ALLITERATION IN THE CHANSON DE ROLAND AND IN THE CARMEN DE PRODICIONE GUENONIS 3l]f MitinrrHtty of Kl)iragn ALLITERATION IN THE CHANSON DE ROLAND AND IN THE CARMEN DE PRODICIONE GUENONIS A DISSERTATION SUBMITTED TO THE FACULTY OF THE GRADUATE SCHOOL OF ARTS AND LITERATURE IN CANDIDACY FOR THE DEGREE OF DOCTOR OF PHILOSOPHY DEPARTMENT OF ROMANCE LANGUAGES BY Alexander Haggerty Krappe IOWA CITY, IOWA 1921 f4 11 K7 THE ATHENS PRINT SHOP Iowa City, Iowa TO Dr. Franz Nobiling ALLITERATION IN THE CHANSON DE ROLAND AND IN THE CARMEN DE PRODICIONE GUENONIS The occurrence of alliteration in Romance prose and verse has for years been a well established fact 1 . Most scholars who have dealt with Old French alliteration have considered only that of coordinated words 2 . For the purpose of finding- alliterating formulas of genuinely popular character, this method of procedure is doubtless the right one, since a com- parative study tells us that popular alliteration is found with coordinated words only. Kohler admits that intentional use of alliteration can be found outside this group 3 , and wherever alliteration occurs as a poetic device, in Greek, Latin, Teutonic, and Provencal poetry, it is not restricted to coordinated words. What is needed, therefore, is a study on the use of alliteration as a means of poetic technique in at least one monument of Old French literature, and for many reasons the Chanson de Roland deserves the first place in our attention. A discussion of the term alliteration, though necessary, may be dispensed with here, since its different definitions have re- ceived adequate treatment in most of the studies mentioned in the bibliography 4 . I take, then, as a basis the definition of Plautinus : Fitalliteratioquoties dictiones continuatae vel binae vel ternae ab iisdem primis consonantibus, mutatis aliquando vocalibus, aut ab iisdem incipiunt syllabis, aut ab iisdem primis vocalibus. To this must be added some supplementary remarks and modifications. Riese, Kohler, and Kriete, following the ex- ample of Wolfflin, went on the assumption that each vowel alliterates only with itself, as a with a, o with o, etc., as in Latin, contrary to Teutonic usage. Scholz pointed out objec- tions to this rule, but added that the Teutonic use of allitera- 1. cf. the works of Fuchs, Wolfflin, Paul Meyer, Grober, C. Michaelis, Riese, Kohler, Kriete, and Scholz. 2. Riese, p. 16; Kohler, pp. 92 and 100. o. p. 100. 4. Riese, pp. 6-8; Kohler, p. 92; Kriete, p. 10; Scholz, pp. 385-92. ALLITERATION IN THE ROLAND tion of all vowels with each other is probably due to the fact that each vowel was preceded by a sort of aspirate h in the Teutonic languages. 1 I do not see any possibility of test- ing whether the occurrence of two or more words, each beginning with a different vowel, is due to mere chance, or whether it must be considered as intentional alliteration. It may be observed, however, that the feeling for quan- tity and quality of vowels was so well developed that at the time of the Roland p did not assonance with o nor e with e or e. 2 There would be a certain contradiction in the principles of versification, should we assume that this feeling was devel- oped to such an extent with regard to the vowels in assonance and was absent altogether with regard to the vowels in alliter- ation. I shall use, then, as a basis for the consideration of vowel alliteration the same principles which are followed in the assonances of the poem, so that, for instance, e alliterates only with e, e with e, etc., bearing in mind, however, that this is an a priori assumption which can neither be proved nor disproved. The study will have a phonetic basis, as at that early time alliteration was supposed to have an effect on the ear, not on the eye. Thus gente and jamais would alliterate, in spite of the different spelling, but not gente and grant. Becq de Fou- quieres justly distinguishes between alliteration forte and al- literation faible, calling an alliteration "strong" if the two al- literating sounds are perfectly identical ; "weak," if the one is voiced, the other voiceless. The latter of these two groups does not appear to play such an important part in Old French as it does in Provencal ; it has therefore been left out of con- sideration in this study. 1. Zeitsch. f. rom. Phil. XXXVII, 1913, 388. 2. An apparent exception to this rule we find in the alliteration, Huns : Hongres, Rol. 3254. There can be no doubt but that initial h was silent at the period of the poem. But in this case there would be no longer any alliteration for the ear, unless we assume that both vowels had the same sound, in spite of the different spelling. On the other hand, alliteration was certainly intended in this particular case, the two words being syntactically coordinated. AND IN THE CARMEN Owing to the difference in stress of Teutonic and Romance words, a requisite which is indispensable in Germanic alliter- ating verse, namely, that tonic syllable and alliteration must coincide, need not be insisted upon in the discussion of Old French alliteration. The prefix does not count in cases where the root syllable alliterates with another word or root syllable, and where it is likely to have been clearly recognized as such, which was cer- tainly the case in compounds, the simple forms of which still existed. Thus demandet alliterates with main, though of course it may also alliterate with Deu. The effect of alliteration can be intensified by an accordance of one or more sounds following the initial, or of the whole first syllables. A still higher degree of intensity is reached when a word alliterates with itself, with one of its derivatives or with an etymologically related word, though some scholars consider these as special cases belonging to the realm of rhet- orical figures. There is a group of words, fulfilling such elementary and absolutely necessary functions from the grammatical point of view that they occur very frequently, being essential parts of every complete sentence. They have been left out of consid- eration altogether. Such words are: 1. Article, definite or indefinite. 2. Pronouns. 3. Common adverbs, such as puis. 4. Positive and negative particles, such as si, ni, ne, ja, guaires, etc. 5. Prepositions. 6. Conjunctions. Cases of elision have not been considered either, because the word, whose vowel is elided, belongs to the above named group of indispensable words, and because the vowel, losing its character as an initial, can no longer be regarded as ca- pable of alliteration. Finally, cases are to be noted where the poet could not pos- sibly have expressed himself differently, even if alliteration ALLITERATION IN THE ROLAND had been absent altogether and where it is therefore very- doubtful whether the alliteration alone prompted him to use the word combination in question. An example of this is found in Rol. 2235: aval : amont. Such cases have not been con- sidered. As a basis for this study I have used for the Roland the read- ings of the forthcoming college edition, The Song of Roland (Oxford Version) by Professor Th. A. Jenkins (D. C. Heath & Co.) 1 ; for the Carmen the edition of Gaston Paris, Rom. XI, 1882, pp. 465 ff. Before entering upon a discussion of the origins of the alliterations of the Roland, we must draw a distinction be- tween the occurrence of alliteration in artless every-day prose and its use as a means of poetic technique by a liter- ary man. I can, for instance, think of a case where a lan- guage possesses a large number of popular alliterating for- mulas which are commonly used in colloquial conversation, but where for reasons of taste or literary tradition they are strictly excluded from poetry. On the other hand, there is a possibility of alliteration being constantly used in the elevated genres, perhaps under foreign influence, and remain- ing entirely strange to the people, though such a case in its extreme form, at least, may never have existed. As this study purports to treat alliteration as a poetic device, we shall not have to answer the question of the origins of popular allitera- tion in Northern France. It is an established fact that at the end of the eleventh century it already existed. The use of alliteration as' a poetic device in the Roland may come from two groups of sources: It may be the invention of the author, who, noticing the vast treasure of popular alliterations current among the people, took over a goodly number of them, increasing them by additions of his own, with the intention of enhancing the 1. Professor Jenkins has numbered the lines just as they occur in the Oxford manuscript, that is, making no transpositions of laisses, either at 11. 