barlaam editor, but regrettably it is A and not h that has preserved the text in its entire length. 4.3.2. Fragment i The Berlin fragment i consists of two vellum leaves in quarto with two columns of 4 6 lines per page. The text preserved extends from 4607 to 4790 and 4975 to 5158 (one leaf is missing in between). The outward appearance of the manuscript is very neat and the text is legible in nearly all parts. S6"hns dates it as of the thirteenth cen tury (Stthns, p. 3), but the catalogue of the Preussische Staatsbibliothek is more credible in assigning it to the 77 fourteenth century. S5hns places I m his "Reihe CLBE" and maintains that i is most closely related to L. All of the five examples cited by him as common readings of iL alone are not only insignificant but can also be found in other manuscripts. Before screening the textual variants, the paragraph divisions in i should be considered. Out of its eight initials, there are three different from the AbCD pattern: at 49 87, i has a common initial with L, at 5009 with GEMLW and at 5143 with GEMk. Furthermore, i has capitulum signs in the margins at 4707, 4756, 4775, 5041 and 5105 which have no equivalent in any other manuscripts (there is a paragraph at 4775 in AK and at 5041 in E, however). These marginal signs occur at the beginning of a parable (in 77 H. Degering, Kurzes Verzeichnis der Germanischen  Handschriften der Preussischen Staatsbibliothek, vol. I (Graz 1970), p. 95. •+775 the Tale of the Man and his Three Friends, and in 5041 the Tale of the King for One Year), and they mark the beginning and the end of the typological exegesis of the parable of the Man and the Unicorn at 4 70 7 and 47 5 5 (the marginal sign at 510 5 stands at the pivotal verse within the parable and is therefore somewhat different). These parables (exempla) are integrated into the didactic dialogue between master and pupil and are not usually marked by ini tials in most manuscripts, especially AbCD. So these mar ginal signs in i could call for the attention of a copyist or a public reader who sought only after these edifying tales. To what extent the Barlaam parables preserved in various exempla manuscripts (Bispelhandschriften, see p. 26) show readings similar to i should be compared in detail*, the Unicorn versions of fragments f and g suggest a possible relationship with i. The comparison of initials had left open whether i belonged to the GEM group or to L. Its text, however, shows a much stronger congruence with GEM than with LW. 4612 Pf: do er was in sorgen vluht; do er also was xn E do er sus was an s. v • 3 do er sus was in GM 4697 Pf: ein kleine honicseimes gan EMfgi: ein lutzel h • 3 ein weninc h. G 4756 Pf: do sprach der guote Josaphat 145 d. sp. die reine J.; d. sp. der reine J. GEMDK 4776 Pf der drier vriunde nam sich an CGEMfi der nam sich d. v. an 5126 Pf sie sluogen in vii sere GEMi sie zogen in v. s. The common variants with L and also with W are of a less significant nature, parallels such as 4620 iLW: bournes; boumelines Pf 4675 iL: an sinen grozen noeten; in sinen g. n. Pf 4716 iW: vertaget; verzaget L; betaget Pf Moreover, L and W do not share any of the above cited variants of i and GEM but have numerous readings of their own: e.g., 5045; vremeder liute; against Pf: vremeder site, at 4745-56 the order of verses is inverted and 4772-73 are omitted. Finally, judging by the number of missing verses between leaf one and two, i cannot have omitted any text. In LW, however, as well as ABb and probably C four verses between 4883 and 4886 are missing. Thus, Sflhns theory of a close relationship between i and L cannot be upheld. Fragment i gives proof of how the wording preserved by more reliable manuscripts has been watered down through the continuous copying process. More easily understood phrases replace older ones or rather unusual ones, e.g., the roaring of the Unicorn sin liiejen (4608) is changed in i 14 6 into sin don (likewise in G: sin stimme), at 5074, instead of in ein vremedez einlant, i reads in ein verre vremedes lant, and at 5134, instead of von maneges mangels arbeit, i reads an manigen dingen arbeit. Fragment i alone does not offer any alternatives for a text reconstitution other than those in common with GE and more so with M. 4.3.3. Fragment k Fragment k consists of three vellum leaves, one leaf is missing between the second and third leaf. Its text covers 5129 to 5448 and 5609 to 5768; it is written in two columns with 40 lines each. Unfortunately, this fragment is in very poor condition, several pages are hardly legible due to chemical stains and two pages have been reduced in size by a vertical cut. S5hns used a transcript for his list of variants (Barlaam, Anhang) but does not attempt to group k: "k ist in Folge seiner zu geririgen Ausdehnung unbestimmbar" (Stthns, p. 41). In the case of k, a look at the paragraph divisions proves advantageous since even at places where the words can no longer be deciphered, the position of initials can still be made out. The initials in k agree entirely with those in GEM, notably at 5143 (as well as i), 5381, 5631 (as well as BLn), and 5663 (as well as B). A comparison of k's textual variants with the other 147 manuscripts confirms the first impression that there is a constant agreement with the versions of GEM. 5142 Pf: die im si brahten . . . GMk: di si im brachten; die sy brachten • • • E 5174 Pf: uns den burgaeren git GEMk: von den b. g. 5213 Pf: dar han wir eine lange vart GEMk: dez han wir ... 5305 Pf: daz er dir ewicliche git GMk: daz dir ewiclicke g. 5665 Pf: daz wort hat er vii schiere vernomen GEMk: daz wort het er schiere vernomen 5677 Pf: daz ich hie von vernam niht e GEMk: daz ich vernam hie von nicht e Within the GEM-group, k is most closely related to th< "Gotha manuscript" G as shown by several common variants with it alone, such as 5183 Pf: und iemer leitlich ungemach kG: und iemerlich ungemach 5273 Pf: bi gotes durftigen hant kG: di gotes d. h. 5286 Pf: dirre broeden welte gelt kG: dirre snoeden w. g.; dirre ploden . . . E C: diser werlde broder gelt 5645 Pf: da kumber arbeit jamer zorn; d. k. arbeit und zorn WL 148 kG: da kuraer lamer arbait zorn 5764 Pf: sin vater mahelt im ein wip k: s. v. vreyet im . . s. v. vrigete im . . . Both k and G were written in the fourteenth century but are not dated more precisely. G cannot have been copied from k as k has a few divergent readings of its own 5292 k: swaz man durch in den alten git; instead of: armen 5624 k: und swaz man ir zu warhait giht; instead of: richeit 56 5 8 k: mit worten wol erkennen; instead of: werken On the other hand, k could have been derived from G, but the basis for such definite grouping is not large enough. 4.3.4. Fragment n This vellum fragment of the fifteenth century, the "Peters burg fragment," was described by the curator of the former 7 8 Imperial Library at St. Petersburg, S. Minzloff. It consists of three leaves in quarto and each page contains two columns with 2 8 lines each on which the text is writ ten continuously with periods separating one verse from another. One leaf has lost two thirds of its inner column 7 8 S. Minzloff, Die altdeutschen Handschriften der  kaiserlich tt'ffe n t'li ch en Bibliothek zu 'St.' Petersburg (Petersburg, 1853), p. 34 was unavailable to us. S&hns quotes from it, p. 43. 14 9 due to a vertical cut. Since SShns' information as to the preserved text is incorrect, a rectification is appropriate. First leaf: 2670-2756. Ira (2670-90) and Ivb (2734-56) have retained only one third of the full width, second leaf: 5555-5645. third leaf: 5999-6084. The paragraph divisions are sometimes marked by alinea with initials of two lines height (2729 and 5631) or they are merely indicated by a paragraph sign and an initial within the line (2693, 5615, 6031, 6061), probably set off in a different colour. Minzloff and Stihns were unable to group this fragment using Pfeiffer's and Kfipke's editions, but there are never theless, some group characteristics to be found in n. The first hint is given by the initial at 5631 which occurs only in GEMkLB. The text of fragment n shows primarily a number of omissions of its own (e.g., 6025-26, me 6034, gulte 60 81, and others), as well as divergent readings not shared by others (e.g., hole instead of loche 5555, richeit instead of schohheit 5617). However, there is definitely a much stronger agreement with the GEM-branch than with any other manuscript as the following variants show: 26 8 3 ADEKW: als ein prophete hat enbart; erbart BbC n: (text cut off) at g_e (. . .); geoffenbart G KbL: offenbart; (no text for M) 5590 Pf: und in der hoehsten armuot swebent 150 6037 6038 GMBn: und in der grosten armuot swebent; in grosser a. E 5623 Pf: den dunket gar diu welt ein niht; 5623-24 omitted LW GEMn: omission of gar; ein wicht n Pf: bi_ grozen schrinen ligen vol EMn: dy grozen schrine 1. v.; s_o grozen . . . G Pf: swaz er dar inne hate hoi GEMCn: omission of er There are also several cases in which n apparently shares a reading with B, in addition to the common initial at 5631. 5631 Pf: der ktinic vraget in viirbaz; omission of in nBC 5638 Pf: da richeit armuot in niht lat nB: do richeit armut nicht enlat; . . . nit inlat L Pf: min zunge hat mir niht verjehen ABn: myn zunge hat des nicht v. Pf: du volgest minem rate nG: du envolgest m. r.; dune volgest m. r. ABM To draw positive conclusions from these common read ings would be unwise since we depend entirely on Pfeiffer's notes as far as B is concerned. The text parallels between n and B could be of a genealogical nature or merely accidental. The first possibility cannot totally be 6020 6065 151 brushed aside, as the congruence of initials had indicated that B might have been influenced in some way by the GEM-branch (see chapter 3.3.4.). In any event, n is a small fragment of no importance whatsoever for a text revision, most likely a very late offshoot of the GEM-branch. 4.3.5. Fragment p This fragment which consists only of the remnants of one damaged vellum leaf was described in detail by its present owner G. Eis. The following remarks are based on his text transcription covering the passage from 1933 to 1992. We have to take issue with Eis's final paragraph in which he follows the stemma and the variants given by SiJhns and groups his fragment together with B and E, but mainly with A close look at the one and only variant on which Eis based his assertion shows how mistakes have been perpetu ated in this field: "Von den unterscheidenden Gruppenmerk-malen fallt eins in die erhaltene Partie: in 51,11 [=1991] lesen (nach SShns, S. 36) die meisten Handschriften unzalhaft, nur B und E unzellich wie das neue Fragment" (Eis, 450). The fault lies not with Eis who quotes correct ly but with SShns whose information on B—as well as on any other manuscript except L—is taken entirely out of 79 G. Eis, GRM 49 (1968), 448-450. 152 Pfeiffer's apparatus (incidentally, B had been destroyed seven years before Stthns' s thesis appeared). Pfeiffer's text reads 1991 (51,11) sin kraft unzalhaft unvurbraht and Pfeiffer annotates for this particular verse: 51.11 unzellich DKEb. There is no mention of B. The correct and complete information should be, however: unzalhaft ACLW; unzallich DGK ; unzellich Ep; sin kraft unz (lberdaht b (on the margin); no text in KaM. There is no textual evidence at all that fragment p should belong to the Strassburg manuscript B. However, there are indications in the text that it can be assigned to the GEM group. 1960 Pf 1975 1981 EGK°p Pf EGp Pf GAbp E (see heilic eine reine heilic und reine die hoerent . . . Sie horent . . .; sie enhorint Einen got . . . Minen got . . ."