[Reprinted from the Pitblications of the Modern Language Association of America, Vol. XVI, No. 1.] THE HOME OF THE HELIAND. - The Heliand is generally called an Old Saxon epic. Its language, however, is not a pure Saxon dialect but presents a peculiar mixture of Saxon with Frisian and Low Franconian forms, for which as yet no sufficient explanation has been offered. At a time when only two manuscripts of the Heliand were known—the Cotton ms. in the British Museum and the Munich ms. in the Royal Bavarian Library—the mixed dialect seemed to present less difficulty than at present. No special importance was attributed at this time to the traces of Frisian dialect in the poem. Most of them, in fact, were reckoned among the Early Saxon forms. The question therefore seemed to lie only between Saxon and Low Franco¬ nian ; and it is easily noticed that the traces of Low Franconian appear to a much larger extent in the Cotton than in the Munich ms. Heyne 1 accordingly advanced the theory that the Heliand was written in Munster in West¬ phalia, and that the Munich ms. preserved on the whole the dialect of the original, while the Cotton ms. represented a transcription of the original into Low Franconian. He assigned the latter to the monastery Werden on the Ruhr, near the Franconian boundary. Meanwhile the well known finds, made in 1880 in the library of the University of Prague 2 and in 1894 in the library of the Vatican in Rome, 3 have furnished us with fragments of two additional manuscripts. By these 1 Zs. f dt. Phil., i (1869), p. 288; cf. his Kleine alts. u. altndfr. Gramm. (Paderb., 1873), p. 2. * Lambel, Pin neuentdecktes Blatt einer Heliandhandschr., Wien, 1881 (repr. from Silzungsber. d. kais. Akad. d. TL'ss., 1880). 3 Zangemeister u. Braune, Bruchstiicke d. alls. Bibeldichtung, Heidelberg, 1894 (reprinted from Neue Heidelberg. Jahrbiicher , Vol. iv). 123 124 HERMANN COLLITZ. new discoveries Heyne’s opinion is definitely set aside. For both new fragments show a dialectic variety like that of the Cottonianus. This is the more important since both repre¬ sent an early stage in the tradition of the text, and are in their readings independent of each other and of the Cotton MS. The condition of our mss. then indicates that the charac¬ teristic mixture of Saxon, Low Franconian, and Frisian forms is not a peculiarity of the Cotton ms., but belongs to the original Heliand. While this view may at present be regarded as generally agreed upon, 1 there is still much difference of opinion as to the locality in which a mixture of these three dialects could have taken place. Several scholars, Koegel 2 and Braune 3 for example, are satisfied with a slight modification of Heyne’s theory. The poem in their opinion was written inWerden, the same monastery which Heyne regarded as the home of the Cottonianus. Kauffmann 4 would prefer to substitute for Werden the monastery of Corvey on the Weser. This would carry us near the southern part of the Saxon territory. Jostes 5 —in a paper which has much stimulated and certainly in some respects benefited the discussion of our problem— finds that for creating an epic like the Heliand conditions were nowhere more favorable than in the northern provinces of the empire, say near Hamburg or in Holstein. As regards our manuscripts, he suggests that the Cottonianus may have been written in Magdeburg, the Monacensis in 1 Cf., e. g., Koegel, Gesch. d. dl. Lit., i, 1, p. 281; Braune, Bruchst. d. alts. Bibeldicht., p. 212. 2 1. c., p. 283 seq., and Erg. heft, p. 21 seq. H. c., p. 220. 4 Germania 37 (1892), p. 368 seq., in a review of Gallde’s Alts. Gramm., written before the Vatican fragments were discovered. In P.-B. Beitr. 12 (1886), p. 358, Kauffmann advanced the opinion that the Cottonianus was written in Eastern Westphalia, and that Paderborn might have been the home of the poet. 5 Zeitschr. f. dt. Alt. 40 (1896), p. 160-184. Cf. H. Tiimpel, Niederdt. Studien (Bielefeld, 189b), p. 130—133. THE HOME OF THE HELIAND. 125 Hildesheira. Finally Wrede, in an able treatise published only about a year ago, 1 has attempted to prove that the poet lived in the southeastern corner of the Saxon territory, in the vicinity of Merseburg. There is good reason to believe that there existed near Merseburg in the Old Saxon period a Frisian colony, since unmistakable traces of Frisian dialect appear (in Low German glosses and in proper names) in this vicinity as late as in the first quarter of the eleventh century. 