239-331, or at 11. 1467-1671. It will be necessary to keep this in mind in verifying references found in these two sections of the poem. AND IN THE CARMEN beauty of his poem by so doing. Or, it may be a borrowing from another literature where alliteration was already used as a poetic device. If the first of these two suppositions be true, it must be con- cluded that the majority of the alliterations found in the poem are of popular character. We know that popular alliteration is restricted to certain formulas consisting of two or more alliterating, coordinated words. The number of examples rep- resenting this kind of alliteration is very small in the Roland. I quote them here in alphabetical order: baston : brief, 341 blanches : beles, 2250 Mans : blois, 999, 1800 bleciede : blesmide, 590 bons : beles, 3064 codart : codardie, 1486 code : crignete, 1494 cors : contenant, 118 deriedre : devant, 1832, 3118 desoz : delez, 114 doel : damage, 2983 didst : detoerst, 772 fer : fust, 1602 filz : fredres, 2420, 3411 forz : fiers, 1879 fous : flambe, 2535 frainst : desconfist, 1247, 1305 frait : fendut, 3604, 3927 longes : larges, 2852 messe : matines, 164, 670 vers : pales, 1979 pers : parenz, 1421 p2ez : poinz, 1969 plorent : plaignent, 3722 porz : passages, 741 regnes : reis, 3408 rei : re'ine, 2713 reluis : reflambes, 2317 ymdgenes : fdeles, 3664 Of these examples some might be considered as not genuine- ly popular. Compared with the large number of other allit- erations, they form a very insignificant fraction of the total, and the theory that the poet took over the device from popular sources can no longer be maintained. It must then be a borrowing from another literature which used it in poetry. There were three great literatures using alliteration in verse which were either flourishing at the time or were at least well known and assiduously studied: Teutonic literature (Anglo-Saxon, Old Norse) 6 ALLITERATION IN THE ROLAND Classical Latin literature Mediaeval Latin literature All three may have furnished some of the elements out of which the Old French epic was built up. In order to gain the criteria necessary to decide which of the three suggested the use of alliteration to the author of the Roland, we must classify the whole material according to the number of alliterating words and their position in the epic verse. The following tables show the alliteration in the Chanson de Roland; the syllables have been designated by their num- ber in sequence; thus an alliteration consisting of two allit- erating syllables the first of which is the first syllable of the verse, the second the sixth, would fall under the heading 1-6. I. Alliterations within one verse line A. Alliterations consisting of two alliterating words: 254 § 1. 1-3 respont : Rodlanz, also: 1701 mielz 292, 1062, 1088, 1106, 1394, 1591, 1752, 2006 1258 ferez : Franceis, also: 1937 1600 tot : trenchat 1695 plaindre : podoms morir, also: 2336, 2738 1851 Rodlanz : reguardet, also: 1978 2454 Charles : chevalche, 3695 Charles : cheval- chet 3729 quatre : confesses § 2. 117 blanche : barbe 248 respont : reis 359 mielz : moerge 790 passez : porz 1132 clamez : colpes 1233 ferez : Franc 1536 brochet : bien, also 1573, 1944, 2128; bro client : bien 3877 1-4 1586 trenchet : teste, also: 1956 1654 Franceis : fierent, also: 1662 2285 ovrit : oeilz 2360 tornat : teste 2383 claimet : colpe, also: 2239 2748 quite : claim 2974 venir : volt AND IN THE CARMEN § 3. 1-5 76 teres : tant 450 dient : desfaimes 1022 tanz : tanz 1311 Margariz : molt 1938 dist : dehet 2093 alquanz : alquanz 2466 molt : merveillose 2724 cZ^s^ : dame 2766 isnelement : issent 3201 beZs : Baliganz 3234 Charles : chevalchet 3300 moZ£ : Monjoie 3471 paiew : Preciose § 4. 1-6 64 Estramariz : Eudropin 1628 efa"ett£ : dechiedent 472 dc??^ : donrat 1799 escwz : espiez 938 Franceis : France, also : 2092 £e£s : troevet 989 2301 ate : doeZ 1080 Franceis : ferront 3535 Gefreit : Jozeran 1454 sep£ : sonent 3744 asemblet : Ais 1519 pramis : prendrom § 5. 1-7 234 saveir : sei£ 2154 lancons : laissoms 921 p^s : Piedres 2185 cerchet : cerchet 1251 guardet : gloton 2475 Franceis : fustes 2020 trestoz : tere § 6. 1-8 206 loderent : legerie 434 pn's : podestet 1146 sire : saviiez 2200 Rodlanz : recerchier 2533 veif : vewz 2730 drodmonz : dire 2733 rendre : recredant 2896 tornez : tenebros 3086 cors : contenances 3447 vai£ : vasselage 3614 repaidret : remembrance 3987 chrestiene : conoissance 52 Charles : chapele 283 vairs : visage 789 i>m#£ : vaillanz 1230 guardet : gloton 7. 1-9 1370 /ier£ : Valferrede 1467 Marsilies : martirie 1744 venget : vengier 1957 brandist : abatut 8 ALLITERATION IN THE ROLAND 1991 tant : troblet 3166 fait : fosset 2895 cors : color, also: 3763 3264 paien : prodo?ne 2953 toz : trovez 3723 Aide : alede 3929 brandit : abatut 47 dist : destre 561 font : Frans 2884 prenent : pin 2998 reclaimet : Rome 1019 veit : venir 2307 tantes : teres 177 Frans : France 289 Deus : donet 356 estraiz : estes 757 mul : mule 956 tient : tere 1299 tot : trenchet 1323 quinze : cols 1969 piez : pomz § 8. 1-10 3133 fier : faillent 3136 so^ez : sachent 3408 regnes : reis § 9. 2-3 2944 dist : dux § 10. 2-4 2208 fustes : filz 2278 fut : forz 2535 fous : flambes 2930 bar&e : blanche 2938 donget : Dews 3382 A*£ : /ori 3739 bien : batent 310 Sarragoce 2589 Tervagant 2722 devendrai § 11. 2-5 sa-i 2875 co£s : comt£ tolent 3269 Chanelius : chevalchent dolorose 3531 quens : codardise 12. 2-6 491 pris : £W^s 502 meillors : meinet 772 dms£ : detoerst 852 Sarragoce : soner 892 conduirai : cors 923 Rencesvals : Rodlant 1346 doze : deivent 2034 par£ : poe£ 2258 paredis : porte 2381 2?oe£ : pZor£ 2635 ?m'e : mer 3239 Valfo'it : -yerai£ AND IN THE CARMEN § 13. 2-7 115 faldestoel : fait 1188 nies : nom 477 podestet : pris 2461 Val-Tenebre : vont 538 Charlemagne : chanuz, 3085 milie : meillors also: 551 3137 tote : tabors § 14. 2-8 863 sire : servit 2084 quatre : cors 878 combatrai : compaig- 2483 Rencesvals : repaidrier nons 2670 Clarifan : Clariien 1658 Sarrazin : soffrir 3018 chevalchiez : c/w'e/ 15. 2-9 155 voldrat : devenir 405 chevalchierent : chemins 1291 vai^ : FaZierne 1294 /ier£ : forceles 1594 Malquidanz : Malcud 1881 monies : mostiers 1882 preierat : pecchiez 2029 /esis : forsfis 2134 quens : codart 2308 Charles : chanude 3067 guiderat : Godselmes 3486 chevaliers : chadir 773 poei : pfor£ 935 £rois : £es£e 1428 cm£ : 2>wi 1115 mifo'e : ??i£s£ 1779 /is£ : /«s£ § 24. 4-5 2876 verte : veit 2892 ?*m : revint 3154 /er : /zts£ § 25. 4-6 2420 filz : fredres 2549 firtiaW : granz 2719 poign : perdut 2824 s^Ve : seignor 3047 betes : bows 3701 mandet : mandet 26. 4-7 239 mandet : mercit 837 pomz : depecout 844 /eZ : /a# 883 riedreguarde : grant 1613 par£ : paiens 1856 /tors : /ace£ 1928 champ : Charles 2087 moi : ?m'e 2232 mori : raoft 2891 pasmet : pout 3229 Nigres : noefme 3721 morte : mercit 3799 pfei£ : preiom 3919 /rowi : /ai^e § 27. 4-8 97 prise : peceiez 1177 pri : 261 barbe : blancheier 2080 cors 1105 /at'£ : /era£ 2472 vow£ 1154 amies : avenanz porpenset : quatre § 28. 4-9 484 poign : paien 696 mains : comandez 741 porz : passages 746 d#£ : diables 1107 coer : codardet 1184 aii : ater 1218 2ne'£ : pow£ 1249 biew : brandir 1690 moergent : molt 1768 reis : Rodlant 1866 fredre : faillir 2103 voeW : vendrat 2147 /or;? : ajornez 2163 Rodlanz : r ernes 12 ALLITERATION IN THE ROLAND 2184 tornet : toz 2234 drecet : dolor 2457 mot : montez 937 per : perdre 1404 magnes : demente 1660 /orce : Franc 1841 irop : £ens 2007 pardoins : Deu 2090 presse : plus 2371 /a# : /u* 2713 rei : reme 3404 veit : voelt 3523 ooc/ie : buisine § 29. 4-10 2852 longes : larges 3064 bons : beles 3122 bronie : barbe 3254 Hums : Hongres 3735 /eZ : /er 3969 Guenles : grant § 30. 5-6 1021 vei : venir 2106 molt : malement 1858 £02 : tens 3800 claimt : quite 443 wis£ : mam 999 oZems : blois 2558 vedeit : venir § 31. 5-7 2563 m'?i£ : veltres 2656 premiers : parlat 3585 mt£ : mt£ 656 Rodlant 1282 Tory's : 1486 codart : § 32. 5-8 : riedreguarde 1842 chevalchet : Charle- Tortelose magnes codardie 1939 Franceis : fierent 2396 portent : paredis 236 £02 : £ofotz 327 £rop : maltalant 3463 chevalchet : champ 33. 5-9 1895 Yvdne : /von 2144 /e£ : ferrat 34. 5-10 3846 pleges : pZa{£ 1118 /orz : /mz § 35. 6-7 2244 £02 tens AND IN THE CARMEN 13 886 molt : males 2347 seignor : saint § 36. 