} Ainen got . . . above, p. .112) It was argued before that Kb must have belonged to the GEM-group, albeit as a codex recentior. In the text passage covered by p, there is one outstanding common variant between Kb and p which would further confirm this theory. 19 36 Pf: daz mileze dir ze heile ergan; zuo guote Bb p: [ ] daz was wol getan (first word cut off) Kb: herre das was wol getan 15 3-Consequently, we suggest that Kb might have descended from the manuscript of which fragment p has preserved a tiny section, or that, at least, they both belonged to the same subgrouping of the GEM-branch of text tradition. 4.3.6. Fragment r The second of the two recently discovered "Schaffhausen fragments" is less spectacular than the first one (q): it is a younger document of the fourteenth century, it has suffered more physical damage, and it presents more problems concerning its position within the manuscript tradition. Its two leaves cover the text between 14192-14314 and 14559-14680 which we consulted in a transcript made by 8 0 P. Ochsenbein's attempt to group r (based on Pfeiffer and SShns) fails, but rather than offer an arbi trary solution, he leaves this question open and presents the contradictory common variants (which would require a number of rectifications and additions, nonetheless). The paragraph divisions in r are indicative of a group ing only insofar as they are different from GEM as well as from LW. In three instancesj r omits a couplet: 14201-202 and 14307-308, both times together with KC, and 14583-584 8 0 For further information on r, see P. Ochsenbein, "Zwei neue Bruchstucke zum Barlaam und Josaphat des Rudolf von Ems," ZfdA 101 (1972), 322-326. 154 together with b (none of which is given by Pfeiffer). As mentioned before, K abounds in text omissions throughout the entire work, and even here, in the passage under study, KC has two more cuts (14199-200 and 14311-312) which r does c a c not share. K belongs, as suggested before, to group DK K . This shows clearly in our passage as well: DKaKG change the order of verses in 14577-578, and in 14577 they read iuwers landes krone instead of des riches krone as do all the others, including r. Thus, there is little reason to believe in a group affinity between r and K°. A common reading in both such as 14284 Pf: von sinem zwivel schiere erlost; K°r omit schiere 14632 Pf: des landes krone und ouch das lant; K Ar omit ouch must consequently be dismissed as accidental. Likewise, parallel omissions of a couplet do not necessarily prove the interdependence of the two or more manuscripts in ques tion. They are sometimes triggered independently by the nature of the text .itself: a couplet written in verse lines with a rhyme identical or very similar to the preced ing one is more likely to be overlooked by a scribe than any other, and also the rhetorical repetition of certain key-words can easily produce a mental slip on the scribe's part. These reasons could be responsible for the common verse omissions in r and Kc as well as b. 155 14572 Pf rb 14290 Pf rb 14234 Pf rb On the other hand, there are also some noticeable text parallels between r and b, apart from the common omis sion. gen sinen hulden missetan gegen ime missetan do du der welte wlirde kunt daz du der werlte ie w. k. diu vorhte die sin sunde im bot d. v. die ime sin slinde bot Moreover, owing to Pfeiffer's apparatus we know that B, the destroyed Strasbourg manuscript closely related to b, shared some variants with r at 14236 Pf: wan owe mir armen we owe mir armen und owei du hast mir versuenet got; d. h. mir gesilenet g. AbCWL rB: du hast mich gisuenit got Pf: des lone er dir rB: dez lone dir On the basis of this sparse evidence it seems possible that fragment r was remotely affiliated with Bb, but on both sides there are also several particular deviations. 14311 ErB Pf 14312 156 4.3.7. Fl "Basel fragment" This single octavo vellum leaf presents the text from 3903 to 4062 in carefully written double columns of 40 verses. On the margin, the rubricator added ornamental designs into which the initials at 3915, 3941, 4013 and 4043 are inter woven. A. Gessler in his brief notice of this fragment dated it from the end of the thirteenth century and claimed 81 that it agreed mostly with the variants of B. However, the examples given by him cannot support this statement. It is remarkable that Fl does not subdivide the passage between 3941 and 4013 (the end of a lengthy speech by Bar laam and his disciple's response and next question), where ACD have two paragraphs. This brings the Basel fragment in line with G, the Fulda fragment (F8), and also with b (the text of M begins again with 3996). In its textual variants, Fl gives proof of being close to the GEM-group. 3909 Pf: nach aller unser guottat GEF8F1: nach unser aller g. 3917 CDWL: diu boteliche lere GEKK°BbFl: diu gotliche lere 3950 Pf: und leidez zil . . . GEKbF8Fl: und leidez ende It is not possible to determine the place of Fl within the GEM—branch more precisely because M and F8 do not have 81 A. Gessler, "Bruchstlick einer Barlaamhandschr if t," AfdA XIV (1888) , 147. 157 a concurrent text with Fl. But the Basel fragment seems to be closer to the latter ones, especially F8, than to the Gotha manuscript G which has a few variants of its own: 3910 FlF8Pf: gedienet; virdinet G 3912 FlF8Pf: daz du . . . solt; must G 3929 FlF8Pf: mit staete; mit vlize G 40.30 FlMPf: er nennet; der nennet G 4052 FlF8MPf: . . . ie niuwe; omission of If the basel fragment Fl is correctly dated as of before 1300, it is one of the earliest representatives of the GEM-branch. 4.3.8. F3 The first of these two vellum leaves has its outer column reduced to one half by a vertical cut, thus the text between 11209 and 11289 is incomplete. The total text pas sage covered by F3 extends from 11169 to 11330 and 12753 to 12914. Each column contains 40 verses with initials set in over three lines at 11223 and 12895 or, in the case of I-initials, drawn out over ten lines on the margin at 11191 and 11317 (due to the vertical cut we do not know if F3 had initials at 11257 and 11287). The fragment is dated as of the (probably late) thirteenth century by Degering's cata logue (see footnote 77). F3 shows a relatively high number of errors and read-158 ings of its own, such as 11184 Pf F3 11208 Pf F3 do kerte ich mine meisterschaft do hatte ich in buoze im buoze 11255-56 Pf: sich schieden gotes degene / mit bruoder-lichem segene F3: (. . .) zu lebene / (. . .) gotes degene (rest cut off) 12821-822 are reversed in F3. Of the two verse omissions occurring in F3 (11325-26 and 12787-88), the first one is shared also by Ka which then rearranges the sequence of the following verses. Throughout F3, there are congruent readings with G and E, and on its second leaf also with M (the text in M begins again at 12351), although these variants may be more convincing by their number than by their quality. 11210 11215 Pf GEF3 Pf -216 DKK^GEF3 11313 Pf GEF3 12779 Pf F3 daz du bist ein alter man omission of daz und zeiget dir demdete / durch sine reine gtiete und z. mit demtlete / dir sine r. g. daz er da vor seiten lie daz er davor doch seiten lie er ist swie du machest in; . . . als du A er ist swere man machet in; swie man 159 GEMF10 12789 ACDKW ein vihe gehoert ez siht ez gat F3 ein vie daz horet daz sitzet daz stat GEMF10 iz sihet iz stat 12854 Pf ein toubez hor; ein tumraes hor GMWF3 Finally, at 12765, we find an interesting variant in ACDEKW: Din got dem du dich ergist (initials in AbCDKcL). This misreading could have originated from a rubricator's mistake whereby the logically correct form Din got was altered into Ein got as in G or into Eyn got as in M. This latter form seems to be the base for the corrupt reading in Thus, we assume that F3 belongs to the GEM group but, due to the small extent of the fragment, a more precise grouping is not possible. 4.3.9. F6 "Breslau fragment" These two vellum leaves contain the Barlaam text from 8477 to 8804, but since the second leaf was cut down in size, its top verses are missing (8644-48, 8684-88, 8725-29, and 8765-69). P. Pietsch described this fragment and gave its variants from Pfeiffer's edition but he did not attempt to F3: Ey dem du dich ergist; instead of F3 . 16 0 assess its position within the manuscript tradition. F6 does not set paragraphs at any unusual places and yet it is significant that the passage containing the long speech of Avenier to his son Josaphat (8571 to 8693) is not subdivided (whereas ACD have three paragraphs), but that it is left in one coherent block just as in GEM. W and L have no subdivision in this passage, either, but their other ini tials do not agree with those in F6. G and M, on the other hand, show a total congruence in their initials with F6, and also E with the one exception at 8 561. The scribe of the Breslau fragment is presumably res ponsible for a few slips of the pen, e.g. , 8522 the repetition of the rhyme-word sere instead of mere 8505 the impure rhyme werde (instead of werbe) / sterbe 8711 and 8712 are contracted into one verse: (daz ein ieglich kint geste/) ze sines vater lere * unde muter ere; instead of Pf: ze sines vater lere / vater und muoter ere. In a number of cases F6 has common variants with GEM as the following selection shows: 8579 Pf: den reinen Josaphaten GEMF6: sinen sun Iosaphaten 8586 Pf: so denken aber vurbaz GMF6: so denke wir aber v.; so gedenken wir a. v. E 82 P. Pietsch, "Fragment einer Hsft. von BuJ," ZfdPh 13 (1881), 163-164. 161 8637 Pf: GEF6: 8704 Pf: GMF6 : 8721 . Pf: GEMBbF6: It is obvious that F6 is an offshoot of the GEM-branch. In one reading, though, it differs from the other manu scripts of this group: 8562 Pf: (diz gebet erhorte got /) im erzeigte gotes gebot/ (einen vreuderichen trost) GEMKb: im erzeigete sin gebot F6: im erzeigte got sin gebot Apparently, the scribe of F6 found the GEM version (replac ing the repetitious got by sin) in his source, but he reiterated got accidentally or on purpose without deleting sin, thus altering the grammatical construction of the phrase. This composite form, however, does not contradict our assumption that F6, through its paragraph divisions, its variants, its middle German dialect and its age (fourteenth century) is closely related to manuscript G. sie sint mir alle noch gelegen si sint noch under gar gelegn; gar is omitted in M vttr die welt vur guot vlir lip vor alle di welt vor gut un lip; . . . welt gut und 1. E min sele mir vii lieber ist omission of vii 16 2 4.3.10. F7 "Freiburg fragment" F7 consists of a vellum double leaf containing the text from 1081 to 1216 and 2043 to 2178. It is quite legible in most parts, only page lv and 2V show some discolouring. The handwritten catalogue notice by Richard Newald dates it from the thirteenth century whereas Worstbrock lists it as thirteenth to fourteenth century. The latter indication seems more likely judging by internal evidence alone. F7 gives proof of a relatively large number of variants from the other manuscripts. In some instances these variants must be scribal errors: Pf: daz mit des ungelouben naht / reht geloube wart bedaht F7: ... ungelouben craft / . . . wart bedaht Pf: drivalten mit der gesiht; drivaltikait ... F7 Pf: (die drie patriarchen sint /) huetaere gewalteclich F7: . . . / huetet er gewaltecliche In some cases, a look at the Latin source might help to decide whether a meaningful but isolated variant in F7 could possibly be considered authentic in spite of the testi mony of most other manuscripts or whether it is just one out of various different versions. On