2 This in Wrede’s opinion would account for the Frisian elements in the Heliand. As regards the supposed Low Franconian forms, Wrede holds that these are not Low Franconian but belong to Eastern or Southeastern Low German. We see then that in this question the East and the West, the North and the South have each found its advocates, and it is for us to take our choice; unless we decide to reject every one of these theories in order to start in a new direc¬ tion, a direction not indicated by any one of the four points of the compass. The fact that one theory has closely followed another, seems to indicate that the proper solution of the problem has not yet been found. Under these circumstances I may refrain, I think, from discussing in detail the different propo¬ sitions and from repeating the objections which each advocate of a new solution has raised against his immediate prede¬ cessor. Let it suffice to consider briefly the latest of the above theories, the one by Wrede. Wrede starts with an argument, in which he follows Jostes and which, at the first glance, seems quite plausible. It is a well known feature of the language of the Heliand that the word burg is often appended to names of foreign cities, 1 Zs. /. dt. Alt. 43 (1899), p. 333-360. Cf. Roethe, “ Heliand und Sachsen- spiegel,” in the Anzeiger of the same vol., p. 387-390. 2 See especially H. Hartmann, Grammaiik d. allesten Mundart Merseburg's. I. (Dissert.) Norden, 1890. 30194 126 HERMANN COLLITZ. so as to form compounds like Nazarethburg, Sodomoburg r Rumaburg} Such names are later on especially common in the eastern part of the Saxon territory (that is to say,, in the district in which at present names like Magdeburg r Blankeyiburg, Quedlinburg , Merseburg , Naumburg are found), although similar names (e. g., Luneburg , Hamburg ) occur also in Northern and sometimes (e. g., Oldenburg , Nienburg , Duisburg) in Western Saxony. These facts in Jostes’s opinion serve as an argument in favor of regarding Hamburg or it& vicinity as the birthplace of the Heliand , while Wrede con¬ tends that the poet more probably lived in the ftitr^-district proper (the “Gegend der Burgwarde”) near Merseburg. There is at the outset a slight chronological difficulty^ We happen to know that Quedlinburg was founded by Henry the Fowler, who reigned from 919-936, and most ot the towns in -burg are perhaps not much older. In fact,, the earliest document in which a considerable number of such names are mentioned is a deed by the emperor Otto II, which dates from May 20, 979 (see Wrede, p. 335). Wrede indeed maintains that a similar list of names (from the abbey of Hersfeld), written toward the end of the eleventh century r is a faithful copy of the original, which belonged to the last third of the ninth century. I am unable to examine the latter statement and am willing to accept it on good faith. But even this would carry us only to a time half a century later than the date of the Heliand. Wrede goes on arguing that with the aid of the Heliand we are able to date the East Saxon towns with -burg farther back: the “ Heliandburgen’ r constitute the earliest testimony for their existence, and judg¬ ing from the Heliand such names were current [N. B. in Eastern Saxony] a century before the original of the Hers¬ feld document was written. But are we not here entirely losing the ground under our feet? If the existence in Eastern Saxony of towns in - burg is warranted for the end 1 Cf. Jostes, 1. c., p. 164. THE HOME OF THE HELIAND. 127 ■of the eighth century only by names like Rumaburg in the Heliand, how can we, without committing a circulus vitiosus, admit that only in Eastern Saxony could the poet have found his models for such names? Moreover, Wrede is apparently not aware of the well known fact 1 that in Old Frisian laws - burch is sometimes added to names of cities in the same manner as in the Heliand , e. g., Colnaburg or Colene = Cologne, as in the He¬ liand Rumaburg (dat. Rumuburg ) or Ruma = Rome. 2 These laws were written not in Eastern Saxony, but in the Frisian country between Bremen and the Netherlands. With reference to the Frisian Colnaburch Siebs (/. c.) has argued against Jostes that the names with -burg are not of much account as to the origin of the Heliand. Judging from Goth, baurgs ‘ town/ O. Norse borg, A.-S. burg, etc., this word was in the Old Germanic dialects the general designa¬ tion for Residence’ or ‘town/ In the Heliand it is added in rather loose composition to the names of foreign cities, in order to relieve somewhat their foreign appearance. Simi¬ larly the poet adds land to the names of foreign countries (e. g., Aegypteo-land, Galilea-land or Galileo-land, Kananeo- land ., Ponteo-land), strom to the names of foreign rivers (Jordana-strom or Jordanes-strom, Nil-strom), folic or liudi to the names of foreign peoples (Ebreo-follc, Ebreo-liudi, Judeo-follc, Judeo-liudi, Romano-liudi). With reference to the origin of the poem, there is no warrant for putting more stress on names with burg, than on those with land or strom, etc. If further confirmation of this view be required, it may be found in the fact that also in Anglo-Saxon poetry the term -burg is used, exactly as in the Heliand, in coined words and added to foreign names. E. g., Finnsburuh (Battle of Finns- 1 Cf. Richthofen, Altfries. Worterbuch , s. v. burch; Koegel, Gesch. d. dt. Lit., I, 1, 244; Siebs, Ztschr.f. dt. Phil., 29, 413. 