6-8 2943 blanche 4001 3021 avrat : altretant barbe, also § 37. 6-9 122 premerains : parlet 174 Geriers : Gerins 196 mar : Marsilie 1011 granz : granz 1502 donat : diables 2317 reluis : reflambes 2587 tere : trestornent 2795 poign : perdiet 2983 doel : damage 3426 dous : desaffret 3714 esforciet : eschange § 38. 6-10 341 baston : brief 727 morst : mals 1201 deseivret : dos 1284 derompit : dobles 1530 baiset : &oc/ie 2193 poe£ : pfor£ 2250 blanches : beles 2010 raor£ : molt 1026 Rodlanz : respont 1329 mcw7e : menude 1683 poe£ : preisier 2512 fame : luisant 2580 corent : cro£e 2966 marbre : mis 3031 Charles : chiere 3332 porprises : parz 3687 veident : vont 3722 plorent : plaignent § 39. 7-8 2361 voeW : veirement § 40. 7-9 2734 ira£ : i£em£ 2746 vienge : vedeir 2858 at'ciwi : a£er 3792 piez : presentet 1602 /er : fust 2720 Rodlanz : riches 2936 voldreie : vivre 41. 7-10 3941 afublent : aZ£res 3993 Dew : dire 14 ALLITERATION IN THE ROLAND § 42. 8-9 303 rit : Rodlanz 1570 mil : mangons 944 porz : passant, 2124 veit : venir 1766 porz : passanz, 3141 son* : saffret 2772 porz : passer 3746 sam£ : Silvestre 1338 champ : chevalchet § 43. 8-10 159 £endre : tref 2844 /men* : Franc 1879 forz : /iers 3530 morz : ra^ie 2334 oaroe : blanche, also: 3784 rcwsc-n : rendre 3618, 3712 § 44. 9-10 1190 mals : moz 1410 *eZ : *ren*e B. Alliterations consisting of three alliterating words 89 mules : amener : Marsilies (4: 7: 9) 657 puis : por* : passage (2: 6: 9) 782 reis : Rodlanz : receut (2 : 5 : 8) 908 Marsilies : mo^ : mercidet (3: 5: 8) 953 venez : vedrez : veirement (1: 6: 8) 1339 **en* : trenchet : *ew7Ze* (1: 7:10) 1579 diewt : -Deits : doel (1: 5: 7) 1604 Franceis : Frans : fus (3: 5:10) 1825 ?n*en : batierent : bostons (1: 3: 9) 1898 dis* : Damnesdeus : doms* (1: 5:10) 2125 /cm* : /orz : /iers (3: 4: 6) 2169 arcevesque : afo£ : aidier (2: 7: 9) 2690 passent : portes : ponz (1: 4:10) 2873 p#e* : poe* : pZor* (1: 6:10) 2955 asez : at : abez (1: 4: 9) 3608 Charles : chancelet : chaduz (1: 3: 9) 3809 clamez : piie : ccwte (2: 4: 6) 3833 /eZ : /eZorae : /is* (4: 6:10) 3898 Dews : oJoms : drei* (1: 8:10) 3998 chrestien : reclaiment : evident (2: 7:10) AND IN THE CARMEN 15 C. Alliterations consisting of four alliterating words 825 pitet : prent : poet : plort 3606 prent : pleine : palme : plus D. More than one alliteration in a verse line § 1. model a : a : b : b 1053 respont : Rodlanz : fereie : fols 1111 fait : fiers : leons : leuparz 1608 dis£ : Deus : tot : tramette (1:4:6:10) (1:6:8:10) (1:3:6:10) (3:4:6: 9) (1:5:6: 9) § 2. model a : b Deu b : a 123 dist : salvez : seiez also: 428, 676 143 dist : messages : molt : dit 199 Pris : Valterne : tere : Pine 1506 cors : trenchet : tres : costet 1694 tanz : vassals : vedez : tere 2597 a^re : mot : molt : haltement 3906 cfo"s£ : Pinabels : placet : Damnedeu (2:5:7:10) (1:3:8:10) (1:4:7:10) (2:4:5: 7) (1:3:5:10) (2:4:5: 6) (1:2:6: 8) § 3. model a : b : a : b 537 paiens : molt : jmws : merveillier 841 magnes : poe£ : muder : plort 1047 £foe>i£ : Franceis : e£e/ie£ : /m£ 1091 miete : voeill : morir : vaignet 1449 Marsilies : vient : mi : valede 1454 sep£ : mzfo'e : sonent : menede 2016 prie£ : Deu : paredis : donget 2241 pne£ : Dew : paredis : doinst 2271 foaW : sow£ : /ioi£ : sont 2641 laissent : Marbrise : laissent : Marbrose 3217 eschieles : trente : establissent : tost 3720 pert : chiet : piez : Charlemagne 3846 dist : pleges : demant : pfrii£ (3:5:7: 8) (4:6:7:10) (1:3:5:10) (1:2:3:10) (1:4:6: 9) (1:2:6: 9) (2:4:6:10) (2:4:6:10) (1:2:7: 8) (1:3:7: 9) (1:4:5:10) (1:5:7: 8) (1:5:7:10) 16 ALLITERATION IN THE ROLAND II. Alliterations extending over two verse lines 1 A. Alliterations consisting of three alliterating words § 1. Two alliterating words in the first, one in the second verse line 409-10 tote 416-7 sals 31-2 mil : mudiers : muls 57-8 trenchier : testes : testes 66-7 Machiner : Maheu : Malbien 125-6 mandet : Marsilies : molt 209-10 prist : puis : enprise 222-3 mandet : Marsilion : mains 227-8 mort : moerjuns : mont 264-5 Reins : renc : rei 285-6 tant : tuit : toz 400-1 tot : talent : teres tint : tot seiez : saintes 480-1 mid : mule : malvais 496-7 dist : dit : dreiz 527-8 riches : reis : recredanz also: 542-3, f-- * 555-6 same 762-3 parlat . : padrastre : pute 809-10 desrenget : destreiz : descendrat 942-3 sont : sosduiant : seignor 969-70 Franceis : France : floride 983-4 dient : diables : dist 1002-3 jut : jut : rejlambeit 1008-9 respont : Rodlanz : rei 1012-3 perdre ; peil : empleit 1047-8 Franceis : juit : jaldrat 1057-8 paien : porz : plevis 1140-1 sont : asols : seigniez 1159-60 cors : cler : compaing 1209-10 proz : porz : perdrat 6: 9: 8) 7:10: 8) 2: 9: 6) 4: 6: 4) 4: 6:10) 4: 7: 8) 7: 9:10) 4:10: 4) 1: 5: 5) 6: 9: 6) 6:10: 1) 5: 6: 6) 2: 4: 7) 2: 8: 7) 2: 4: 5) 4: 8) 9: 9) 6: 2) 8: 5) 6: 9) 6: 2) 2: 7: 9) 1: 3:10) 4:10:10) 3:10: 8) 3:10: 3) 2: 4: 9) 1: 7: 3) 4:10: 3) 1. It goes without saying that both lines belong to the same laisse. AND IN THE CARMEN 17 1255-6 Franceis : fo'ir : ferons (3: 9: 5) 1285-6 met : more : mist (8:10: 9) 1308-9 douze : dis : dous (2: 6: 4) 1315-6 costet : conduist : cors (3: 6: 6) 1325-6 cheval : Chernuble : charboncle (2: 9: 7) 1371-2 tote : teste : trenchet (1: 4: 1) 1417-8 moerent : milliers : mort (1: 6: 6) 1439-40 mort : milliers : milliers (4: 6: 3) 1472-3 males : molt : Marie (4: 6: 9) 1483-4 molt : medisme : molt (1: 9: 8) 1539-40 mei : -more : misi (8:10: 9) 1543-4 paien : presse : prodome (3: 7: 9) 1560-1 Charles : chicrs : chevalier (7:10: 8) 1614-5 Capu'el : Capadoce : claimet (2: 8: 7) 1618-9 /e?^r : force : fraint (3:10: 6) 1643-4 voelt : volt : vertuosement (4: 8: 6) 1698-9 fredre : faire : comfaitement (4:10: 2) 1703-4 porz : passanz : plevis (8: 9: 3) 1725-6 mielz : mesure : mort (1: 3: 4) 1730-1 mort : Marsilies : mar (4: 9: 7) 1800-1 Mans : blois : frarow (5:10: 7) 1873-4 voeillet : vengier : vait (7: 9: 6) 1883-4 respont : Rodlanz : recomenciet (1: 3: 7) 1992-3 pres : poet : poissct (4: 6: 5) 2011-2 teste : tornent : tote (7:10:10) 2070-1 presse : paiens : piet (4: 9:10) 2099-2100 quens : combat : cors (2: 9: 3) 2173-4 plaies : pairs : piz (4: 6: 4) 2227-8 petit : pas : ?;>oe£ (2: 4: 7) 2288-9 fiert : M : froisset (2: 8: 1) 2298-9 piez : puet : perdude (4: 7: 9) 2310-1 tens : tenude : tel (8: 9: 4) 2373-4 destre : Dett : descendent (2 : 8 : 6) 2400-1 p^em : pie£ : paien (9:10: 9) 2419-20 pitiet : plort : plorent (3:10: 1) 2422-3 pasment : plusor : proz (6: 9:10) 2426-7 podez : poldros : paienor (3: 9: 8) 18 ALLITERATION IN THE ROLAND 2447-8 veit : vespre : verte 2505-6 more : mercit : manovrer 2517-8 poet : plort : priet 2555-6 altre : avison : Ais 2576-7 moillier : Bramimonde : molt 2600-1 fait : felonie : faillirent 2686-7 Baliganz : baron : baston 2695-6 demeinent : dolor : deus 2707-8 message : mantels : montet 2744-5 fUz : fille : fut 2770-1 Marsilies : mandet : mort 2780-1 Durendal : donat : destre 2787-8 quite : claime : comencet 2863-4 raison : Rodlant : regnet 2868-9 plus : poet : pui 2878-9 piet : plein : prist 2905-6 parenz : prot : pleines 3038-9 sont : sont : sont 3068-9 eschiele : establide : est 3087-8 barbes : blanches : bronies 3152-3 tient : Maltet : tinels 3167-8 piez : poet : paien 3184-5 bels : Baliganz : bons 3226-7 Sorbres : Sorz : siste 3237-8 eschieles : establissent : est 3298-9 paien : Preciose : perte 3370-1 conduit : cors : cols 3411-2 filz : fredres : furent 3532-3 vassals : vestit : vit 3557-8 dites : Deu : demandereiz 3596-7 pais : paien : apresentet 3617-8 trenchet : teste : tot 3710-1 per : prendre : pesance 3713-4 mort : demandes : molt 3740-1 altre : ait : atent 3896-7 Tiedris : tendrai : toz 3963-4 moerget : merveillos : amener [2: 6: 4) ;6: 7: 8) [2: 6: 2) [1: 8: 6) ;6:10: 5) [7: 8: 9) 2: 9: 9) 9: 4) ;5 :3 :3 7 7 9 3) 6) 8) [2: 6: 2) 5) 9) 8) 7) [1: 4 [3: 7 1: 4 [4: 9: ;3:10: 5) [4: 7: 4) ;3: 8: 4) [7:10: 7) 1:10: 9) [4: 6: 1) 1: 8: 6) ;6:10: 3) ;i: 5: 4) 1: 5:10) [1: 4: 6) ;4: 6: 5) [3: 8: 9) |1: 6: 7) 1: 8: 9) ;i: 4: 2) ;8:10: 9) 7:10: 5) ;6:10: 7.) [3: 7: 1) [4: 6: 7) AND IN THE CARMEN 19 § 2. One alliterating word in the first, two in the second verse line 32-3 chargiez : charre : chareier (9: 4: 8) 106-7 Gefreiz : Gerins : Geriers (1: 6: 9) 163-4 matin : messe : matines (7: 1: 3) 242-3 deit : client : dus (6: 1:10) 252-3 seignor : Sarrazin : Sarragoce (1: 2: 6) 333-4 did : dient : Deus (3: 1: 5) 349-50 tant : twit : tant (5: 2: 5) 361-2 part : Pinabel : per (4: 2:10) 410-1 wAlie : mot : mot (6: 6: 9) 418-9 message : mains : mont (3: 4:10) 437-8 morrez : Marsilies : molt (2: 3: 7) 439-40 /%£ : ferir : /to (7: 1: 7) 508-9 dist : destre : deiz (2: 8:10) 576-7 corteis : conte : cre££ (9: 4:10) 584-5 riedreguardc : Rodlanz : riches (2: 7:10) 589-90 bataille : bleciede : blesmide (3: 6: 9) 620-1 meillor : mil : mangons (5: 8: 9) 632-3 bien : baisierent : boches (1: 3: 6) 665-6 a£ew£ : treiit : iere (6: 3:10) 669-70 matin : messe : matines (7: 1: 3) 742-3 riedreguarde : respont : Rodlanz (8: 3: 5) 780-1 tendut : trovez : ires (9: 3: 6) 787-8 dit : Dews : desment (2 : 1 : 9) 807-8 faire : Franceis : France (10: 3: 6) 821-2 pulceles : pitet : plort (3: 7:10) 822-3 at : altres : angoissos (4: 4: 8) 847-8 muls : Marsilies : mandet (1: 1: 4) 853-4 p^s : paien : prit (7: 3: 6) 973-4 podrom : paiens : parfondement (3: 3: 5) 1035-6 mesure : medismes : raoZ£ (6: 3: 7) 1072-3 retorneront : respont : Rodlanz (6: 7: 9) 1090-1 valor : voeill : vaignet (8: 2:10) 1136-7 descendent : Deu : benedist (3: 6:10) 1190-1 Franceis : felon : Franceis (3 : 1 : 3) 1195-6 destre : Dews : doe£ (6: 5: 8) 20 ALLITERATION IN THE ROLAND 1210-1 France : ferez : Franc 1246-7 ferir : frainst : desconfist 1266-7 cors : contreval : quat 1275-6 ferir : frainst : flors 1306-7 mist : mort : mil 1517-8 malvaisement : mielz : moerjom 1522-3 sainz : serez : sedant 1596-7 soleil : siet : Saltperdut 1625-6 vait : Valerie : Viviers 1631-2 doel : dist : Dens 1816-7 conte : comandat : cous 1924-5 fel : ferez : forbides 1940-1 out : ont : orgoeil 1979-80 descolorez : clers : cors 1983-4 compaign : cors : contrevaillet 2030-1 morz : mot : marchis 2042-3 vencuz : voeillet : vals 2063-4 recredwnz : recomencent : crit 2085-6 resailit : Rodlant : reguardet 2087-8 vencuz : vassals : vis 2147-8 pesmes : perdut : pers 2445-6 trovez : tienent : tuit 2613-4 bries : Babilonie : Baligant 2616-7 sorvesquiet : Sarragoce : socorre 2774-5 riedreguarder : Rodlanz : rentes 2778-9 reis : Rodlanz : remes 2829-30 pernez : poign : pris 2839-40 doel : degrez : descent 2934-5 mise : mort : molt 2945-6 dist : dolor : demenez 2970-1 comandet : Milon : marchis 3115-6 cors : cZer : contenant 3155-6 montez : Marcides : mer 3206-7 podez : pan : pais 3208-9 Val-Marchis : Malprimes : mercit 3218-9 merveillos : menor : milie 3239-40 travers : T^rs : fr'erce (5: 1: 4) (9: 4:: L0) (7: 5: 10) (6: 4:: 10) (10: 2: 7) (5: 4: 6) (1: 7: 9) (3: 1: 8) (2: 3: 6) (3: 1: 5) (6: 2: 6) (4: 1: 9) (9: 4: 6) (6: 4: 8) (3: 7: 8) (4: 4: 9) (9: 1: 8) (3: 3: 10) (7: 1: 3) (9: o . 7) (2: 1: 10) (9: 1: 5) (7: 2: 5) (2: 2: 6) (7: 9) (2: 3: 9) (1: 2: 6) (2: 3: 9) (10: 4: 5) (6: 3: 6) (4: 1: 9) (4: 1: : 8) (9: 5:10) (7: 6: : 9) (9: 3: : 9) (6: : 3 : 6) (9: : 4 : 7) AND IN THE CARMEN 21 3243-4 siste : sedme : Samuel (3: 3: 8) 3466-7 derompiet : dous : deseivret (6: 2: 6) 3479-80 damage : doel : departet (9: 1: 9) 3537-8 paiens : placet : portez (3: 4: 7) 3558-9 Franc : /eZ : fierget (4: 4: 6) 3603-4 fiert : /raft : fendut (1: 7: 9) 3621-2 mot : montez : magnes (4: 5:10) 3646-7 doel : diables : donet (4: 7:10) 3772-3 mort : messages : Marsilion (6: 1: 7) 3824-5 sire : savez : servit (2: 2: 9) 3845-6 destre : dist : demant (6: 1: 7) 3908-9 mortel : mielz : morir (9: 1: 3) 3916-7 fiert : M*s : fait (2: 4: 8) 3926-7 fiert : /ra^ : fendut (1: 7: 9) 3932-3 penduz : parenz : plaidiet (9: 3: 6) B. Alliterations consisting of more than three alliterating words § 1. Two alliterating words in the first, two in the second verse line 49-50 Francels : desfaire : Franc : France (3:10: 1: 6) 889-90 poignant : Malprimes : plus : piet (3: 6: 1: 4) 1272-3 empeint : passet : pleine : place (2:10: 1:10) 1505-6 colp : quit : cors : costet (4: 6: 2: 7) 2016-7 Deu : donget : benedist : dolce (4:10: 4:10) 2363-4 quens : conquerant : claimet : colpe (4: 8: 1: 4) 3172-3 bien : baron : blanche : barbe (5: 9: 1: 4) 3481-2 fierent : Franceis : froissent : f orbit (4: 5: 1: 9) 3615-6 fiert -.France : fraint : reflambent (1:10: 4:10) § 2. One alliterating word in the first, three in the second verse line 3122-3 at : amor : altretel : altre (6:3: 5:10) 3860-1 messes : molt : metent : mostiers (4: 1: 5: 9) 3890-1 dementent : Deus : dist : dreit (3: 2: 3: 6) 22 ALLITERATION IN THE ROLAND § 3. Three alliterating words in the first, one in the second verse line 181-2 Marsilies : tramis : messages : masse (3: 7: 9:10) 540-1 tantes : teres : travailliet : tanz (2: 4: 8: 1) 1030-1 Sarrazins : sont : asemblet : sont (2 : 7 : 9 : 8) 1438-9 Franceis '.France : ferut : fols (1: 4: 6:10) 1843-4 bronie : barbe : blanche : baron (4: 8:10: 7) 1891-2 brocket : bien : Bevon : Belne (1: 4: 9: 6) 3125-6 passent : puis : plus : parfondes (1: 4: 9: 3) § 4. Two alliterating words in the first, three in the second verse line 3846-7 pleges : plait : parenz : plevist : Pinabels (5:10:3:6:8) § 5. Two alliterations, the one of three, the other of two al- literating words 3980-1 creidre : Deu : chrestientet : demandet : Deus (1:4:5:9:7) III. Alliterations extending over three verse lines A. Alliterations extending over four alliterating words § 1. (model: a/a : a/a). 91-3 mises : montet : message : mains (10:3: 7: 4) 700-2 font : Franc : font : France (5: 1: 5: 4) 1304-6 fiert : frainst : desconfist : fort (6: 4:10: 2) 1353-5 ferir : frainst : flor : fors (3: 4:10: 1) 1399-1401 tante : tant : tante : tant (5: 1: 8: 1) 1408-10 perdit : plait : pendre : parenz (3: 2:10: 3) 1420-2 perdent : pers : parenz : porz (4: 6: 9: 7) 1715-7 dist : dis : deignastes : damage (1: 4: 9: 9) 2306-8 tantes : tantes : teres : tient (1: 2: 4: 4) 2308-10 at : ait : altre : at (8:3:7:6) 2332-4 teneit : teres : tantes : tient (7: 8:10: 4) AND IN THE CARMEN 23 2381-2 muder : medisme : metre : mercit (3: 3: 7: 9) 2593-5 porter : phisors : peintes : ploret (3:1:6:6) 2640-2 mer : Marbrise : Marbrose : amont (4: 3: 9: 4) 2994-6 Marsone : mort : Malpalin : molt (9: 4: 5: 5) 3011-3 repentent : Rodlant : rendre : respont (9: 3:10: 1) 3393-5 afichiede : fut : fort : fins (9:6:8:7) 3400-2 paien : pleniers : perdent : plus (1: 3:10: 1) 3691-3 sages : sarcous : seignors : saint (7:3:9:2) 3724-6 pasmede : pitiet : ploret : prent (9: 1: 6: 1) 3905-7 parlet : Pinabels : placet : parentet (9: 2: 6: 8) B. Alliterations consisting of more than four alliterating words § 1. model: a:a / a:a / a 553-5 tantes : teres : tanz : trenchanz : tanz (2:4:1:9:1) § 2. model : a / a : a / a : a 1022-4 flambeios : feront : Francels : fel : faite (8:3:6:4:6) § 3. model :a : a / a :a :a/a 2229-31 at : at : oiftz : aias£ : arpent : avant (2:6:1:3:7:9) IV. Alliterations extending over more than three verse lines 794-8 vint / vint : vint / vint / vint : vieilz / venuz (1:2:7:1:1:10:1) 1491-4 colpez / corte : quisse : crope / costez / code : crignete (3:1:4:7:3:4:7) 1636-40 vaillanz / vertuos : vassals / veie / vit : veirement / visage (9:2:6:4:4:8:3) Let us now determine the characteristic traits of Teutonic alliteration. Lachmann 1 denned it as follows : "Alliteration — 1. Ersch und Gruber, I, 166. 24 ALLITERATION IN THE ROLAND nennt man die in der nordischen Dichtkunst gebrauchliche Art von Assonanz, die durch gleiche Anfangsbuchstaben der Worter hervorgebracht wird. Alle Selbstlauter reimen auf- einander ohne Unterschied ; hingegen manche besonders horbare Verbindungen von Mitlautern, wie st, sp, erfordern genaue Wiederholung, so dass z. B. ein einfaches s nicht als Reim darauf gelten wiirde. Es ist naturlich, dass die Buch- stabenreime, wo moglich, auf die bedeutenderen Worter fallen miissen: sie konnen selbst in der Mitte der Worter sein nach weniger betonten Vorsilben. Auf den Versbau hat die Alli- teration den bedeutendsten Einfluss. Ein strenges Silbenzah- len kennt zwar die nordische Poesie nicht, aber jede Halbzeile erfordert zwei Hebungen, welche eben durch die Reimbuch- staben bezeichnet werden. Auf dem ersten ruht die erste Hebung; darauf reimen gewohnlich zwei andere, einer, der auch fehlen kann, in der zweiten Hebung des ersten Halb- verses, der andere, notwendige, auf einer der beiden Hebun- gen des zweiten. Nur die Hebungen, aber nicht die Silben vor oder zwischen ihnen werden genau gezahlt; oft konnen die letzteren sogar fehlen." H. Jordan 1 characterized Germanic alliteration in this way : 1. Only the same consonants, but all vowels alliterate with each other. 2. Alliteration coincides with the arsis. 3. It must recur in both hemistichs. 