three occasions within our passage, this seems a reasonable approach: a) In the brief medical explanation on the roots of sickness 2129 -130 2149 2171 163 (1192-1202), line 1200 reads in most manuscripts like Pf: und ir materje ersterben; as opposed to WF7: und ir nature ersterben; das ir nature mus verderben b (b jumbles up the lines, the above quoted is 1199). The term materia, however, occurs at this particular spot in the Latin text and can therefore be regarded as authentic: "illi has humanas calamitates esse responderunt quae ex corruptae materia . . ." (Migne, t. 73, p. 458), b) 1117 Pf: ich wolde gerne vlir diu tor bLF7: ich wolde gerne vur daz tor The Latin text suggests the plural form as authentic: "etenim gestit animus ea quae extra has ianuas sunt perspicere . . ." (Migne, p. 457). c) 1094 Pf: hant dich beswaeret die hie sint F7: Beswerent dich ... Only in this case, the Freiburg'fragment is supported by the Latin text which shows a present tense: "quisnam sit hie moeror qui te obsidet . . ." (Migne, p. 457). These samples should make clear that F7, in spite of its given age and its numerous deviations from Pfeiffer's text, is by no means closer to what might be considered the "authentic" version. The few paragraph markings in F7 do not show any noticeable group characteristics but seem to agree more with AbCD which, differently from G and E, set paragraphs at 2051 and 2089. The divergent readings are manifold and contra-161+ dictory in F7. However, there is one important variant in F7 and b which gives a lead: 2175 Pf: Jacob zwelf sline do gebar / von den daz -77 geslahte gar / kam der israhelschen diet bF7: Jacob zwelf sline do gewan / von den daz geslehte kan / der israhelschen diet, (b omits 2177-78) This speaks very strongly for a genealogical affinity between F7 and b of which, there are further indica tions, such as the variants at 1200 and 1117 (see above). To what extent the destroyed Strasbourg manuscript B might have shared these parallels can no longer be determined. But there is evidence that Bb as well as WL have some common characteristics. All four manuscripts omit the couplet 2177/78 and also 1209/10. For the latter omission, we do not have any information on B, and WL leave out the preceding couplet 120 7/0 8, too. F7 does not share any of these omissions, but in a few common readings one could detect some wider relationship between F7 and LW along with bB. 1111/12 bLF7 1192 Pf AWF7 3113 Pf erden I . . . werde der smaehen siecheit ungemach der smehen siechen ungemach macht er im sa ein bilde; omission of sa in F7bL The picture does not become entirely clear. The only 16 5 conclusion we can draw is that the Freiburg fragment was part of a manuscript belonging to the same branch or sub-grouping as b, probably B, and in a wider sense WL. Assum ing some form of dependence, Bb and WL could have been derived from a source related to F7. That source would have had the gaps common to WL and Bb where F7 has the full text. The Freiburg fragment does not offer any valuable alterna tives to Pfeiffer's text. 4.3.11. F8 "Fulda fragment" and F16 "Prague fragment" The Fulda fragment consisted of a vellum double leaf which was cut into strips for bookbinding purposes. E. Schrttder succeeded in partially reassembling it and thus restoring 8 3 some of its text. In his description of the fragment, he gives its measurements and main variants but refrains from grouping it. Schrttder concludes: "Das bruchstuck hat keinen textkritischen wert und kaum textgeschichtliches interesse. es mag nur als weiteres zeugnis fur die grosse verbreitung des werkes dienen . . .: es m6gen gut 40 handschriften und fragmente bekannt sein. ein fragment, das wie das unsere 36 zeilen auf der spalte hat, ist, soviel ich sehe, nicht darunter." This last statement, however, is not correct, Schr5der must have overlooked an article by 8 3 E. Schrttder, "Fuldaer Bruchstuck von Rudolfs von Ems Barlaam," ZfdA 54 (1913), 23-24. 16 6 V. E. Mourek in which the author presented a Barlaam frag ment kept in the former Bohemian Museum in Prague (our sigle: F16). Both fragments, F8 and F16, show many similarities: they are written in two columns of 36 lines with nearly the same measurements (width: 11 cm; F8 seems to be some what longer than F16: 20 cm instead of 17) and the hand writing and rubrication present a very similar picture. Both fragments date from the fourteenth century. Their dialectal provenance is hardly different, either: F8 is, according to Schrtider, "oberdeutsch, alemannisch," but he admits that it does have Middle German forms as well. He mentions forms such as sal for sol, saltu for solt du:, e for Alemannic ae in words like swere, and he could have named several others, e.g., monophthong u for Upper German iu and lie in buzen for btiezen or stete ruwe for staetiu  riuwe, -Id- for Upper German -It- in forms like werlde. The Prague fragment F16 on the other hand shows Middle German forms more consistently, but has preserved a number of Upper German forms as well. Consequently, it is difficult to decide whether the Fulda and the Prague fragment belong to the same otherwise lost manuscript or not, the very limited text sample pro vided by F8 makes a comparison by photocopy nearly impos-84 V. E. Mourek, "Prager bruchstlick einer pergamenthand-schrift des Barlaam und Josaphat von Rudolf von Ems," Sitzungsberichte der Ktiriigl. Bflhm. Ges. der Wissenschaften, (1893), 1-16. 167 sible. It seems, though, that the differences between the two outweigh their common features: the different height of the column, F16 having a second vertical line to separate the capitals at the beginning of a verse (perhaps bleached out in F8?), and some slight differences in handwriting (e.g., the capital D) and dialect forms. Thus, it seems doubtful that F8 and F16 were parts of the same manuscript. However, they must have been written at approximately the same time and not far from each other (according to their linguistic features), and they belong to the same branch of text tradition, namely that of GEM. The Fulda fragment covering very sporadically the pas sage from 3894 to 4181 cannot always be compared with the London manuscript M, which has a gap until 3995, but F8 partly overlaps with the Basel fragment Fl. Of the GEM-group characteristics, it is sufficient to name the long sections without paragraph markers in GF8F1 between 3941 and 4013 and again between 406 7 and 4175 in GEMF8 (Fl ends with 40 61) and to recall the two textual variants given before (see p. 156): 3909 nach unser aller guttat, and 39 50 und leidez ende. The Prague fragment F16, which consists of less than half a leaf and one double leaf, contains the text between 6483-6554, 12969-13112 and 13691-13834. It pro vides a long text passage, a speech by Josaphat (1299 3-13112), without any paragraph divisions, just as G and M 168 do. Moreover,the rubricator in G and F16 forgot the mar ginal initial J at 12993. Likewise, the passage between 13681 and 13811, which is subdivided in AbCDL by three com mon paragraphs, is left in one piece in F16, whereas E and GM have only one initial in this whole section (E at 13731 and GM at 13781). Thus, in its paragraph pattern, F16 appears to be in line with the'GEM-group, just as the Fulda fragment was. This is also documented by the following variants: 6506 Pf: wan du so kumberliche; sit du s. k. F16GEM 12994 Pf: sinneloser Theoda; vii sinneloser t. F16GEM 13029 Pf: diner gote werdekeit KbF16: mit siner gute werdekeit; go'te M; gotlichen G 130 32 Pf: die liute twungen . . .; di lute trugen . . . F16GM 13089 Pf: der stleze der gewaere Krist GMF16: der suze und der gewere crist;. . . geware c. E 13097 Pf: von gotes kreften sprichet sus GMF16: von gotes worten sprichet sus; von des sp. alsus E 13754 Pf: der unheil der andern heil; der ain unhail ... C GEMF16: dirre heil der andern unheil The Fulda and the Prague fragment deserve little 169 attention in themselves; they both show signs of how the text at this younger stage in a different dialectal form gradually loses its precision. Blatant mistakes are rare, e.g., 13695-696 swer umbe gelt gevangen wart / (instead of: lac) daz gelt er balde vur in wac but slight alterations and omissions occur frequently. F8 and F16 enhance the impression that the Barlaam text as represented by the GEM-group must have been widely spread in Middle Germany in the fourteenth century. 4.3.12. F9 The ""Gottingen. fragment" These two vellum leaves of a late thirteenth century manu script contain the Barlaam text between 13811-13946 and 14903-15038. It is written in two columns with 34 lines each and quite legible, with one faded passage on page 1 . Its initials at 13811 and 13907 (thus not subdividing the section presenting Avenier's letter to his son), as well as 14905, 923, 973, 999, coincide entirely with those in G and M. The text in F9 is written with great care, there are no obvious scribal errors and hardly any deviations from Pfeiffer's edition other than: 13876 Pf: min herze ein kleine erliuhtet ist GMBF9: m. h. e. k. entluhtet ist; m. h. ein lutzel erlaucht ist E 170 13927 -28 13944 14938 14959 14973 15015 Pf: GMLWF9: Pf: GEMbF9 Pf GMF9 Pf GMWbF9 Pf GEMF9 Pf F9: siner guete lobetin do / sin munt sin herze. er sprach also Siner guete lobeter in do / s. m. s. h. sprach also (daz der verhertet adamas /.../) dir vorgeweichet suezer krist; gewichen AB von dir geweichet suezer crist und niht mit dem rehten gant daz si niht mit rehten gant; . . . mit dem r. g. E bi einem alten armen man bi einem armen alden man sus truoc er dan sin haerin kleit sus truoc dan s. h. k. ze gote schrei er alle stunt; z. g. rief er DKaK° ze got screi er zaller stunt; zu aller st. GMb There can be no doubt that the Gttttingen fragment belongs to one branch with G and M and to a noticeably les ser extent with E (also WLBb show occasional agreement with F9). F9 could not have depended on either G or M. G has two divergent readings of its own where F9 and M agree with the other manuscripts: 13 854 MF9Pf: unde im rehter volge jach; rehter warheit G 13884 MF9Pf: die min lip gevrumet hat; begangen G 171 On the other hand, the London manuscript M leaves out 149 34-937 where F9 and G have the full text, thus ruling out the possibility that F9 might have been copied from it. The opposite, however, that G or even more so M might have descended from F9 or a very closely related manuscript, seems rather likely, although the evidence to prove such a suggestion is far too slim. This would also be in accord ance with the fact that F9 is dated slightly earlier than M and G. 4.3.13. F10 "Hannover fragment" This fragment consists of only one and a half leaves in quarto, the verso of which is so badly destroyed that it is not worth photocopying. The description by Wolfgang Stammler lists a number of variants on which we have to 8 5 rely for the verso parts. Stammler dates the fragment as of the second half of the thirteenth century, but the form of some letters seems to point to a later period, presum ably after 1300. The text covered by F10 is: lr 10965-11006, 1V 11091-132; 2r 12643-726, 2V 12727-810. Thus, the text of the first leaf overlaps with the Zurich frag ment d and the Wurzburg fragment h, the second leaf with F3; M can be compared only for the text of the second leaf. Stammler concludes from his collation of variants: 8 5 W. Stammler, "BruchstUcke einer Barlaamhandschrift," Beitrage zur Gesch. d. dt. Sprache u. Lit. 43 (1918), 554-55. 