5 Richthofen, Fries. Rechtsquellen (Berlin, 1840), pp. 3 and 4: Colnaburch het bi aide tidem Agrippina (Emsigo ms.) = Colnaburch hit bi alda tidon Agrip (Riistringer ms.) = Colene het bi aide tidem Agripina (Hunsigo ms.). 128 HERMANN COLLITZ. burg 38), Mceringa burg (Door’s Complaint 38), Romana burg (Boet. Metr.y ix, 10), on Romebyrig (Fata Apost. 11), Troia burg ( Boet. Metr., ix, 16 and xxvi, 20), Sodome burh (Gen. 1975), on (or of) Sodoma byrig (Gen. 1925, 2013, 2558), Aethanes byrig (plur., Exod. 66), in Caldea byrig (Dan. 95), Babilone burh (Dan. 601), Babilon burga (plur., Dan. 694), on Sione by rig (Psalm lxxvii, 67). It is quite probable that the agreement of Frisian, Anglo- Saxon, and Old Saxon in this peculiarity is not incidental, but inherited from an earlier stage of West Germanic poetry. But it certainly disposes of Wrede’s conclusions, since we cannot very well assume that, e. g., the author of the Battle of Finnsburg or Cadmon lived near Merseburg. As regards Wrede’s grammatical arguments, they are scarcely more convincing than the one based on the use of -burg. To be sure, his treatise is ingenious and brilliant, and contains much valuable information, derived especially from the comprehensive map of German dialects, at which he is working in conjunction with Dr. Wenker. But as to the main issue he has followed a wrong track, and the result is a theory whose shortcomings even his skilful treatment is unable to disguise. Wrede, e. g., endorses (p. 342) Jostes’s view as to the form fon. Jostes wrote in the Zs. f. dt. Alt ., 40, 173: “ In my opinion the one little word von may suffice to show (as against the reasons advanced for Westphalia) that the home of the poet must have been in the East.” We are assured by Wrede that this view is confirmed by the map of German dialects, and that according to the same source and in keeping with Wrede’s theory von and van are both found to-day (just as they are found alternating in the Heliand ) in the principality of Anhalt to the right of the Saale and further on beyond the Elbe. Wrede also states that van is the North Frisian form, and finally refers to Tiimpel’s Niederd. Studien , p. 11 seq. He does not inform us that both fon and fan occur in THE HOME OF THE HELIAND. 129 the Old Frisian laws 1 and that fon is the current form in Saterland Frisian. We may reckon fon in the Heliand among the Frisian forms, or we may assume with Holt- hausen 2 that in Low German originally both fan (accented) and fon (unaccented) were found. The latter theory is perhaps recommended by the fact that fon occurs in Middle Low German too frequently to be explained (as Tiimpel proposes) simply by the influence of High German. In any case the little word von is not entitled in this question to the prominent place which Jostes and Wrede are willing to bestow upon it. Wrede assumes that the Heliand originated in a part of Germany in which Low German is no longer spoken to-day. He consequently reconstructs the dialect of what he regards as the home of the poet, with the aid of the neighboring Low German and Midland German dialects. Since Frisian, as we have seen, was probably at some time also spoken in the same vicinity, the result is a reconstructed dialect from which Wrede is able to produce almost any variety of dialectic forms, whether commonly called Saxon, or Frisian, or Franconian. And yet, this remarkable dia¬ lect—or rather combination of dialects—does not account for some of the most notable peculiarities of the Heliand. Not, e. g., for a number of preterits in st, which deserve our attention the more since they are not mentioned by either Jostes 3 or Wrede. l fon in the Rustringer, Brokmer, Emsigo, Fivelgo, and Hunsigo MSS., fan in the two printed texts from Westerlauwer Friesland. See Richthofen, Allfries. Worterb. s. v. fon. 2 Allsdchs. Elementarbuch, \ 127. 