4. It must recur in all verse- lines. To this should be added that the groups sk, sp, st are treated as one consonant, at least until the later Anglo-Saxon period. Latin alliteration, on the other hand, is characterized by H. Jordan in the following words 2 : "In der romischen Kunstpoesie besteht die Alliteration in der unmittelbaren Aufeinanderfolge consonantisch gleich anlautender Worter: die vokalischen Gleichklange treten dagegen ganz in den Hin- tergrund, desgleichen die Alliteration nicht unmittelbar sich folgender Worter. — Ist im Trimeter des Plautus von einer an- 1. Kritische Beitr., p. 170. 2. Ibid. p. 171. AND IN THE CARMEN 25 deren Art der Alliteration als der bald hier, bald da auftau- chenden Aufeinanderfolge gieich anlautender Worter keine Rede, so sucht man auch in der hexametrischen Dichtung ver- gebens nach einer nur einigermassen haufigen Verteilung alli- terierender Worter auf beide Vershalften, geschweige dass hier oder dort jeder Vers alliterierte oder irgend ein Zusammen- hang mit der Hebung hervertrate. Es muss also geradezu geleugnet werden, dass die Dichter dieser Periode fur die Al- literation als versbildendes Element auch nur dass geringste Gef iihl gehabt haben : sie wenden sie an, wie es in der unge- bundenen Rede geschah, wenn man bedeutsam, spruchartig, feierlich, eindringlich redete und um so haufiger, als diese Elemente in der gebundenen Rede vorwalten, aber doch im- mer in verhaltnismassig geringem Umfange, da hochstens ein Siebentel aller Verse mit Alliterationen bedacht ist." J. Maehly 1 formulates the difference between Teutonic and Latin use of alliteration in a similar way: "Hier muss gieich zu Anfang als Unterschied von der deutschen Alliteration der Umstand betont werden, dass diejenige der classischen Volker sich nicht auf die Sphare der Consonanten beschrankt, son- dern auch die Vocale in ihren Bereich zieht. Ueberhaupt ist sie hier an kein strengeres, kunstmassigeres Gesetz gebunden, wie in jener altsachsischen, angelsachsischen und nordischen Poesie. Wenn hier in zwei zusammengestellten Verszeilen wenigstens zwei, meistens drei betonte Silben mit gleichem Anlaut beginnen miissen, so kommt bei romischen und grie- chischen Dichtern die Zahl der alliterierenden Silben eben- sowenig in Betracht, als Arsis oder Thesis, Quantitat oder Accent, sondern die Anwendung oder Nichtanwendung der Alliteration in beliebiger Ausdehnung ist so willkiirlich als moglich und dem Takt des Schriftstellers vollig anheimgege- ben." Such is also Wolfflin's opinion on the subject 2 : "Die Poesie der Romer dagegen, welche in dem Quantitatenprinzipe ein Geniige fand, hat zur weiteren Erhohung der Schonheit die 1. Ueber Alliteration, p. 209. 2. Ueber die alliterierenden Verbindungen, pp. 21-2. 26 ALLITERATION IN THE ROLAND Alliteration nur in freierer Weise iiber die Verse ausgegossen, ohne sie zum notwendigen Bindemittel je zweier Verse oder Vershalften zu erheben und im Verlaufe der Jahrhunderte auch diese Beigabe immer mehr eingebiisst." Jordan, Maehly, Wolfflin, and most of the scholars who have collected and classified examples of Latin alliteration considered only the poets of antiquity, probably because no work exists to-day which deals adequately with the use of allit- eration in Mediaeval Latin literature. However, a superficial glance at the examples given by Fuchs 1 will suffice to convince us that alliteration, far from going out of fashion, became more and more prominent after the fall of the Roman empire. At the same time it underwent certain modifications, so that the statements of the above mentioned authors concerning the use of alliteration by the classics no longer hold true for Me- diaeval Latin authors. Kawczynski 2 quotes a passage from the Etymologies of Isidore of Seville to this effect: "Paro- moeon est multitudo verborum ex una littera inchoantium quale est apud Ennium: o Tite tute Tati tibi tanta tyranne tulisti, sed hoc bene temperat Virgilius dum non toto versu utitur hac figura ut Ennius, sed nunc in principio versus tan- tum, nunc in medio, ut : Quaeque lacus late / liquidos qwaeqwe aspera dumis nunc autem in fine, ut : Sola mihi tales casus / Cassandra canebat." As early as 1882 Wilhelm Meyer 3 wrote : "Alliteration und Assonanz finden sich als rhetorisches Kunstmittel in alien Sprachen bisweilen angewendet, allein erst die regelmassige Wiederholung macht dieselben zu gesetzmassigen Bestand- teilen der poetischen Technik. Alliteration findet sich in sehr alter Zeit zuerst bei Virgilius Maro." Let us then examine the alliterations found in the Roland and try to determine their origin, according to the criteria giv- en above. As was stated on p. 2, we have no means of ascer- taining whether or not two words beginning with different 1. Die romanischen Sprachen, pp. 287-95. 2. Essai comparatif, p. 104. 3. Gesammelte Abhandlungen, I, 190. AND IN THE CARMEN 27 vowels can be considered as alliterating, or whether their oc- currence is due to mere chance. The first criterium of H. Jordan is then of no avail for our research. As the second point, we shall mention the one not given by Jordan, but contained as an essential in Lachmann's definition. It concerns the treatment of s followed by a consonant. It is obvious that if there are fewer instances of sp and st alliterat- ing with s-j-vowel than there are of sp and st alliterating with themselves, we must say that there exists a characteristic common to the alliterations of the Roland and those of Teuton- ic poetry. The opposite result would, however, not exclude al- together the possibility of Teutonic origin, since in late Anglo- Saxon poetry this separation of the three sounds, s followed by a vowel, sp, and st, was no longer observed. As a matter of fact, we find no alliterations with s+consonant in the Old French poem, a situation which is certainly due to the scarcity of words beginning with sp or st. Thus also the second cri- terium fails. The second point in Jordan's summary expresses the close connection which exists between alliteration and the rhythm of Teutonic verse, which is based on the number of arses. The French verse being the representative of an entirely dif- ferent principle of versification, that of counting the sylla- bles, it is clear that even if the use of alliteration were taken over from the Teutonic literatures, it must have been modified and adapted to the Romance verse. The Old French ten- syllable verse, such as appears in the Roland, has two fixed, stressed syllables, the fourth, followed either by the caasura directly or by an unstressed (feminine) syllable, which does not count, and the tenth, followed by the end of the verse line or by a feminine syllable. We should, therefore, expect these syllables, the fourth and the tenth, more than any others, to take the place of the Teutonic arses which carry alliteration. Since the Germanic long verse line (Langzeile) is not limited to any number of syllables, it is conceivable that a French poet writing in ten-syllable verses should, when adopting the use of alliteration, either consider one ten-syllable line as his 28 ALLITERATION IN THE ROLAND verse unit, making it take the place of the Teutonic verse line and the caesura after the fourth counted syllable that of the Teutonic caesura, or unite two ten-syllable verses, regarding the verse end of the first line as having taken the place of the caesura of his model. Which of the two ways the poet follows will incidentally depend upon the length and structure of his Teutonic model. Let us consider the former of the two possibilities. In short Teutonic verse two alliterating words, one in each hemistich, are the rule. We should then expect alliteration of the type 4-10 to be most frequent, although even this would be against the norm of the Teutonic verse, which generally requires that the second alliterating word be at the beginning of the second hemistich, not at the end. Our expectation is far from being realized on examining the facts: the types which are repre- sented by the largest figures are 1-4 with 22 and 1-3 with 24 examples, whereas the type 4-10 stands in the sixth place. The type 1-4 and 1-3 represent entirely un-Germanic allitera- tions, the two alliterating words occurring in the same hemi- stich. If the supposed verse model were a "long line," that is, a line containing many syllables, he would have found alliter- ation of three alliterating words to be the general rule. Ac- cording to our assumption that the fixed stressed syllables of the Romance verse were the only ones which could have taken the place of the Teutonic arses, the three alliterations would have to be of the type 4-10-4, since in the most common type of this kind