172 "Die handschrift steht mit ihren lesarten keiner der bekannten nahe, sondern stellt offenbar einen besonderen zweig der uberlieferung dar. Eine reihe kleiner besserungen und anderungen zeigte einen spateren selbstandigen {iberar-beiter an." Stammler's assumption points in the right di rection. The hitherto unknown branch of text tradition to which F10 belongs is the one represented mainly by Q and M, to a lesser degree by E, and by a number of fragments. One indication of this affinity is given by the para graph sign in F10 at 10999 which it shares only with G and W (W shows a different initial pattern in the rest of the comparable text and can therefore be disregarded). The following textual variants support our theory sufficiently: 10965 Pf: mit dienestlichem werde / . . . uf der erde -966 GF10: mit dinstlichen werden / . . . uf der erden 109 86 Pf: lachte sin, herz unde muot GEdFlO: lachete herze und mut 10996 Pf: der warheit ie verkerte GEdFlO: der di warheit i verkerte 11104 Pf: daz ich dich wise uf daz leben F10: daz ich wise uf d. 1.; daz dich wise uf d. 1. G 11124 Pf: wirf durch den guoten got von dir GF10: omission of guoten 17 3 11127 Pf: unz an dines libes zil GEF10: biz an d. 1. z. 12681 Pf: do sie gesazen an den rat; sie gesazen GF10: do gesazen an d. r.; di gesazen . . . M 1274-7 Pf: du vervluochter alte unwis; . . . alter greys E GF10: der vervluchter alder u. 12789 Pf: ein vihe gehoert ez siht ez gat GEMF3F10: . . . iz stat (see p. 159). It follows that the Hannover fragment F10 is part of the GEM-group and stands closest to the Gotha manuscript G. It is too small and too badly damaged to be of any value for a critical text reconstitution. 4.3.14. Fll "Oettingen fragment'-' This vellum fragment of the fourteenth century consists of a double leaf in quarto. The text has become scarcely legible in large parts; it is written in two columns of 32 lines each and covers the text between 6285-6412 and 6670-6796. Its initials are placed at 6297, 6329, 6361, 6397 and 6697, 6733 which already gives a hint as to its pos sible group characteristics: at 669 7, only manuscripts GEM plus K (which could mean: Kb) set a paragraph. A comparison of the textual variants that Fll shares 174 with other manuscripts points into the same direction: 6287 Pf: rehte leben daz ist daz leben GEMF11 : r. 1. daz ist leben 6295 Pf: so sol daz tot ouch heizen niht EF11: so sol der tot doch h. n.; so sol daz t. doch h.n. GM 6301 Pf: so wirt des libes ende erkant GF18F11: so wirt des ende erkant; s. w. daz ende e. ME 6347 Pf: unser koch ist gotes segen GEMFll: unser koch ist der g. s. 6348 Pf: sunne tou hitze regen GMF11: s. t. hitze unde regen 6357 • Pf: swie vii ir der eine hat GEMFll: swie vii der eine hat 6765 Pf: in broeder menschlicher maht; in broder m. m. M EF11: in bloder m. m. ; in snoder m. m. G 676 8 Pf: do vraget er in aber sa GMF11: omission of in; do sprach er und fragt in aber sa E The Oettingen fragment Fll can be considered a furthe member of the GEM-group. It has a few minor textual diffe ences with either G, M, or E, but because of its large deletions it cannot be determined more precisely. 175 4.3.15. F13 "London fragment" Together with the larger incomplete London manuscript M, one leaf of another vellum manuscript unrelated to the first one has been preserved. This leaf has lost three lines in the middle of each column due to its being cut in half, but other than that the text between 5174 and 5 29 5 is well preserved. John Koch in his article on both London frag-8 6 ments seems to be more fascinated by this small piece than by M, probably because of its age (he dates it from the middle of the thirteenth century). On the other hand, Koch states rightly that its variants are of little inter est, so he contents himself with giving a mere transcript of the text. In it, he accidentally leaves out one verse: 5 210 frumet dc vor hin wirt gesant / (in daz gedende lones lant). A fragment of such great age would deserve particular attention since it might promise a version closer to the original one. However, in the case of F13 any such hopes are unfounded. The variants which F13 offers as opposed to the other text witnesses seem deteriorations rather than improvements. There are two clumsy additional verses by which a scribe intended to bridge the abrupt change of speaker after 5272: "Barlaam der sprach do / Der frage waz er vii fro." 8 6 J. Koch, "Fragmente von R.v.E. Barlaam . . .," ZfdPh 13 (1881)  78-89 . On F13 in particular pp. 87-89. 176 The true indication for the change in speaker comes after 5276 in all manuscripts, however ("sprach des herren lerer do"); in F13 this phrase has become redundant after the additional scribe's verses. Further examples of a dete rioration in the text of F13 are 5175 von den werden wir erslagen; instead of geslagen Pf as required by the context ("beaten," not "slain") 5225 (die lebent als ain tube tuot /) als sie ain her besezzen hat; instead of ein ar Pf 5 249 sint ein wol in solher not; instead of eine wile Pf. The three initials in F13 at 5189, 5223 and 5265 are in line with all manuscripts except GEM which leave out 5189. Only the following variants can be compared with other man uscripts : wie disiu welt sol zergan w. d. welt muoz zergan den•iemer mere staeten hort den iemer maere waerenden hort (see p. 134) daz almuosen dest daz guot daz almuosen ist daz guot The picture is not clear at all. F13 could be related to virtually any manuscript or grouping, its 112 verses simply do not contain sufficient evidence for any theory. 5196 5272 5275 AbCWLi DKKCGEMF13 ADKKCGF13 MCm ADKKC CGEMWLkmF13 177 4.3.16. F18 "Berlin fragment" The two quarto leaves of this fourteenth century fragment are in a very poor state: not only was the lower half of both leaves cut off for bookbinding purposes, but pages lv and 2r have become virtually illegible in part due to the imprint of another strip of the same otherwise lost manu script, with which our fragment had been affixed to the inside cover of a volume. Since P. Strauch gave a detailed description of its physical condition and of the text pre-8 7 served, it may suffice to say that the two leaves would have covered 5881 to 5992 and 6215 to 6326. Due to the damage sustained by F18, not even half of the text mater ial has survived and even less is still legible. Strauch remarks justly that F18 shows little deviation from Pfeiffer's text and indeed, nearly all of the variants listed by him are negligible. The only exception is at 6 301 ACDKK°LW: so wirt des libes ende erkant GF11F18: so wirt des ende erkant; s. w. daz ende EM Furthermore, F18 shows an initial at 59 7 3 together with GEMB, whereas ACDKKCLW set their initial at 59 6 5 (b has no text). Based on this evidence, the common paragraph and the common variant, we can assume that F18 probably belonged to the GEM-group, probably more closely to manu-87 P. Strauch, "Fragment aus Rudolfs Barlaam," ZfdA 52 (1910), 354-356. 178 script G and the Oettingen fragment Fll. One objection could could be raised: Strauch noticed that the missing leaf between the two of F18 could have contained only 222 verses (5993 to 6214) instead of 224 (eight columns with 28 verses each), and he concludes that F18 must have left out 6159-60, just as ABC did (Strauch erroneously adds E as well). How ever, since GEMKb omit the preceding couplet 6157-58, we can assume that the Berlin fragment F18 would have shared this omission. 5. Conclusion 5.1. Final grouping of manuscripts based on text and rubrication Having studied the relationship of the larger manuscripts on the basis of their rubrication and that of the smaller frag ments using a more textual approach, we shall now attempt to sum up our results, verify them by taking into account com mon text omissions throughout the work, and discuss the value of the "initial method." A parallel text omission can indeed be highly indicative of a manuscript relationship, but it is not the ultimate criterion of manuscript grouping. It can be caused, just as a common reading or a common initial, independently by a sometimes psychologically expli cable error of the scribe (see above, section 4.3.6.). Thus, one has to note whether parallel text omissions occur 179 repeatedly and how they relate to data gained by other methods. The first grouping which can be established without any doubt is that of DKaKc. In three instances, at 2817/18, 12243/44 and 13487/88, these manuscripts leave out a couplet (mentioned neither by Pfeiffer nor by Stthns). Furthermore, after 376 DK K leave out a verse and read differently as follows: (umb daz iemer wernde guot/) Daz dir din leben gevromidet hat/ Daz dir unlange in vroden stat. Pfeiffer follows the other manuscripts which read: daz iemer wert und hiht zergat/ daz dir din lebn gevremdet hat. A similar case at 14577/78 was cited already (see p. 154). As far as the reversal of 1602 3/24 in K and K is con cerned, D can no longer be compared, it ends at 15743. D does not share any omissions with other manuscripts, and presents a very reliable, early text version with only one major omission of its own (4213-4396, see p. 128) and three reversed couplets in a short sequence at 13563/64, 13571/72, 13593/94. In these instances, Ka and K° have the "normal" text, that of the consensus of manuscripts, there fore they cannot have been derived from D. Also, the source of Ka must have contained the author's digressions, since Ka has cut out only parts of the courtly praise of 180 ladies, whereas it is altogether missing in DKC. Ka and especially K have a number of individual text omissions which do not indicate any other affiliation. The only fragments that we found related to this group are dq and el (see 4.2.1 and 4.2.2) . The study of initials helped only partially in deter mining this group of manuscripts, it is true. As far as the position of Ka is concerned, the fault lies with the unreliable paragraph transcription in Kfipke's edition which distorts the agreement figures. However, the high congru-ence of initials between D and K indicated a close rela tionship between those two manuscripts. As we extended our scope and took into consideration divergent readings at places of rubrication, we found cases which pointed in the right direction (see pp. 115-16), which suggested that Ka as well as K belong to one group together with D, but that they could not have been copied from D itself. S6hns had claimed "that one class of manuscripts (Hand-schriftenreihe) was constituted by ADK as opposed to the BCLE-class. However, neither the study of initials nor textual comparisons yielded any conclusive evidence that A should belong to DKaKC (as to Kb, see pp. 109-10). A does not show any consistent agreement with any manuscript other than the Wurzburg fragment h (see 4.3.1). A has left out a verse or a couplet on its own here and there, but only the following omissions and reversals are shared 181 by other manuscripts and appear to be significant: 4 215/16 omitted in ALW. These verses have the same rhyme words as the preceding couplet. 4883-86 omitted in ABbLW; C has a missing leaf here on which four verses must have also been left out. These verses are plainly redundant. 4885/86 repeat the rhyme words of 4881/82. 6159/60 omitted in ABC (b has missing leaves). The preced ing passage leads to a rhetorical repetition of key-words and a double rhyme at 6157/58 which is left out by GEMKb. 