3 Jostes (1. c., p. 77) says: “ The number of reasons therefore which point for the origin of the Heliand toward the East is quite considerable, whereas such as would speak for the West do not in reality exist” (“ wahrend solche, die fur den Westen spreehen , in Wirklichkeit gar nicht vorhanden sind”). This statement, it seems to me, would be more correct if Jostes had reversed the terms East and West. 130 HERMANN COLEITZ. The equivalent of Engl. ‘I could 9 is in the Heliand konsta , subj. kunsti or konsti. Similarly we have from the verb unnan Ho grant ’ the preterit onsta , and from far-munan Ho disdain * the preterit far-munsta or far-monsta. Such pre¬ terits occur only in the Low, Middle, and Rheno-Franconian dialects. They are not used in modern literary Dutch, where the preterit of ik lean is, in the written language, ik konde or ik kon. But their modern offshoots 1 are found in Belgian and Dutch dialects, and on the borderline between the Nether¬ lands and Germany south of a line connecting Leiden with Uddel in the Veluwe (near Utrecht) and running from there to Muhlheim on the Ruhr. These preterits are not, as is sometimes assumed, old forms, but are new formations, shaped after the analogy of the preterit dorsta which belongs to the old verb dorsan ‘ to dare.’ The old and genuine forms are found in Goth. kuri\>a , A.-S. cu\>e , MHG. kunde; in A.-S. wfe, MHG. g-unde ; and in Goth, munda , A.-S. munde 2 Here then we have in the Heliand an unmistakable trace of Franconian dialect, and one on which the more stress is to be laid since these preterits are found ki our mss. —as far as the st is concerned—without a variant. 3 If the preterits in - st - are Franconian and cannot be anything else, there is no reason to abandon the derivation from the Franconian dialect of the diphthongs uo and ie (e. g., in muodar mother = Sax. modar , or in hie he = Sax. he) in favor of the one suggested by Wrede (p. 342). Nor can I regard Wrede’s complicated hypothesis as to mi and 1 Viz., forms like ik kos or kosl ‘I could,’ plur. kossen or kosten (subj. kos, pi. kosten) and ik begos ‘ I began.’ 3 See on the above preterits my introduction to Bauer’s Dictionary of the Waldeck Low German dialect (which is to appear within a few months in the series of dictionaries published by the Low German Dialect Society), p. 69* 3 It happens that no preterit of kunnan, unnan , or munan occurs in the Prague or Vatican fragments. But since Cottonianus and Monacensis are, as to the st, in complete harmony, there can be no doubt that the sHorms belong to the original. THE HOME OF THE HELIAND. 131 mile as an improvement oh the simple explanation given recently by Tumpel. 1 Onr result then is that the language of th ^Heliand points to the Western part of the Saxon territory, or rather to that part of Germany where from the earliest times we find the Low Franconian, Frisian, and Saxon dialects in close proximity. But the difficulty begins as soon as we attempt to identify the dialect of our poem with that of a particular locality. For, although the three dialects have been neighbors for many centuries, there exists nowhere now, and as far as we can see there has never existed, in actual speech, such a com¬ bination of various features from the three dialects as is found in the Heliand. The difference between the Heliand and the spoken dialects is seen, e. g., in the pronoun ‘ other/ which in the Heliand form is othar. 2 This form is identical with Old Frisian other , and is characterized as Frisian (or Anglo-Frisian) by the change of the original group an\ to oj?. The original sounds, short a followed by a nasal, are preserved not only in Gothic an\ar , but also in the modern Low Franconian and Low Saxon dialects, where we find ander (or in some dialects anner or arj e r). There is no modern dialect to warrant the opinion that the pronoun dthar was ever found in a district in which the preterit of leunnan is Iconsta. The area of these forms is at present separated by a neutral zone in which neither the st of Iconsta nor the long 6 of dthar occur. I have l Niederd. Sludien, p. 131. 8 dthar is both in C and in M by far the most frequent form. In M it occurs, according to Schmeller’s Glossar . Saxon., 91 times. The regular Low German form andar (which however occurs, besides dthar, also in Old Frisian) is found only in two instances ( andran 1263, ander 1444) in C alone, and cannot be ascribed to the original. A third form athar or adar, which occurs twice in C ( athres 1478, adron 1536), three times in M (adrurn 1271, athrana 1434, adrom 2985), and once in Gen. (a'Sur 211), looks like a combi¬ nation of the two other forms and is perhaps merely a graphical variant of dthar. 132 HERMANN COLLlTZ. mentioned before that preterits developed from Iconsta are found south of a line which connects Leiden with Utrecht and Muhlheim. Here the pronoun ‘ other’ is at present generally arfr. „ North of this line we have a belt of dialects in which the nth of Goth. an\ar and kuri\>a has become nd, as in Dutch ander and wij konden. Finally we meet further north with the Frisian dialects, in which the n is in both forms lost before the following spirant, as in English ‘other’ and ‘ I could;’ e. g., Modern West Fris. oar ‘other’ and ik hoe ‘ I could.’ The difficulty cannot be solved by asserting that at the time of the Heliand there may have existed between Frisian and the present northern boundary line of the preterits with st a dialect which combined the forms honsta and othar. If konsta had ever extended northward into Frisian territory, this would have led in Modern Dutch to a preterit hos or Icoste instead of kon or konde. Nor can othar have extended southward beyond the boundary line of the preterits with sf y because this again would be incompatible with the existence of konde in Modern Dutch. For the same phonetic law which has done away with the nasal in the pronoun ari\>ar would have applied to the nasal in the preterit kon]>a (Goth. kun]>a). Regularly then the preterit konde goes together in Dutch with ander , as in Middle High German and Middle Low German kunde with ander; and on the other hand in Modern Frisian hoe (= Old Fris. *kuthe) with oar (— Old Fris. other), as in A.-S. cu\e with o\er y and in English ‘ could ’ with ‘ other.’ 1 But why not assume that the mixed dialect of the Heliand is due to various scribes or perhaps to a compromise between x As regards the former boundary between Franconian, Saxon, and Frisian, I may refer to K. v. Richthofen’s map, “ Friesland im 9. Jahrh.,” in his Unlersuchungen zur friesischen Rechlsgeschichle, Vol. 2 (also published sepa¬ rately in Zivei Karlen von Friesland im 9. und im. IS. Jahrh., von K. v. Richthofen. Berlin, 1882). Maps of the modern Dutch dialects are found in Jellinghaus, Die niederland. Volksmundarlen (Norden, 1892), and in Paul’s Grundriss d. german. Philologie, Vol. i, 2nd ed. (Nr. 4, Strassb., 1899). THE HOME OF THE HELIAND. 133 the dialect of the poet and that of his scribe? We might say, e. g., that a Saxon poet, not versed in the art of writing, availed himself of the assistance of a Frisian scribe, who perhaps lived on Franconian soil, or whose manuscript was soon afterwards copied by a Franconian. The chief objection to this or similar views is the fact that a mixed dialect, closely resembling that of the Heliand , is found in various other ‘ Old Saxon’ writings, e. g., in the fragments of a Commentary to the Psalms 1 and in the Essen Confession. 2 In both the characteristic Frisian other is found ( 5ther[imu] Ps., dthra Conf.); and in the Conf. there occurs the Franco¬ nian preterit bigonsta, while in the Comm, to the Psalms the Saxon d (e. g., in tote) is generally replaced by the Franconian diphthong uo (e. g., tuote , guodlica , bluodo , fuoti , duonne). Similar forms might be quoted from other ‘Old Saxon’ texts, e. g., from several of the manuscripts which contain Old Saxon glosses. It is scarcely probable that all these different texts should have been written under similar conditions and should presuppose the same complicated situation : an author unacquainted with writing, and a scribe who made it a point to write in three different dialects:—his own, that of the author, and a third which was neither his nor the author’s. Even if we modified the theory so as to limit the activity of the first scribe to two dialect^ and make for the third dialect a set of later scribes responsible, as a steadily recurring combination this would not appear credible; nor does it agree with what we know of the circumstances in which some of these texts were written. 3 There seems to remain then only one possibility. We shall have to acknowledge in the language of the Heliand a mere literary and artificial mixture of dialects, similar to the com- 1 E. Wadstein, Kleinere altsachs. Sprachdenkmaler (Norden, 1899), Nr. ii. *lbid., Nr. hi. 3 E. g., the Confession was written in a Westphalian convent (Essen); see Wadstein, 1. c., p. 124. 134 HERMANN COLLITZ. bination of Low Franconian with Middle High German in Veldeke’s poetry, or to that of Aeolic with Ionic and other Greek dialects in the Homeric poems. Such a blending of different dialects is in no case merely arbitrary. As a rule it is rather forced upon the poet by circumstances, and is generally due to a compromise between the dialect of the poet and that of his public, or more frequently that of an inherited poetry. In the latter case the mixture of dialects generally furnishes a valuable aid for tracing the different stages through which a certain species of poetry has gone. In case, e. g., of the Homeric poems the mixture of Aeolic, Ionic, and other dia¬ lects indicates that epic poetry was first developed among the Aeolic tribes in Asia Minor, that from these it passed to the neighboring Ionians, and afterwards to the Greeks of the islands and of the continent. It seems to me that similar conclusions may be drawn from the language of the Heliand . For the Heliand belongs only to the latest stage in the development of Early Germanic epic poetry. The poet may have drawn on the heathen poetry of his people not only for his metre and rhythm, his style and his vocabulary, but also for his dialect. Not