of alliteration the third alliterating word stands at the beginning of the second hemistich of the Germanic verse line. In the Roland we find only one example of this type in a total of 106 alliterations consisting of three alliterat- ing words, the first two of which are found in the first, the third in the second ten-syllable line. Thus conditions 2 and 3 of Jordan's summary remain unfulfilled. As for the fourth point of Jordan, we shall see later that only 18.3 per cent of all verses of the Roland contain allit- eration. We come, therefore, to the conclusion that the use of AND IN THE CARMEN 29 alliteration in the Chanson de Roland cannot be of Teutonic origin. Here a special reservation must be made for the six pairs of alliterating personal proper names which occur in the poem. They are the following: Clarifan : Clariien 2670 Gerins : Geriers 107, 174, 2404 Estramariz : Eudropin 64 Machiner : Maheu 66 Gefreit : Jozeran 3535 Yvorie : Ivon 1895 Professor Pio Rajna 1 , in 1884, called attention to the fact that these names find their parallel in the Germanic epic, as Hettel and Hilde, Lindegast and Lindeger, Sigemunt and Sige- lint, etc. Kristoffer Nyrop 2 held the same opinion, adding that in Old Norse legends and partly also in real life the names of members of the same family, or of such persons as were bound together by some other tie were derived from the same root, as, for instance, Sigmund and Sigurd, Sigar and Signe, Gun- nar and Gudrun, etc. Numerous examples are given by Karl Weinhold, 3 who distinguishes four classes of alliterating proper names : 1. Such as are merely alliterating: Buri : Bor; Ask : Embla; Yngvi : Iorund. 2. Such as are besides differentiated through ablaut: Bivur : Bavur; Ani : Oni, etc. 3. Such as have the whole first part in common : Thord : Thorgerd; Vebiorn : Vestein, etc. 4. Such as have related meanings beside alliteration: Kott : Kisi. The occurrence of alliterating proper names in Latin is well 1. Origini, p. 54; cf. on the same subject, Ludwig Uhland, Schriften zur Geschiehte der Dichtung und Sage, Band I, Stuttgart, 1865, p. 365. 2. Storia dell' epopea francese, pp. 193-4. 3. Altnordisches Leben, pp. 264-9. 30 ALLITERATION IN THE ROLAND known. But it appears 1 that only group 1 of the above quoted classification is represented there, while the rest seem to be characteristically Teutonic. In the Roland we find class 1 represented by two examples only: Estramariz : EucWopin and Gefreit : Jozeran; whereas we have at least four examples for class 3 : Clarifan : Clariien, Gerins : Geriers, Machiner : Maheu, Yvorie : Ivon. One, the place-names Marbrise and Marbrose, belongs to both classes 2 and 3. There is no example for class 4 in the poem. Thus seven alliterations out of ten would show decidedly Teutonic origin. Is it possible to determine whether these alliterating per- sonal proper names are of continental Germanic, Anglo-Saxon or Old Norse origin? Alliterating proper names of all the three types which we find in the Roland occur in the three Germanic literatures in question, as can be seen from the ex- amples quoted by Rajna and others. Thus it is impossible to find a definite answer to this question. Did the author of the Chanson know the use of alliterating proper names in the actuality, or did he take it from one of the three Teutonic lit- eratures? Pio Rajna supposes the latter of the two possibili- ties, seeing in these pairs of names a connecting link between the Germanic and the Old French epic. However there is no need for such a conclusion. The name of the supposed author of the Roland, Turoldus, goes far to show that Christian names had not altogether succeeded in taking the place of the Scandinavian ones, in Normandy. If such were the case, there is no reason to suppose the old custom of preferring alliter- 1. So Kalbow, Die germanischen Personennamen. This seems to be, however, an a priori assumption of his, since the whole problem of allit- erating proper names in the ancient languages has not yet received ade- quate treatment. Whether Kalbow's statements with regard to this are true in their entirety is still open to doubt. The names of Semiramis' sons, Hypates and Hydaspes, which I met in Diodorus Siculus (lib. II, cap. 1) and those of Ohola and Oholiba which occur in the Old Testa- ment (Ezekiel, XXIII) seem to belong to group 3 of Weinhold and would be an example of the occurrence of this class of alliterations in Greek and Hebrew. AND IN THE CARMEN 31 ating proper names to have been abandoned by the Norse set- tlers. Moreover, the list of the Anglo-Saxon kings of Alfred's family show it well preserved, in spite of the Christian influ- ences, as we see from names such as Ethelbert : Ethelred, Edgar : Edmund, etc. Thus the occurrence of these names in the Old French poem is in itself no proof for a connection between the Teutonic epic and the chansons de geste. Returning now to the alliterations of other than proper names, we have to determine whether the poet's model is to be sought in Classical or in Mediaeval Latin literature. To facili- tate the examination, let us summarize the points brought out by Jordan, Maehly and Woelfflin as characteristic of the use of alliteration by the Classical poets : 1. In the large majority of cases only successive words alliterate. 2. The cases of consonantic alliteration far outnumber those of vocalic alliteration. 3. The alliterating words are not distributed over both hemistichs. 4. Not every verse contains alliteration, at highest one- seventh. 5. Alliteration and arsis seldom coincide. 6. Alliteration does not serve to link two verses. Some of these points need a more detailed discussion. It is doubtless true that in most cases only consecutive words al- literate; still there are some examples in the Aeneid where the alliterating words are distributed over both hemistichs, and we have seen above that this fact did not pass unnoticed, as when Isidore of Seville quotes Virgil. This was enough to form the starting point of a new school, as we shall see later. Point 2 does not mean much, because, if the individual vowels alliterate only with themselves, as is the case in Latin, the number of vocalic alliterations must needs be inferior to that 32 ALLITERATION IN THE ROLAND of consonantic alliterations; consequently no conclusion can be drawn from the small number of vocalic alliterations in the Roland. As for the proportion of the verses containing allit- eration, it appears to me to be a dangerous principle to carry purely mathematical computations into this argument. In the particular case of the Aeneid, Kvicala 1 came to an en- tirely different result after such a mathematical procedure, as he found 72 per cent of all verses of the poem to contain alliteration. In such a method all depends upon the judgment of the computer in denning the conception of the term "in- tentional alliteration as a poetic device." But even if we ac- cept the proportion given by Jordan as the correct one, it would certainly not lead us very far in our task. Let us sup- pose we know the exact percentage of the verses of the Roland which contain alliteration. This would not allow us to con- clude that the model was or was not a classical Latin poet, since too much depends upon the taste of the author of the French poem. If he possessed little, he could easily exagger- ate a device which he found used with moderation by a classic, say Virgil. On the other hand, if he had good taste and the "sentiment de la mesure," which is the second nature of every true Frenchman, there is a possibility that the model was not a classic, but a second-rate Mediaeval author who overdid the use of alliteration, the author of the Chanson having enough tact and taste to follow his model with due reserve. There are left, then, points 1, 3, 5, and 6 to be used in the following examination. When considering point 1, we must bear in mind the fact that the genius of the Latin language was far more favorable to the formation of series of two or more consecutive alliterating words than is the French lan- guage. The Latin had the asyndeton which has not come down into Romance; it further had a very elaborate inflex- ional system, whereas French generally had to use preposi- tions for Latin genitives, datives and ablatives. It would be a wrong method, therefore, should we merely count the ex- 1. Neue Beitrcige zur Erklarung der Aeneis, pp. 293ff. AND IN THE CARMEN 3 q amples of alliterations of two consecutive words; it is neces- sary