6863/64 reversed in ACGLW, with C having a slightly differ ent reading. All these examples would range A on the side of Bb, LW and C rather than, of DKaK°. One rather-interesting omission should be mentioned here. A leaves out verses 7863-66, again repetitious in contents and rhyme, whereas L omits 7 86 2-6 5. It is intriguing to see that codex C had initially left out the same verses as A, but that they were later added on the margin. Were they added by the same scribe or his corrector at the time of the manuscript production, or were they filled in at a consider ably later date? And could the faulty version, in the meantime, have been copied by other manuscripts? At a few places C does indeed show corrections and additions of a later stage, which partly coincide with gaps in other manu-18 2 scripts.- Firstly, C had left out the first two paragraphs of the prologue; the overly large and illuminated initial at 63 proves that this is where the scribe started his work. The missing section (with the omission of 33-36) was later added on the inner cover, according to Pfeiffer "von anderer nicht viel spaeterer Hand" (Barlaam, p. 407). There is no parallel for this omission in any other manuscript. On the other hand the Heidelberg codex b has omitted verses 32 37-40, and so had C originally; however, they were later added on the margin "von anderer Hand" (Barlaam, p. 421). In this case, just as in the above cited common omissions, the gap in Cb could also be explained as independent scribal error: 3235/36 have the same rhyme words as 3239/40, and so the scribe's eye could have easily slipped down four lines and taken up the text from there. But it would seem unlikely that this type of scribal error should have occurred repeatedly only in some manuscripts and not in others. It is more convincing to assume that C and ABbLW might have drawn some of these omissions from a common source or, for that, matter, that they do not show repeti tious verses filled in by later scribes which the other manuscripts contain (this alternative is admissible espec ially at 4883-86 and 6159-60, see above). The information which we gained from the study of the above mentioned manuscripts is too contradictory to esta blish a stemmatic relationship. It seems that not only C 183 but also other manuscripts were later corrected from other sources and that this contamination cannot be disentangled. To quote two more examples: C has left out 9 2 75/76 at the end of a paragraph; in W line 9276 is initially omitted and later added at the bottom of the page. Likewise A has left out 2542; W also did so originally, but has added the mis sing verse at the bottom again. Coincidence or a sign of interdependence? We believe that the versions of A, b, C, L, and W have absorbed some interference and that, in spite of their loose group relationship, they cannot be traced in a straight line from one source. The "Gflttweig fragment" m is too close to C to be considered a key element for this conglomeration. The problem of dependence could not be solved by the comparison of initials, either, which showed only the same basic paragraph pattern in AbCD. Before abandoning this group altogether, we shall attempt a conclusive assessment of WL and Bb. The compar ison of the rubrication had also indicated that L and W have a strong similarity in the first half of the work, but that afterwards this special relationship ceases entirely and L follows the pattern of C. We considered the two parallel initials in CL at 8869 and 8895 to be a likely turning point. Taking now common text omissions into account, our tentative judgment appears fully corroborated. There are parallel omissions in LW at 1207-10 (1209/10 also in b), 1281/82, 2177/78 (also in Bb), 2295-98 (also in a), 184 3775-80, 3784/85, 4113/14, 4215/16 (also in A), 4772/73, 7616/17 (also in B), 8320 (W leaves out also 8321, L rever ses 8321/22) and 87 21. Furthermore, there are several coup lets with reversed lines, in one case, at 1709/10, together with Bb. The dividing line in L, however, shifts by about 300 verses. W and L show the last noticeable parallel at 9195/ 96, where both reverse the order of lines. It is only after this point, that L and C share the same text omissions, namely 9949-10046, 10055-58, 10065/66, 10083-250, 10333-54, 11587-94, 11921-24, 12235-38, 12259-90, 16089/90, 16105-110 and 16129-45. Furthermore, it is only in this second part of the work, that L contains some of the Latin bible quota tions which C has preserved after 13065, whereas its Latin verses in the first half of the text (between 2449 and 3785, at 6191, 6207, and 6930) do not occur in L. The question was raised earlier (see p. 10 3) whether W might have ultimately been derived from D, as the agree ment between their initials is slightly higher than that of W and C. A look at their text omissions, however, rules this possibility out. W and L share common text omissions with ABb and C (all of which occur in the first half, see above), but none with the DKaKc-group. As we pointed out earlier, C does not always represent the old paragraph pattern as carefully as D and sets initials of its own. Also, the major text omissions in the second half of C, 185 where W has the full text, would account for the lower agreement figure between C and W. On the basis of a textual comparison, we can now defin itely reject Sfihns's claim that the Heidelberg codex b was alternately copied from B and A. The study of the rubrica tion in b and of the relationship between the alternation of the two scribes and the various gatherings of the codex suggested that b does not follow two different sources, but rather that B and b originated from one source which was loosely related to A (see p. 100). This preliminary state ment has to be modified to some extent. It appears that within the first 5000 verses B and b are closely related, but not dependent on each other (e.g., B leaves out 1697/98 where b has the full text). Their common omissions consist of 733-36, 1369, 2064, 2971, 4157-60, 4411/12, 4752/53, 4774/75. There are also several parallel cases of reversed lines, the last such reversal occurring in B and b at 5289/90. Hereafter, the parallelism between the two manu scripts ceases completely and both have omissions entirely of their own. Codex b continues its regular paragraph pattern in agreement with that of ACD, whereas B seems to join the GEM-group in a number of common initials. However, in the final 150 0 verses of the work, the parallelism between B and b is re-established (recognized correctly by Sfihns, p. 37). Again, they have a common initial differ ent from all other manuscripts at 14613, they reverse two 186 lines at 14821/22 and omit others at 15371/72, 15481/82 (double rhyme, also omitted in LKaKb), 15729/34 and 15985/86. The most plausible explanation for this change in the rela tionship between B and b is to assume that the lost Strassburg manuscript B (and not the Heidelberg codex b) might have drawn its middle part from a different source, a source that obviously had some of the rubrication of the GEM-group. A great handicap in further pursuing this hypo thesis is the fact that for all information on B we are com pletely dependent on Pfeiffer's sketchy and not very reli able apparatus. It seems hardly necessary to compile further evidence in order to establish a third manuscript group, which we called after its three surviving codices the GEM-group. A number of fragments had to be included in this group, as well as the second Koenigsberg manuscript Kb, known to us through K6pke's few variants. As mentioned before (see p. 109), Kb shows important criteria by which it can be associated with this group, several verse omissions with GEM and also one additional couplet after 9400 with EG (section missing in M). While it is clear that the incom plete London manuscript M and the Gotha codex G are much more closely related with each other than with E (e.g., G and M omit paragraph 12933-62), it should be stressed that E does indeed belong to this group, albeit to a lesser extent. This affinity is suggested by the agreement of its 187 initials as well as several cases of divergent readings at rubricated paragraphs (see p. 114). E was written consid erably later than G and M and its scribe treated the text with some liberty, inserting a whole didactic poem, the Magezoge, into it and adding occasional paragraph titles. Apart from some text cuts of its own (e.g., 7729-58 and 15466-75) E shares the following omissions with other manu scripts, 3895 with b, 9176-78 with C, and 16089-90 with CL. It seems that E or its source could have also used a manu script related to the C-group. Besides these omissions E would have derived from this source the text of the two author's digressions which are otherwise not documented in the GM-group (with the exception of the Schimpfrede in Kb, according to Kttpke, p. 416). E has also preserved some of the Latin quotations which appear only in CmL, but they do not correspond entirely with those in C. The only Latin verses in E occur towards the end, after verse 1306 5; they are missing in the earlier passages, after 2449, where C and m abound in them. Moreover, E contains one Latin quo tation which is left out in CL: 13250 "Vivo ego dicit dominus nolo mortem peccatoris etc." Thus, there could have been no direct influence of C and E. And yet E gives proof of an amalgamation of two different text versions, that of the GM-group which prevails and that of the C-group. In view of the nearly total agreement in the paragraph divisions of G and M, the possibility that one manuscript 188 might have been copied from the other had to be left open (see pp. 104-05). G could not have descended from M, since the latter has several omissions where G shows the full text (5641-42, 6067-68, 6135-36, 8171-72, 8247-48, 13163-64, 13497-98, 14117-18 and 14934-37). M, on the other hand, cannot have been copied from G either, for the following reason alone: the Latin bible quotation which C and the "Gottweig fragment" m insert after 6206 ("celi enarrant gloriam dei et opera manuum ejus annunciant firmamentum") appears also in M but not in G. It is the only Latin verse in M; none of the others which we encountered in Cm, E, or L have been preserved in M. Certainly M cannot be derived from C or m, otherwise it would have copied their faulty omission of himeln at 6208, in the German translation of this Latin quotation (see p. 136). Consequently both M and G can merely be derived from one source. This brings us to a final point. We noticed that G shares a common omission with A at 119 7 (a redundant cou plet with word repetition), and that it reversed the order of verses at 6863-6.4 together with ACLW and at 16141-42 together with ABEW (CL have omitted the paragraph). M, in spite of its close affinity to G, does not coincide with it in the first two instances and has no text for the last one. With respect to these common readings, to the Latin quota tion in M, to E's occasional parallels with bCLW and to its Latin verses,should we regard the GEM-group as a sub-group-189 ing within the larger conglomerate group represented by manuscripts such as C, m, A, B, b, L and W? Or should GEM be rather considered an independent "branch" with its own original readings? If we rigidly adopted the first position, we would have to account for the preservation of 4883-86 which is omitted in ABbLW (and most likely in C), whereas these verses appear in DKaKC as well as in GEM (see p. 181). If we adopted the other position, we would have to assume an archetypal source of the GEM-group, which would have con tained the Latin quotations as well as the two authorial digressions. The few parallel text omissions in GEM and manuscripts of the "C-group would have to be explained as coincidental scribe's errors. Let us, at this point, recall Bedier's warning, that, the further back an investigation into a