he then but the Old Germanic heroic poetry would be responsible for the admixture of Frisian and Franconian. We might claim that this view’ was possible, or probable, even if there existed no remains of an earlier poetry with which to compare our poem. Yet we are fortunate to possess, in the song of Hildebrand and Hadubrand, at least one fragment of German heroic poetry from the time before the introduction of Christianity, and in this fragment we meet with a mixed dialect quite similar to that of the Heliand. We need not concern ourselves here with the controversy whether this lay was originally composed in Low r German or in High German. Nobody will deny that in its present shape its language forms a combination of the two dialects, and it suffices for our purpose that its ‘ Low German ’ ele¬ ments show significant Frisian (or Anglo-Frisian) in addition THE HOME OF THE HELIAND. 135 to the Saxon forms. E. g., the word for ‘ other ? is in the Hildebrandslied oder (1. 12, ibu du ml enan sages , ik ml de odre uuet ‘if you tell me one, I know the others'); Mod. Germ. hund is child (= O. Fris. Jcuth); O. High Germ, gund ‘ com¬ bat ’ is gud or gift (= A.-S. gift). 1 If the Hildebrandslied is a Low German poem, copied by a High German scribe, its language furnishes immediate proof of the existence in Low German poetry of Frisian forms. If it be a High German poem, transcribed (with frequent traces of its original dialect) into Low German, the conclusion would be that the Frisio- Saxon dialect in which it was clothed, was that of Low German heroic poetry. 2 In either case the mixture of Frisian and Saxon form appears as a significant feature of heathen poetry in Northern Germany. Whether Low Franconian forms occurred in the Hilde¬ brandslied to the same extent as in the Heliand it is impossible to decide. Since Low Franconian resembles in its conso¬ nantism the Old Saxon, in its vocalism the High German language, the Low Franconian forms cannot as a rule, in a text like the Hildebrandslied , be distinguished from those 1 The loss of n before th is generally regarded as a peculiarity of Saxon as well as of Anglo-Frisian, and in every Old Saxon grammar (e. g., Holt- hausen’s recently published Allsachs. Elementarhuch , $ 191) words like othar , soth, kulh are quoted as genuine Saxon. Yet in Middle Low German and in the Modern Low German dialects only the word for ‘south’ (MLG. suden ) has this syncope, and here it is shared by Middle High German. The phonetic law, therefore, which does away with n before th, is not Saxon but Frisian. Cf. Bauer’s Wald. Wtb. (see above, p. 130, note), p. 70* seq., and Bremer in Paul’s Grundriss , m 2 , p. 866. 2 The former alternative seems to me the more probable, and I trust that the theory set forth here may perhaps serve to weaken some of the objec¬ tions which have been raised against Koegel’s views (Paul’s Grundriss, II, 1, p. 175 seq. of the first edition). We may, e. g., readily admit that the vocabulary of the Hildebrandslied agrees as much with Anglo-Saxon as with Old Saxon (see especially F. Kauffmann in Philolog. Sludien, Festgabe fur Sievers , p. 127 seq.). Considering the near relationship of Anglo-Saxon and Frisian this would not militate against Old Saxon origin, if we assume that Old Saxon heroic poetry preserved largely the vocabulary of its- Frisian models. 136 HERMANN COLLITZ. which exhibit a mixture of Low German consonantism and High German vocalism. E. g., the diphthong uo in words like cnuosfes or muotti may be regarded as Low Franconian, or it may be in line with the High German ch in chud or the t in gihorta and many other examples. There is, however, as far as I am aware, nothing in the Hildebrandslied to contra¬ dict the opinion that its 1 Low German ? dialect compares as to the Low Franconian elements with that of the Munich manuscript of the Heliand. Our manuscript of the Hildebrandslied was probably written between the years 809 and 817, 1 while the song itself is probably at least half a century older. The Heliand may be dated, in a round number, about 830. It follows then that there existed previous to the time of the Heliand an epic dialect, characterized by the same mixture of Low Saxon with Frisian—and, we may add, probably Low Franconian— elements. Thus the problem which the mixed dialect of the Heliand offered, is shifted back to the history of Early Germanic epic poetry, and it seems to me that on this ground we are able to arrive at a satisfactory solution. For several centuries Germanic heroic poetry flourished especially among the Franks. To the Franks is due, more than to other Germanic tribes, the development of the great and complicated legend of the Nibelungen, whose historical elements incorporate (in the characters, e. g., of Dietrich and of the Burgundian kings) earlier Gothic and Burgundian traditions, while its mythical elements (viz. that part of the story which centres around the characters of Brunhild and Siegfried) seem to rest chiefly on Frankish or more particu¬ larly Rhinefrankish 2 legends. We are told that Charles the 1 These dates have been ascertained by F. KaufFmann in Festgabe fur JSievers, p. 136 seq. 2 Cf. Sijmons, in Paul’s Grundriss, ill*, p. 656. KaufFmann has recently (Zs. f. dt. Phil. 31, 1899, p. 5) suggested that the Siegfried legend may have been combined with the story of the Burgundians as late as in the tenth THE HOME OF THE HELIAND. 