to include also those cases where the two alliterating words are separated by a monosyllable, or even by two mono- syllables, if in Latin the latter would have been replaced by the asyndeton or by a genitive, dative or ablative. The total number of this type of alliterations in the Roland amounts to about 130 out of 465, considering the alliterations consisting of two alliterating words only. Thus it is clearly seen that the author of the French poem followed a principle different from that of the Classical Latin poets. As for the third point, we find 168 alliterations within the same hemistich against 298 extending over both hemistichs, likewise considering only alliterations consisting of two allit- erating words. Again the result is contrary to the rule of Classical Latin tradition. Quantity playing little or no part in Romance verse, point 5 cannot serve as a criterium. In the Roland, alliteration is decidedly a means of linking two or more verses, there existing altogether 243 such cases against 488 where alliteration is confined to a single verse line; in other words, the first group forms a percentage of 33.2 of the total number of alliterations found in the poem. We see, then, that in all points which can be regarded as true criteria the technique of the Roland differs in the use of alliteration from that of the Classical Latin poets. There is consequently only one possibility left: the influence of Me- diaeval Latin poetry. Since a study dealing with the use of alliteration as a poetic device in the works of Mediaeval Latin literature is still lack- ing, there is only one means left to help us in the next task, that of determining the characteristics of Mediaeval Latin al- literation : we must limit ourselves to selecting a certain num- ber of verses from different authors who are known to have used alliteration, and to deducing the underlying principles. We shall examine: 34 ALLITERATION IN THE ROLAND 1. a poem ascribed to Saint Augustine, quoted by Fuchs 1 : Claustra carnis praesto frangi clausa guaerit anima, Gliscit, ambit, eluctatur exul frui patria 2. Sidonius Apollinaris, (5th century), Carmina VII, 1-2 2 : Phoebe peragrato tandem visurus in orbe quern possis perferre parem, da lumina caelo 3. a piece of rhythmical hexameters, published by W. Meyer-Speyer 3 I, 2-4: et ductus animae det focum vera dicenti? Licet in Zacrimis singultus verba erumpant de te certissime tuus discipulus Zoquar 4. Venantius Fortunatus, De Childeberto Rege, (6th cen- tury) , published by W. Meyer-Speyer 4 I, 1-3 : .Rex regionis apex et supra regna regimen qui caput es capitum, vir capitale bonum, ornamentorum ornatus ornatius ornans 5. Eugenius, Bishop of Toledo, (7th century), quoted by Fuchs 5 : Monastica de decern plagis Aegypti. Prima plaga Aegypti lymphas in sanguine vertit, Altera ranarum crepitum tabemque creavit 6. Aldhelm, (7th century), De laudibus virginum, 1-2 6 : Omnipotens #enitor mundum ditione grubernans, Lucida stelligeri qui condis culmina coeli 1. Die ro^utnischen Sprachen, p. 288. 2. Mon. Germ. Hist. Auct. antiqu., VIII, 203. 3. Ges. Abh., I, 229. 4. Ibid., II, 368. 5. Die romanischen Sprachen, p. 290. 6. Sancti Aldhelmi Opera quae extant, ed. J. A. Giles, Oxford, 1844, p. 136. AND IN THE CARMEN 35 7. Walafried Strabo, (9th century), Vita S. Mammae Mo- nachi, cap. IX, 1-3 1 : Caesaream post ista puer venerandus adivit Repperit et portas juxta residere patentes Quos dudum praeses capiendum misit ad ipsum 8. Milo of St. Amand, (9th century), Vita S. Amandi, lib. I, cap. 1, 2-3 L> : Errorum £enebris 7??undo pereunte, misertus, Virginis mtravit £halamum, m£actumque paravit 9. Heiric of Auxerre, (9th century), Vita S. Germani, lib. 1, cap. 2, 2-3 3 : Praesul amator erat, werito cognomine raactus, Doctor magnificus, weritorum luce coruscus 10. Arnulfus, Delicie cleri, (11th century), 798 4 : Parvi parva petunt, magni ??zaiuscula malunt 11. Roger of Bee, (end of the 11th century), De contemptu mandi 5 : Terram contemnas, qui coelum guaeris habere ; Si mansura placent, haec /ugitiva /uge. Alliteration as a poetic device is of frequent occurrence in the Christian-Latin hymns of the middle ages, as has been shown by J. Huemer 6 . We have space for only a few exam- ples: 1. Daniel, V, 235 7 : Novitate partus casti virgo cuncta renovasti cum paris clauso cardine 1. Migne, Patrol, lat., t. CXIV, c. 1051. 2. Acta Sanct. Bolhmd, ed. Renschen, Februarii Tomus I, p. 884. 3. Acta Sanct. Bolland. Julii Tomus VII, p. 238. 4. Rom. Forsch. II, 243. 5. Migne, Patrol, lat, t. CLVIII, c. 6910, 5-6. 6. Untersuchungen, etc. cf. also B. M. Dreves, Hymnologische Studien zu Venantius Fortunatus und Rabanus Maurus, Miinchen, 1908, p. 23. 7. Thesaurus hymnologicus, V, 133. 36 ALLITERATION IN THE ROLAND 2. Mone, 533 1 : Ave stella ???atutina, peccatorum medicina, ?wundi princeps et regina 3. Mone, 136 2 : spinas poenae, non peccati portas Jesu volens pati 4. Morel, 202 3 : i>otrus Cypri fralsamatus austro pneumatis perflatus, in pacis exemplum. In the Latin sequences of the middle ages we likewise meet with the use of alliteration as a poetic device 4 . One example 5 : Plebs parentis pietatis, plausu plaude pravitatis putando propaginem, Petrum 2^recare pastorem, peccatorum piscatorem, pacis plenitudinem. From these examples we can deduce the following conclu- sions concerning the use of alliteration as a poetic device in the Latin literature of the middle ages : 1. Alliteration as a means of poetic technique never went out of existence, but was handed down from generation to gen- eration, constituting an important factor in literary tradi- tion. 1. Lateinische Hymnen des Mittelaltcrs, II, 321. 2. Ibid. I, 180. 3. Lateinische Hymnen des Mittelaltcrs, p. 127. 4. cf. Bartsch, Die lateinischen Sequenzen des Mittelalters, p. 240. 5. Ibid. 6. cf . Fuchs, Die romanischen Sprachen, p. 257. G. Korting's statement in his Encyklopaedie und Methodologie der Romanischen Philologie, Teil II, Heilbronn, 1884, p. 411: "Audi die Alliteration wurde von den Kunst- dichtern nicht selten verwertet, in weiterem Umfange aber nur von denen der vorklassischen Zeit," must therefore be modified. AND IN THE CARMEN 37 2. The frequency of its occurrence in one poetic unit is bound by no fixed rule, but depends entirely upon the taste and predilections of the individual authors. 3. It is used in a two-fold way: either the alliterating words are consecutive, or they are separated from one an- other by words other than mere particles, the examples of the first group slightly outnumbering those of the second, though much depends here also upon the individual taste of the poet. 4. The alliterating words may or may not be distributed over several hemistichs. 5. In many cases, two or even three consecutive verse lines are linked by alliteration. 6. As for the relationship of alliteration and arsis, I do not feel competent to undertake such a research, as it would require a detailed knowledge of Latin rhythmical verse. The material used as a basis here would be insufficient for such a task, and the preliminary work necessary would be in too great a disproportion to the main object of this study. 7. Alliteration is bound, however, to no particular kind of Latin verse nor to any definite place within the verse line. In all the important points, 3, 4, 5, and 7, its use in Mediae- val poetry agrees with that observed in the Roland. There can be no doubt, then, but that the model or models used by Turoldus are Mediaeval Latin writers. This result is in ac- cordance with the observations of Wilhelm Tavernier, who demonstrated the great influence exercised by the works of erudite Latin literature, such as the Waltharnis, on the Old French poem. Of all this literature there is one poem which stands in a peculiar relationship to the Roland: the Carmen de prodicione Guenonis, a piece consisting of 482 distichs, written we do not know by whom, or in what part of France, or at what date. A superficial glance at this monument of erudite poetry suf- fices to show an immense number of alliterations, all through the poem. As there exists a certain possibility of the allitera- 38 ALLITERATION IN THE ROLAND tions helping in the solution of the question as to which of the two poems is the older, the Roland or the Carmen, it will be worth while to collect and to classify the alliterations of the latter. The definition and system of classification will be the same as those adopted in the case of the Roland, with three exceptions: First, vocalic alliteration has received the same treatment as in Wolffiin's studies, that is, all vowels are con- sidered as alliterating with each other. Second, in view of the great artificiality of the Carmen and its immense number of rhetorical figures and alliterations, it appeared probable, in more than one case, that even in the poet's choice of conjunc- tions, prepositions and particles we have examples of inten- tional alliteration. Third, the subdivision according to the position of the alliterating syllables within the verse line is no longer possible and has been replaced by another, as will be seen. I. Alliterations within one verse line. A. Alliterations of two alliterating words. § 1. Simple alliterations. 