text tradition is pushed, the more it becomes mere speculation. Throughout our study, we found manuscripts GEM plus a number of frag ments to be a group clearly distinct from all the others, not only by a different rubrication pattern (which we con sider as of a more recent stage), but also by textual criteria, common variants, omissions, and additions. It would serve no purpose to draw a stemma that pretended to trace the genealogical relationship of all manuscripts back to an archetype. The image of a tree with larger and smaller branches would, in our view, be a mirage rather than a useful model. After all, we know only very few 190 surviving manuscripts out of a probably far greater number, and we do not always know them in their original form. Therefore, whether GEM and their related fragments ultimately go back to an ancient text version equivalent to that of the "D- or the *C-group, or whether their source is already the product of a mixture of these two groups, cannot be esta blished. In any case, it is an important group in its own right within the Barlaam text tradition and should be given credit accordingly. In our attempts to group and assess all existing Bar laam manuscripts and fragments, in order to help pave the way for a new text edition, we found the method of tabu lating and comparing the position and reading of all rubri cated paragraphs very useful. As shown in the individual cases, this method has its limitations. Sometimes the per centage values did not constitute clear indication of a manuscript's group affiliation, and they certainly do not determine a genealogical relationship. But which other methodical' approach in this field is free of ambiguities? On the other hand, adopting it as one of several methods, it draws attention to parallels between manuscripts, it indicates when a manuscript changes its paragraph pattern and thus possibly its source, and it can channel, support, or contradict results gained by a comparison of textual variants. Moreover, it provides data which can hence be used in structural studies. We hold that an inquiry into 191 the manuscript relationship of a medieval work, which ignores the transmission of its rubrication, is incomplete. Moreover, such an evaluation of the paragraph agreement could be efficiently organized in order to save time, the mechanical work could be delegated to assistants, the data computerized. Naturally the figures would have to be interpreted, but the basis for a division into groups would become broader and more solid. And that means also: . the areas, in which a positive judgment cannot and should not be be ventured, would be more clearly defined. Finally, what principles for a new Barlaam edition can be derived from our study? We have come to the conclusion that a new edition should follow the Freiburg codex D as a Leithandsehrift. Wherever D has a text omission or an obvious mistake, the reading of the manuscripts and frag-ments belonging to its group (dq, el, K , and Kttpke's edi tion representing Ka) should be given priority for the necessary emendations; their different dialectal forms would have to be adapted to those in D. Only when the group as a whole has a gap or an obvious misreading would the main representatives of the two other groups be used, namely C (and m) on the one side, and G (or M) on the other. Ideally the variants of C and G would be continu ously given together with the text of D, either in the margin or at the bottom of the page, while the readings of the other manuscripts could be printed in an appendix. 19 2 The passages containing the author's digressions would have to be filled in according to the text in A and b. One of the advantages in following D would be that D gives a reliable, early version of Rudolf's Barlaam in the author's own Alemannic dialect; another, that D has best preserved the early paragraph pattern. The greatest weak nesses of Pfeiffer's edition are, in our opinion, his inconsistent text selection (with undue consideration given to E) and the inaccurate and sometimes confusing material in the apparatus. If these shortcomings are successfully avoided, H. Rupp's sceptical remark that a new Barlaam edi tion might not be able to improve much on Pfeiffer's (Bar laam, p. 512), would consequently be proven wrong. 5.2. Geographical distribution of Barlaam und Josaphat Judging by the number of its preserved manuscripts, Rudolf von Ems's Barlaam und Josaphat must have gained a consider able popularity throughout the German speaking territories between the thirteenth and fifteenth century. This impres sion becomes the more convincing by contrast with the two other Middle High German Barlaam versions written in the thirteenth century, the so —called Laubacher Barlaam by bishop Otto of Freising, of which just one manuscript is 193 preserved, and the so-called Zurcher Barlaam, known to us 89 only'm two short fragments. An attempt to outline the spreading of Rudolf's Barlaam version on the basis of its manuscript tradition meets with grave obstacles. It is a rare exception that the manuscripts themselves give any direct information as to the date and the place of their origin; this occurs only in the cases of manuscripts C—written by a scribe Chunrat in 1284 E —written in 1459 for Veit von Egloffstein P.—the illuminated manuscript of 1469 , produced in the atelier of Diebolt Lauber in Hagenau (Alsace). For all other manuscripts and fragments we depend on secon dary information as provided by catalogue descriptions and articles, or even, in several instances, on our own dialect definition of a text. In the case of shorter fragments, such probing into dialectal characteristics could yield questionable results, due to the inconsistent spelling of most scribes and the shortness of the sample. Furthermore, some subtler aspects of such a linguistic analysis (e.g., the differentiation between the dialect of the source, the native dialect of the scribe, and the target dialect of the copy) could not be studied within the framework of this outline. 88 See Adolf Perdisch, Der Laubacher Barlaam: Vor-studien zu einer Ausgabe (Marburg^ 1904) . 89 See J. Klapper "Barlaam und Josaphat," m Verfasserlexikon, I, 171. 194 It has been established by various scholars that Bar laam und Josaphat was written about 1225, at a time when Rudolf lived at Hohenems in the alpine Rhine valley as a Ministeriale of the Lords of Montforir. One major aid in determining when the work was composed is its reference to the Cistercian monastery of Cappel near Zurich and its abbot Wide who provided the author with the Latin source (Barlaam, 144-149 and 16057-16074).90 It is most likely that Rudolf at the beginning of his career was known only in his home territory, the area south of Lake Constance, if we judge by the literary patrons whom he mentions in his earlier Per guote Gerhard and in his Barlaam. "Rudolfs Horizont ist gleichsam noch alpin begrenzt, seine Auftrag-91 geber leben in der unmittelbaren Umgebung seiner Heimat." There are no manuscripts left that could be ascribed to this early period. A, which was kept at the castle of Hohenems until the nineteenth century (see pp. 8-9), is definitely of a much later origin and was transferred to 9 2 Hohenems at a later stage. The only remaining witnesses to a Barlaam text tradition in this area, although dating from about half a century later, are the two matching 90 See X. v. Ertzdorff, Rudolf von Ems, pp. 80-89. 91 H. Brackert, Rudolf von Ems, p. 29. 9 2 See Edward Schr5der, "Rudolf von Ems und sein Litteraturkreis," ZfdA 67 (1930 ), 211: 11. . . unser Rudolf hat damit nichts zu tun." 195 fragments kept at Zurich and Schaffhausen, d and q. We can expect that literary works at that time would be propagated along the upper Rhine valley with its close interrelations between courts, cities and monasteries. The Freiburg manuscript D which belongs to the same group with dq but does not show their marked Swiss-Alemannic features would illustrate this point. Other, younger documents of the greater Alemannic area including Alsace and Swabia are the two destroyed Strassburg manuscripts B and a, the above mentioned Hohenems codex A, the second Schaffhausen frag ment r and the (missing) Stuttgart fragment F17, all of the fourteenth century. However, as indicated in chapter 5.1, these manuscripts have not, or not primarily, derived their text versions from dq or D. In this context, the two fragments e (in Munich) and 1 (Nuremberg), which in our opinion belong together, pose a problem. K. Roth, in his reference to e, states that 9 3 the dialect is "alamannisch (schweizerisch)." This assertion is probably provoked by the handwritten notice of a later owner (Roth suggests of the fifteenth century) underneath the final verses: "Wer dys buch findet, der sol es Hans von Wintertur wyder geben." The dialectal features of these 138 verses, however, show some unmis takably Middle German characteristics: the Umlaut is not 9 3 K. Roth, Deutsche Predigten, p. xxii. 196 indicated in words such as vroliche, horen, luge, truge, the Upper German diphthong ae is not differentiated from e, and an occasional d for t occurs in initial position (er det). The same features can be found in the Nuremberg fragment 1 which Stthns classified correctly as Middle German. On the other hand, this fragment 1, just as e, contains also a few minor Alemannic traces, e.g., the grapheme ch. for k within the word (dunchet, bechant, gechundet). In one case in 1, there is apparently even a remnant of the Old High German strong adjective declension which, with its full final vowels, was kept on in Alemannic dialects: 4598 unde disem valschan bi gestant. Pfeiffer mentioned that he received fragment 1 from a friend (Wilhelm Wackernagel) in Basel (see footnote 71). There is not information on its former owners, but one might assume that 1 was kept in the Alemannic-Swiss area. Could it have belonged, together with e, to the same "buch" which was owned in the fifteenth century by a certain Hans of the town of Winterthur (between Zurich and the Rhine), before it was cut up and its parts dispersed? Whatever the answer may be, both fragments, e and 1, give proof of an early transition from the Alemannic dialect to the prevail ing Middle German forms. The close relationship to D would account for this influence. By the second half of the thirteenth century, Rudolf's Barlaam had apparently spread not only to the North-West, 197 but also eastwards into the Bavarian region. The Munich codex C, written by a Bavarian scribe in 1284, was already preceded by a manuscript in Tyrolian dialect of which the "Gttttweig" and the Berlin fragment m and F2 bear witness. C was usually downgraded by nineteenth century scholars because of its several text omissions; Pfeiffer called its scribe "fahrlassig" (Barlaam, p. 408), which is a mild reproach compared to K. Roth's verdict: "Er anderte, strich oder schob ein, was ihm gut dunkte . . . Wenn es Alle machten, wie Chunrat ... so wUrden wir von den alten guten Texten bald Nichts mehr haben" (Deutsche Predigten, p. xxii). It seems to us, however, that the text omissions in C, whether they were introduced by the scribe Chunrat or already existed in his source, are not a sign of negli gence, but rather of vigilance. What is left out are ele ments foreign to a typical saint's legend, everything that detracts from its otherwise religious and edifying nature: the author's praise of ladies, his rather flippant digres sions (schimphrede) and even the brief mention of courtly love and aventiure in the epilogue (16105-09), the refer ence to the Guote Gerhard where the author addresses his literary public, and also the explicit description of pagan love and of the ungodly conduct of the Greek gods (9949-10047, 10083-251 and 10333-54). These cuts bear the stamp of a clerical censor; moreover, the preservation of the Latin bible quotations in the text could also point 198 to a monastic sphere. Helm and Ziesemer quote a passage of the Apocalypsis by Heinrich von Hesler in which the author encourages the reciter to skip the following para^ graph if he deems it too offensive for his audience. It would even seem conceivable that Rudolf himself might have written or authorized a version for a monastic audience (since "von Kapelle der abbet und al diu samenunge" encour-9 5 aged his work), as well as one for a courtly circle. Rudolf's Barlaam is otherwise very scarcely documented in the Bavarian area, with the exception of the two later fragments k, Fll ("Oettingen fragment"), and the above men tioned Munich codex E, all derived from the GM-group. It is possible that the slightly earlier Barlaam poem by bishop Otto of Freising with its more churchly character hampered the dissemination of Rudolf's work in this area. Nevertheless, the early Bavarian documents M and C do not constitute a dead end in the text transmission, but, as we showed before, they exerted a direct influence on the Middle Franconian manuscript L in its second half. Fur thermore, they are loosely related to ABbW, perhaps descend ing ultimately from one common source. We assume that by the fourteenth century, manuscripts of this "Bavarian line" had spread west, into the Rhenish regions, and that their 9 4 K. Helm and W. Ziesemer, Die Literatur des  Deutschen Ritterordens (Giessen, 19 51), p. 29. 