137 Great had the epic songs of the Franks written down. But the interest in these songs seems not to have been as strong during Charles’s reign as formerly; and a century afterwards, at the time of the monk Otfried, they were completely for¬ gotten,—for Otfried 1 tells us that the Franks have no poetry and that their language is not accustomed to the restraint of metre. Meanwhile, however, the main body of Frankish heroic legends had found their way to the Northern countries, where they were embodied later on in the collection of alliterative songs which is familiar to us under the name of the JEdda. Opinions differ as to the exact line on which the migration of these legends proceeded. But this much is certain that we have to distinguish in the Norse tradition at least two different layers, an earlier and a later one. As regards the latter there is no doubt that it is based on Low German sources and reflects the form in which the legends were current in Northern Germany at the end of the ninth or in the first half of the tenth century. It is probable, however, that also the earlier set, which seems to belong to the eighth century, goes back—directly or indirectly—to a Low Saxon source. 2 The share which fell to the Saxons in the cultivation of epic song, reminds us of the part which they played at the end of the middle ages in the propagation of the beast epic. The Low German ReinJce de Vos , destined to become the most popular form of the beast epic and the source of numerous translations, was nothing more than a skilful translation of a century. His chief reason is that the obvious diversity in character be¬ tween the two ought to prevent us from dating their union too far back. But do the two differ more fundamentally than the mythical and the historical elements in the Beowulf epic? It seems to me that stronger reasons would be required to convince us that a combination which here¬ tofore has been regarded as one of the characteristic features of Early Germanic epic poetry, could militate against an early date. 1 Liber Evangeliorum, i, 1, 33-36. 2 See for the particulars Sijmons, l. c., pp. 632 and 663. 138 HERMANN COLLITZ. Flemish work. Similarly most of their heroic songs appear to have been mere adaptations from those of their western neighbors. For with the exception perhaps of the legend of Wieland the blacksmith, which is with some probability claimed as Low German, 1 there is apparently not a single subject in the earlier heroic legends which could be regarded as originally Saxon. This lack in originality is easily ex¬ plained, if we assume that the Saxons became acquainted with the epic poetry of the Franks at a comparatively recent date, when the principal legends had obtained their definite poetic garb. Not so their western neighbors, the Frisians, in spite of the unjust saying Frisia non cantat —which we may confidently change into Frisia cantat , or at least Frisia cantabat —and in spite of the unfortunate fact that not a single alliterative poem has been handed down in pure Frisian dialect. 2 3 Frisian heroic poetry has left its traces in Anglo-Saxon epic songs. It is generally admitted that the fragment of the Battle of Finnsburg and the Finn-episode in Beowulf are derived from a Frisian source. But we are allowed to go further and to maintain that whenever subjects from conti¬ nental epic poetry are met with in Anglo-Saxon poems, the 1 Sijmons, 1. c., p. 725. I should like to say, however, that even in this case the evidence of Saxon origin is far from being conclusive. It is true that in most of the later versions the scene is laid in Westphalia. But there remains the possibility that the legend was fixed only later on in a certain locality, or that the scene was changed to Saxony. In the earliest version (Deor’s Complaint) there is no indication of Saxon origin, and even in the V0lundarkvitha the local names are partly fictitious. I do not see why under these circumstances the legend should not have originally been Khinefrankish or Frisian. [I have not been able to consult the recent discussion of the Wieland legend by Jiriczek in his Deutsche Heldensagen .] 