9 5 X. v. Ertzdorff admits only the latter possibility pp. 88-89. 199 text versions merged with others. Where this influx took place (if as far south as Strassburg where B and a were preserved) and through which channels, we do not know. Manuscript W, formerly of the famous Ambras library, did not originate there, but displays Middle German character istics, just as the Wurzburg fragment h, which might have used the same source as A. Approximately a hundred years after the composition of Barlaam, the mainstream of its text tradition had moved northwards into the Middle German regions. These various linguistic areas, heterogeneous as they may be, share cer tain overall dialectal features, which set them apart from the Upper German areas, e.g., the trend towards monophthong-isation, the lacking of Umlaut, the lesser impact of the High German Sound Shift. Regarding this Middle German area for our purpose as one large unit, it appears that the number of Barlaam manuscripts and fragments originating from there is about equal or even superior to those known of Upper German origin. In the West, codex L must have been written at that period; its later owners up to the eighteenth century were the lords of Blankenheim in the Eiffel. Two further manu scripts of which only parts have been preserved show Middle-Franconian features. One is the brief fragment F10, kept at Hannover, which Stammler defines as Middle 200 German, perhaps Middle Franconian (see p. 172)- The other one is the incomplete London manuscript M which in the eighteenth century was rediscovered in a monastery near 9 6' Ltineburg. Whether it was actually written there or not is impossible to say, but it should be noted that even in regions where the vernacular was Low German, poetic texts of Middle German dialect were copied and understood at that time. Most of the Middle German Barlaam manuscripts, which we know mainly in the form of fragments only, do not come from the Western parts of the territory but rather from Central Germany (north of the Main River) and from the Northeastern boundaries of the Hohenstaufen empire. The best preserved representative of this category is manuscript G, formerly kept at Gotha, which proved to be a very impor tant factor for the determination of the- GEM-group. How ever, G is not the oldest manuscript of its group from which the others descended, fragments such as Fl (now in Basel) or F9 (now in Gfittingen) have probably preceded G. It is remarkable that by far the greatest number of Middle German Barlaam texts belong to the GEM-group, namely i, n, p, Fl, F3, F6, F8, F9, F10(?), F16 and F18. One of the reasons that might account for the existence of so many Middle German manuscripts is the popularity which Rudolf's Barlaam apparently enjoyed within the Order 96 See J. Koch, ZfdPh 13 (1881), 80. 201 of Teutonic Knights in the easternmost parts of the Holy Roman Empire. Inventories made around 1400 list a Barlaam manuscript in the libraries of the Teutonic Order at Konigsberg, Elbing, Osterode, and two at 97 ,a b Marienburg. Of this heritage, ¥r and K survived until the end of the Second World War at Konigsberg, and only K°—which by virtue of its close affinity to K and its dialect must have sprung from the same tradition— is still preserved intact. The other manuscripts have perished or were dispersed after the order was incorporated into Prussia; probably one or another of the above mentioned .Middle German fragments might have 98 belonged to such a manuscript. The esteem in which Rudolf's Barlaam must have been held among the influential circles of the Teutonic Order is easily understood. Since the order's mission was to carry on the ideals of the earlier Crusades, subjugate the pagan population of Eastern Europe, and convert them to the Christian faith, there was a strong need to strengthen the motivation within its own ranks. Religious literature, at the same time entertainment (miracles and saints' legends) and instruction, played an important role in that respect. Lectures and recitations from suitable works were 97 See K. Helm and W. Ziesemer, p. 34. 98 The official written language of the Teutonic Order was Middle German, see G. Ehrisrrann, Geschichte der deutschen Literatur  bis zum Ausgang des /llttelalters (Munich, 1935), II, 2.2., 67^7 202 prescribed for various occasions, such as communal meals, and the libraries were well provided with manuscripts. Not only must the ideal of asceticism in Barlaam und Josaphat have been considered very appropriate for this community of men (E. Kantorowicz characterized the Teutonic Knight as 99 a mixture of monk and warrior), but also the theme of overcoming heathen beliefs and ways of life would have appealed to them. The passage in which Josaphat's activity as a Christian ruler over a formerly pagan country is des cribed (13467-13755), reflects very well the idealized role which the Teutonic Order believed itself to play, and which was depicted in its own chronicles. Another factor, which could also account for the exis tence of Barlaam manuscripts within the Teutonic Order of Knights, is a political one. Rudolf in his later years is known to have been closely associated with the Hohenstaufen party and with Konrad IV in particular to whom he dedicated his Weltchronik. Since the Hohenstaufen rulers, from the days of Frederick II on, were the greatest benefactors of the Teutonic Order, Rudolf's position at the court would have facilitated the introduction of his work. The fact that the Koenigsberg manuscript Ka is so closely related to the Alemannic manuscripts D and dq, despite the differ ence in time and place, would suggest that a reliable 99 E. Kantorowicz, Kaiser Friedrich der Zweite (19 2 7; rpt. DUsseldorf, 1963), pp. 81-88. 203 manuscript close to Ddq might have been transmitted directly to the Order and copied there. K°, on the other hand, shows signs of a long copying process by which the text greatly deteriorated. Kb has derived from the GEM-group, and it is possible that the readings of the DK-group and of the GEM-group mingled within the Teutonic Order (Kb, as much as we know of it, has a common omission of a verse with K at 5088 and with Kc at 8341). The final major offshoots of the Barlaam manuscript tradition occur in the Southern German area again. It is not established whether Rudolf's composition lived on in the Bavarian region after its early appearance there, or whether it was reintroduced with E which drew its text pri marily from the Middle German GM-group. E, as mentioned before, was written in 1459 for a nobleman in Northern Bavaria, Veit von Egloffstein. With self-assurance, he names himself on the first page: "Das puech hab ich Veit von Eglofstain pfleger zue Vochburg mir schreibn lassen yn dem LVIIII jar der gepurt Xpi." A decade later, in 1469, the atelier of Diebolt Lauber in Hagenau (Alsace) produced the last known hand written manuscript of Rudolf's Barlaam. While we were unable to study its text, it is obvious that the emphasis was entirely on the pictorial side; the manuscript is richly illustrated with 138 full-page water-colour drawings. The commercial and yet sometimes highly artistic 204 production of codices such as this one presupposes a parti cular clientele. Diebolt Lauber's advertisements appealed to a more traditionally oriented aristocracy who still cherished the overcome courtly values and the old German literature. In a notice which Lauber addressed to his lord "hertzog Ruprecht" (who, according to Konrad Burdach, had become bishop of Strassburg in 1439 and died in 1478)}-^0 Lauber praised his illustrated copies of such works as Wilhelm von Or lens by Rudolf von Ems, Parzival and Iwein. At about the same period, the interest in classical phil osophy and literature was awakening in Southern Germany through the influence of Italian poets and scholars. The list of book acquisitions of the Heidelberg library ("Palatina") reflects this swift change in literary taste which was summed up by Burdach as follows: "um und vor 1440 steht die alte mittelhochdeutsche Litteratur auch im Sudwesten, im Elsass, in Baden, in der Pfalz bei Adel und Ftirsten in Beliebtheit und Ansehen: um 146 0 , kein Men-schenalter danach, ist sie dort bereits von der modernen, durch neue Einflusse Frankreichs, der Niederlande und Italiens bestimrnten Litteratur zuriickgedrangt and ver-X^ K. Burdach, "Die pfalzischen Wittelsbacher und die altdeutschen Handschriften der Palatina," Centralblatt fur Bib1iothekswesen, 5 (1888), 126. See Rudolf Kautzsch, "Diebolt Lauber und seine Werkstatt in Hagenau," Centralblatt fur Bibliothekswesen, 12 (1895), 5. 205 dunkelt" (loc. cit.). Furthermore the poetic attraction which the rhymed couplets with their fixed metre had held for more than two hundred years had finally waned by force of stereotyped repetition. Prose versions were taking the place of the original rhymed versions. For the Barlaam, such an anony mous prose version of the late fifteenth century is pre served in Berlin (Staatsbibliothek Preussischer Kultur besitz, MS. Germ. Fol. 1259). Another very obvious reason for the abrupt ending of the Barlaam manuscript tradition lies in the rapid spread ing of book printing. As far as we know, Rudolf von Ems' s work was never printed at that time. On the other hand, only seven years after the completion of the Diebolt Lau-ber manuscript (1469) the first incunabulum of a prose version of the Barlaam story (independent of Rudolf's text) appeared in Augsburg under the title: "HIE vahet ann eyn gar loblich unnd heylsam christglaubigen cronica. Sagend von eynem heiligen klinig mit namen Josaphat. wie der ward bekeret von eynem heyligen vatter unnd aynsideln genant Barlaam." A second printing followed shortly thereafter. While Rudolf's composition had fallen into oblivion, the Barlaam story with its characteristic parables stayed alive in Germany as a Christian legend over the following centuries. New translations from later, abridged Latin 206 versions were published, probably in connection with the Counter Reformation, around l600. But this subject did not stimulate in Germany such rich a literary output as in Spain, where several Baroque Barlaam dramas were written and produced by, among others, Lope de Vega. Performances of various Barlaam plays by Jesuit seminaries or schools are also docu mented in Germany until the middle of the seventeenth century, but these texts were usually in Latin. In the German vernacular no further adaptations of the Barlaam and Josaphat story seem to have existed in the late seventeenth and eighteenth 102 centuries, with the exception, perhaps, of a rhymed Jewish-German version. It was not until the early nineteenth century philologists and their fore runners Bodmer and Gottsched rediscovered the Middle High German Barlaam  und Josaphat, that interest in this work and its author was revived. For a short time, this rediscovery produced nearly euphoric reactions from some 101 literary scholars. J Rudolf von Ems was held in such great esteem that he was widely credited with the creation of the Hibelungenlied. Such extreme praise could not fail to draw adverse critique on the work itself and further its subsequent downgrading as "epigonal." This may partly 102 We were able to consult fragments of a Hebrew manuscript (Cod. hebr. monac. 3^7) in the transliteration by Dr. M. S. Batts. 'This interesting version does not seem to have drawn its material, mainly parables, from the known Christian sources and can therefore be disregarded in this context; see also H. Peri, pp. 23^-35. See e.g., K. Roth, Deutsche Predigtcn, p. 6. 207 explain why, after its editions in the first half of the nineteenth century, so little research has been done on it. A better founded and more balanced judgment on Barlaam und Josaphat, as mentioned before, would necessitate a new, reliable text edition. Our study has attempted to contribute to this aim. 208 List of Works Cited Batts, Michael S. "Poetic Form and Medieval Scribal Practice." Journal of English and Germanic Philo logy , 62 (1963), 697-702. Bedier, Joseph. "La tradition manuscrite du Lai de 1'Ombre. Reflexions sur l'art d'editer les anciens textes." Romania, 54 (1928), 161-96 and 321-56. Bodmer, Johann J. Chriemhilden Rache und die Klage: Zwey  Heldengedichte Aus dem Schwabischen Zeitpuncte, samt Fragmenten aus dem Gedichte von den NibeTungen  und Aus dem Josaphat. Zurich; C. Orell, 175 7. Bonath, Gesa. Untersuchungen zur Ub erlieferung des 'Parzival' Wolframs von Eschenbach. Germanische Studien, 238/239. Hamburg and Lubeck: Matthiesen, 1970/71. de Boor, Helmut. Die htifische Literatur: Vorbereitung, BlOte, Ausklang, 1170-1250. Geschichte der deut schen Literatur, vol. II. 6th ed. Munich: C. H. Beck, 1964. Brackert, Helmut.. Rudolf von Ems: Dichtung und Geschichte. Heidelberg: C. Winter, 1968. Burdach, Konrad. "Die pfalzischen Wittelsbacher und die altdeutschen Handschriften der Palatina." Central-blatt fur Bibliothekswesen, 5 (1888), 111-3 3. Czizek, Hannah. "Rudolfs von Ems Barlaam und Josaphat und seine lateinische Vorlage." Diss. Vienna, 19 31. Dain, Alphonse. Les manuscrits. 2nd ed. Paris: Belles-Lettres, 1964. Diemer, Joseph. "Kleine Beitrage zur alteren deutschen Sprache und Literatur. XI. Bruchstticke von Bar laam und Josaphat des Rudolph von Hohenems." Sitzungsberichte der Akademie der Wissenschaften, Vienna. Philosophisch-Historische Classe. XI (1853) , 650-53 . Docen, Bernhard J. Review of Barlaam und Josaphat. Ed. Friedrich K. Kttpke. Jahrbucher der Literatur, Vienna, 11 (1820), 110-38. Ehrismann, Gustav. Geschichte der deutschen Literatur  bis zum Ausgang des Mittelalters. 2. Teil: Die mittelhochdeutsche Literatur. 1935; rpt. Munich: C. H. Beck, 1959. II, 2. 209 Eis, Gerhard. "Ein neues Fragment von Rudolfs von Ems Barlaam und Josaphat." German is ch-roman i s che  MonatsschrTft, 49 (1968), 448-50. Erben, Johannes. "Zu Rudolfs Barlaam und Josaphat." Germanistische Studien, ed. J. Erben and E. Thurnher. "Innsbrucker Beitrage zur Kulturwissenschaft, 15. Innsbruck, 1969. von Ertzdorff, Xenja. Rudolf von Ems: Untersuchungen zum  httfischen Roman im 13. Jahrhundert. Munich: Wilhelm Fink, 1967. Fischer, Hermann. "Fragmente aus Barlaam und Josaphat." Germania, 30 (1885), 102-03. Gessler, Albert. "Bruchstucke einer Barlaamhandschrift." Anzeiger fur deutsches Altertum.und deutsche  Literatur, 14 (1888), 147. von der Hagen, Friedrich H. and Busching, Johann G. Literarischer Grundriss zur Geschichte der Deut schen Poesie "von den altesteri Zeiten bis in das  sechzehnte Jahrhundert. Berlin: Duncker und Humblot, 1812. Helm, Karl and Ziesemer, Walther. Die Literatur des Deutschen Ritterordens. Giessener Beitrage zur deutschen Philologie, 94. Giessen, 1951. Hofmeister, Rudolf A. "Lachmann's Role in the Transmission of Parzival." Seminar, 10 (1974), 87-100. . Kantorowicz, Ernst. Kaiser Friedrich der Zweite. 192 7; rpt. DUsseldorf: Helmut Kttpper, 196 3. Kautzsch, Rudolf. "Diebolt Lauber und seine Werkstatt in Hagenau." Centralblatt f(lr Bibliothekswesen, 12 (1895), 1-113. Klapper, Joseph. "Barlaam und Josaphat." Die deutsche  Literatur des Mittelalters: Verfasserlexikon. Ed. Wolfgang Stammler. Berlin und Leipzig: de Gruyter, 1933. I, 167-72. Koch, John. "Fragmente von Rudolfs von Ems Barlaam und  Josaphat in einer Handschrift des Britischen Museums in London." Zeitschrift fur deutsche  Philologie, 13 (1881), 78-89. K6pke, Friedrich K., ed. Barlaam und Josaphat von Rudolf  von Montfort. 2nd ed. Leipzig: Gebhardt, 1838. 210 Lachmann, Karl, ed. Iwein. Eine Erzahlung von Hartmann  von Aue. 6th ed. Berlin: de Gruyter, 1962. Leitzmann, Albert, ed. Briefwechsel der Bruder Jacob  und Wilhelm Grimm mit Karl Lachmann. Jena: Frommann, 19 27. Linke, Hansjurgen. Epische Strukturen in der Dichtung Hartmanns von Aue: Untersuchungen zur FormkrTtik, Werkstruktur und Vortragsgliederung~ Munich: Wilhelm Fink, 1968. Maas, Paul. Textkritik. 19 27; 4th ed. Leipzig: B. G. Teubner, 1960. Migne, Jacques P'. Patrologiae cursus completus. Series  latina. Paris: Gamier 1844-80. Mourek, Vaclav E. "Prager Bruchstlick einer Pergamenthand-schrift des Barlaam und Josaphat von Rudolf von Ems." Sitzungsberichte der KtinTglich Bflhmischen  Gesellschaft der Wissenschaften. Classe fUr Philo-sophie, Geschichte und Philologie, Jahrgang 189 3. Prague, 1894, pp. 1-16. Nagler, Charlotte. "Studien zu Barlaam und Josaphat von Rudolf von Ems." Diss. Karlsruhe, 1972. Neumann, Friedrich. Studien zur Geschichte der deutschen  Philologie. Berlin: E. Schmidt, 19 71. Ochsenbein, Peter. "Zwei neue Bruchstucke zum Barlaam und  Josaphat des Rudolf von Ems." Zeitschrift filr  deutsches Alterturn und deutsche Literatur, 101 (1972), 322-26. Perdisch, Adolf. Der Laubacher Barlaam: Vorstudien zu  einer Ausgabe. Marburg: Elwert, 1904. Peri, Hiram. Der Religionsdisput der Barlaam-Legende: Ein Motiv abendlandischer Dichtung. Acta Salman-ticensia, Serie de Filosofia y Letras, vol. XIV, 3. Universidad de Salamanca, 1959. Pfeiffer, Franz, ed. Barlaam und Josaphat von Rudolf von  Ems. Mit einem Anhang aus Franz Stthns, Das Handschriftenverhaltnis in Rudolfs von Ems 'Barlaam,' einem Nachwort und einem Register von Heinz Rupp. Deutsche Neudrucke. Reihe: Texte des Mittelalters. 1843; rpt. Berlin: de Gruyter, 1965. Pfeiffer, Franz. "Bruchstucke aus Barlaam und Josaphat von Rudolf von Ems." Anzeiger fur~Kunde der deut  schen Vorzeit. N.F. 2 (1854), 108-09. 211 Pietsch, Paul. "Fragment einer Handschrift von Barlaam und Josaphat." Zeitschrift fur deutsche Philologie 13 (1881) , 163-61+. Piper, Paul. Httfische Epik. 3. Teil. Nachahmer Wolframs  und Gotfrids . Klemere Epen und Chroniken. Deut sche National-Literatur, ed. J. Ktirschner. 4. Bd., 1. Abtlg. Stuttgart: Union, n.d. Re uss, Eduard W. "Beitrage zur deutschen Handschriften— kunde." Zeitschrift f ttr deutsches Altertum und  deutsche Literatur, 3 "1X843), 432-46. Roth, Karl, ed. Deutsche Predigten des XII. und XIII. Jahrhundertes. Bibliothek der gesammten deutschen National-Literatur, XI, 1. Leipzig: G. Basse, 1839 . Rupp, Heinz. "Rudolf von Ems und Konrad von Wurzburg." Per Deutschunterricht, 17, No. 2 (1965), 5-17. Rupp, Heinz. "Rudolfs von Ems Barlaam und Josaphat." Dienendes Wort: Festgabe fllr Er ns t Bender. Ed. Walter Franke. Karlsruhe: G. Braun, 19 59, -pp. 11-37. Schanze, Heinz. "Beobachtungen zum Gebrauch der Dreissiger-initialen in der Willehalm Handschrift G (cod. sang. 857)." 'Wolfram-Studien. Ed. Werner Schrttder.-Berlin: Erich Schmidt, 1970, pp. 170-87. Schirock, Bernd. "Der Aufbau von Wolframs Parzival." Piss. Freiburg, 19 72. Schnell, Rudiger.. Rudolf von Ems: Studien zur inneren Einheit seines Gesamtwerkes. Bern: Francke, 19 69. Schrttder, Edward. "Fuldaer Bruchstuck von Rudolfs von Ems Barlaam." Zeitschrift fur deutsches Altertum und  deutsche Literatur, 54 (1913), 23-24. Schrttder, Edward. "Rudolf von Ems und sein Litteraturkreis." Zeitschrift fllr deutsches Altertum und deutsche  Literatur, 67~Tl930), 209-51. Schwab, Ute. Pie Barlaamparabeln im Cod. Vind. 2705: Studien zur Verfasserschaft klemerer mittelhoch-deutscher Gedichte. Istituto Universitario Orientale di Napoli. Quaderni della Sezione Ger-manica degli Annali III. Naples, 1966. de Smet, Gilbert. "Pie AusdrUcke flir leiden im.Althoch-deutschen." Wirkendes Wort, 5 (1954) , 69-79 . 212 Stthns, Franz. Das Handschriftenverhaltniss in Rudolfs von Ems 'Barlaam.T Diss. Erlangen, 1878. Erlangen: Junge, 1878. Sonet, Jean, S. J. Le_ Roman de Barlaam et Josaphat. Vol. 1: Recherches sur la tradition manuscrite  latine et francaise. Bibliothdque de la Faculte de Philosophie et Lettres de Namur, Fascicule 6. Paris: J. Vrin, 1949. Stackmann, Karl. "Mittelalterliche Texte als Aufgabe." Festschrift fur Jost Trier. Eds. W. Foerste and K. H. Bock. Cologne: Bohlau, 1964, pp. 240-67. Stammler, Wolfgang. "BruchstUcke einer Barlaamhandschrift." Beitrage zur Geschichte der deutschen Sprache und  Literatur, 43 (1918), 554-55. Strauch, Philipp. "Fragmente von Wolframs Willehalm und Rudolfs Barlaam." Zeitschrift fur deutsches Alter-tum und deutsche Literatur, 5 2 TT910), 351-56. Vinaver, Eugene. "Principles of Textual Emendation." Studies in French Language and Mediaeval Literature  presented to M. K. Pope. 1939; rpt. Freeport, N.Y.: Books for Libraries Press, 1969, pp. 351-69. Ward, Harry L. D. Catalogue of Romances in the Depart ment of Manuscripts in the British Museum. Vol. 2. London: 189 3. Whitehead, Frederick. "The Textual Criticism of the Chan son de Roland: An Historical Review." Studies in Medieval French presented to Alfred Elwert. Oxford: Clarendon, 19 61*7 pp. 7 6-89. Wisbey, Roy. "Zum Barlaam und Josaphat Rudolfs von Ems." Zeitschrift ftir deutsches Altertum und deutsche  "Literatur, 86~TT955/56) , 293-301. Worstbrock, Franz J. Review of the reprinted edition of Barlaam und Josaphat von Rudolf von Ems. Ed. F. Pfeiffer. Anzeiger fur deutsches Altertum  und deutsche Literatur, 7 7 TT966), 111-15. Wyss, Ulrich. "Rudolfs von Ems Barlaam und Josaphat zwischen Legende und Roman." Probleme mittelhoch-deutscher Erzahlformen: Marburger Colloqium 1969. Eds. P. F. Ganz and W. Schro"der. Berlin: Erich Schmidt, 1972. 213 i Geographical distribution of Rudolf von Eras's Barlaam und Josaphat The dialect division follows in its major outlines the map "Schriftdialekte in mit-telhochdeutscher [-1350] und mittelniederdeutscher [-1600] Zeit," contained in Herrn. Paul,' Mi11elhochdeu't sche  Grammatik, 18th ed. (Tubingen, 1960) . 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items