3 From alliterative formulas, which occur frequently in the Old Frisian laws, Koegel, Gesch. d. dt. Lit., i, 1, 242 seq., has attempted to reconstruct portions of a Frisian legal poetry. We need not follow Koegel in these experiments. But we may justly hold with Miillenhoff ( Beovulf , p. 105) that the important part which alliteration plays in the legal prose of the Frisians, favors the view that it had also taken a firm hold of their poetry. See on this question especially Siebs in Zs. f. dt. Phil. 29, p. 405 seq. THE HOME OF THE HELIAND. 139 immediate sources were as a rule Frisian poems. 1 Among the texts which come under this point of view, belong espe¬ cially the fragments of Waldere , the account of Siegmund’s heroic deeds in Beowulf (1. 875 seq.), and Deor’s Complaint. As regards the Waldere fragments, I agree with Learned 2 that they are based on an early ‘ Low German ? version of the legend. Learned is inclined to ascribe this version to the Saxons, although he himself is in doubt as to this point. Waldere certainly differs somewhat from the later Saxon tradition, which is found in the Thidrekssaga and which in Mullenhoff’s opinion 3 goes back to a Frankish source. Matters may perhaps be adjusted if we assume that Frisian poems formed the connecting link between the continental and the Anglo-Saxon version on the one hand, and between the Frankish and the Saxon form on the other hand. Of the passage on Siegmund in the Beowulf and of Deor's Complaint we may say that they represent a peculiar Anglo- Saxon or Anglo-Frisian development of legends which apparently took an intermediate position between the early continental and the later Norse tradition. There is, therefore, at least some probability that here, as in the case of Waldere , the source of the Anglo-Saxon songs is to be sought in Frisian tradition. The influence of Frisian heroic poetry is furthermore noticeable in the Middle High German popular epic. The well known poem of Gudrun, next to the Nibelungenlied the most important popular epic in Middle High German, is derived from Frisian heroic poetry and preserves the traces of its origin in its scenery, its principal characters, and in the very name of Gudrun. 4 For the genuine High German form of this name is Gundrun or K.undrun ) while Gudrun (= Guftrun) points to a dialect in which n was lost before a following p, with compensatory lengthening of the preceding 1 Cf. Miillenhoff, Beovulj , pp. 104-108. 3 Publ. of the Mod. Lang. Assoc., vii (1892), pp. 181-185. 3 Zs.f. dt. Alt. 12, p. 273 seq. 4 See Miillenhoff, Zs. f. dt. Alt. 12, p. 315, and Sijmons, 1. c., p. 716. 140 HERMANN COLLITZ. vowel, just as in 5]>ar, gvft-hamun ( Hildebrandslied ), and in the other examples discussed above. Finally it is of interest in this connection that the only North German rhapsodist whose name has been handed down to us from the time of Charles the Great was a Frisian. His name was Bernlef, and he was a friend of the Frisian bishop Liudger (■)■ 809), the well known founder of the monastery of Werden on the Ruhr. 1 The above data, how r ever few in number, allow of the interpretation that in heroic poetry—or at least in certain branches of heroic poetry—the Frisians were the pupils of the Franks and later on became the teachers of the Saxons. Looked upon in this light, the Frisian and Franconian forms 2 in the Heliand (as in the Hildebrandslied) find their natural explanation in the language of Saxon epic poetry, which in its dialect preserves the traces of its earlier history. We cannot in these circumstances draw from the lan¬ guage of the Heliand any definite conclusions as to the home of the poet, just as we cannot tell from the language of the Homeric epic to which of the seven cities belonged the honor of having produced a Homer. This much may be said, however, that more general reasons—e. g., the close relation of Saxon to Frisian poetry, and the fact that most of the Old Saxon literary productions come from the Western part of the country—point to Western rather than to Eastern Saxony. Hermann Collitz. 1 See on Bernlef especially Mullenhoff, Beovulf p. 105, and Koegel, Gesch. d. dt. Lit. i, 1, 141 seq. and 283. 3 More exactly: those Frisian and Franconian forms which belonged to the original text of the poem and are accordingly found in most of our mss. The preponderance of Franconian forms in V calls for a different explanation. If we may assume with Mullenhoff ( Denkm ., l 3 , p. xxvii seq.; cf. Koegel, Gesch. d. dt. Lit., i, 2, p. 558 seq.) that Kheno-Franeonian was spoken at the Carlovingian court, it seems possible to suggest that perhaps a copy of the poem was rewritten in Franconian dialect (without, however, effacing every trace of Saxon and Frisian) for the emperor Ludwig the Pious, and that from this manuscript the Vatican fragments were copied.