31 legatum : leges 293 auditur : armorum 39 judice : jubetur 313 mortis : Mars 91 sedere : sinistra 318 carnis : cogit 140 species : sola 324 fuit : ferus 151 Rollandum : rex 328 summam : superstes 153 donans : daturum 355 ferunt : ferre 175 letificant : legatum 359 sexaginta : supersunt 178 accipit : adit 387 arma : actus 212 scandit : scopulos 397 obliquo : Oliverus 213 vexant : valles 427 stat : secat 220 premia : polliciti 432 primus : putat 246 manum : munit 453 Rollandus : referta 262 cogit : condicione 465 jama : fuisti 278 Rollandi : ruunt 466 dux : decus 280 timidus : tutus 472 gentis : gloria 289 majora : minans AND IN THE CARMEN 39 § 2. Rhetorical figures. a) a word alliterates with itself in the same form. 12 idem : idem 77 tutus : tutus 111 rex : rex 228 dedecus : dedecus 314 gravis : gravis 342 vix : via; 345 veZwi : velut 421 comes : comes 445 herba : herba b) a word alliterates with a form of different inflexion. 33 fieri : factum 183 iz(.izts : faita 51 par at : paratus 202 querit : querat 101 re^es : regis 206 vacet : vaca^ 109 honor : honore 218 Marsilii : Marsilium 118 multis : multa 226 subveniat : subveniendo 124 millia : mille 258 victus : vincere 130 credere : crede 277 diffugiunt : diffugientes 131 Ze#e : ledum 291 omnes : omnis 132 a#as : a#e 301 instemus : instant 133 inspicit : inspecta 325 adnichilat : adnichilatis 142 ^eo;^ : /Zexa 389 visus : videt 166 movent : mover e 393 regem : regi 168 /a7fo'£ : jailer e 415 equo : equorum 169 audere : audet 417 gravis : gravior 172 possit : posse 433 regem : regis 174 scelere : scelus 443 exsanguem : exsanguis c) a word alliterates with one of its derivatives or a deriva- tive of its root. 55 talis : tanta 65 festinus : festinare 89 spaciantem : spaciosa 129 rex : regna 146 sineret : sinus 241 Gero : Gerinus, also : 267 247 latitans : latenter 248 hostes : hostili 375 regna : rex 444 immundat : munda 40 ALLITERATION IN THE ROLAND B. Alliterations of three alliterating words. § 1. Simple alliterations. 145 Karolo : carus : cordi 245 caput : cassis : cllpeus 339 aspectus : auris : ardiia 348 ferit : fugat : feros 388 prelia : posse : pati 473 patricios : proceres pedites § 2. Rhetorical figures. 8 decus : digna : decens 22 veils : -yeZte : veils 64 parens : parare : parans 73 Sme : Insidlas : Siriorum 94 decorat : decus : decor 136 medio : medium : medio 155 donat : donis : donantur 163 jusjuratus : jure : y^raf 193 comltes : comltes : comltantur 309 raoa; : mortl : mox 317 Anseum : acrlter : acer 381 omnibus : omnis : omnes 391 efoi£ : dampna : dat 411 wosce : wosce : noscls 426 metum : mortis : metus 480 ereptus : equo : equls C. Alliterations of four alliterating words. multa : 43 minis : mlnas : mlnatur 297 m'res : revocate viri : victoria 305 /or£e : /er^s ferit : fugat D. More than one alliteration in one verse line § 1. model: a : a : b : b 61 regem : re#i paratus 82 meei : remeet studet 273 mos : mo# : fugantur stans : 274 Mars : Marte gravis 113 Karolus : Karolo : regna 439 gravlter : gravis : regenda sis : exhlnc 191 rex : regnum : tentorla : 477 summos : summo turmam ocres : medlocri 259 ;am : yace^ : truncus : fugant : graves : exta- medl- AND IN THE CARMEN 41 § 2. model : a : b : b : a 59 furor : vincit : victa : 292 decern : rex : retinet : furore decern 128 nunc : parcat : parcere : 333 hostibus : invisus : w'sos nidlus : hostes 201 phalanx : tutela : tutela 336 equus : tutus : fotftts : : phalangis eques 223 hostis : patet : procedit 382 causa : scire : submonet : hostes : cito 235 omnibus : inquit : i£a : 395 minus : audax : ausus : omnes remanere 286 wisse : gentis : gens : 431 diffugiunt : celeres : cer- modo tamine : ducti § 3. model: a : b : a : b 85 incessanter : a&ii : cessat que : fugantur : abire 310 passus : eques : precipi- 121 plures : reges : plurima tatur : equo : regna • 366 desine : pudeat : desine 173 Karolo : tradat : claves : pudor : tradit 377 comperto : scelere : com- 179 miratur : reditum : mi- perta : sedicione ratur : redeuntem 390 utrumque : fuge : 200 collectis : vm's : con- utrumque : fugans sulit : viros 457 adnixus : petit : aspera : 239 probat : esse : prius : passus esse 458 supplex : veniam : sup- 240 fugando : mori : fugien- plice : voce do : mori 'ill promam : satis : pro- 273 quinque : fugant : quin- mere : solum § 4. model: a : b : a : b : b 244 armis : pereant : arma : 256 regum : remanet : rc#e parare : parant : manente : manus 42 ALLITERATION IN THE ROLAND § 5. model : a : b : b : a 344 equos : celeres : cursibus : equat : eques § 6. model : a : b : b : a : 265 Rollandi : turma : tur- a mas : ruit nes omnis : om- § 7. model : a 123 rex : regibus : subito : regna a b : a : b II. A. § 1. 3-4 10-1 27-8 32-3 38-9 45-6 59-60 69-70 90-1 103-4 subibit Alliterations extending over two verse lines. Alliterations of three alliterating words. Two alliterating words in the first, one in the second verse line. Karolus : clipeus : contemptor mage : magis : mini- tans perdi : cem possit pit minor tante : furor tere vincit errore error pinu : placet : parte decies : duo : dant perdunt : pa- posse : preci- : minor : mox tanta : tumet furore : flec- victa : velle : pererrat : 107-8 salutat : salutem : salutato 129-30 brevi : abreviare : brevi 165-6 ceca : cupida : cuncta 205-6 Rollandum : rex : rege 229-30 Mars : mors : millia 231-2 optat : optat : omni- bus 239-40 probat : prius : posse 259-60 jam : jacet : jam 265-6 Rollandi : ruit : ruunt 270-1 minus : minor : magis 271-2 instant : ingens : instat 284-5 rege : rationis : robora AND IN THE CARMEN 43 300-1 terga : timere : tali- 373-4 capitis : conamine : bus cruor 302-3 mira : movent : Marte 383-4 Marte : Marsilius : 306-7 obest : obit : Oliverus mori 329-30 Marsilius : Marti : 428-9 vulnera : viros : mente adventum 349-50 parcens : parcere : 446-7 vix : vix : volvens prosternit 471-2 satis : solum : solus 350-1 solus : solo : similis 477-8 mediocres : mediocri minimo § 2. One alliterating' word in the first, two in the second verse line. 28-9 dolus : dicas : dicat 36-7 nullus : nullus : nullus 52-3 vota : videt : videtur 55-6 gravis : gravat : gravem 56-7 monet : remanere : Minerva 73-4 timet : timet : timor 99-100 multa : multa : magis 109-10 talis : talem : talia 125-6 ungebunt : urbes : urbanos 127-8 nullo : nunc : nullus 172-3 credere : Karolo : claves 178-9 means : miratur : miratur 199-200 cura : collectis : consulit 239-40 multo : mori : mori 282-3 fluens : festinans : fremit 291-2 dimidiat : decern : decern 306-7 leso : lateri : Zoca£ 329-30 vacai : vacat : vacans 335-6 Turpinus : £wiws : tutus 337-8 singula : singula : sufficiant 340-1 costo : cr?is : cavus 343-4 clavo : celeres : cursi- bus 346-7 penam : proximus : primordia 375-6 agendum : audit : agnoscit 376-7 causa : comperto : comperta 379-80 recfo'i : redeunte : 384-5 vulnera : mros : vzres 389-90 utrique : utrumque : 394-5 mortem : minus : rewcmere 400-1 postremo : plenus : pos£ 44 ALLITERATION IN THE ROLAND 406-7 ictus : ictum : ictus 455-6 vivere : visum : 434-5 mori : metuunt : vivere remanente 457-8 scopulo : supplex : 437-8 finit : finitis : finis supplice 448-9 crure : circumquaque : 469-70 miratur : mors : mori corpora 470-1 potuisse : promam : 452-3 corpus : cedem : cede promere 476-7 sim^Z : summos : summo B. Alliterations of more than three alliterating words. § 1. Two alliterating words in the first, two in the second verse line. 5-6 presignis : prestans : pius : potens 19-20 regis : rex : regna : redire 41-2 amoris : amore : amoris : amor 47-8 minas : minantem, : minante : minas 185-6 jubet : juberi : jubere : y^&ei 223-4 2J«^£ : procedit : reputans : posse 273-4 mos : mox : Mars : Marte 274-5 graves : gravis : gentis : gentilis 285-6 mira : mira : misse : modo 356-7 plurima : prava : potest : plangere 430-1 dampna : redire : diffugiunt : dwcii 441-2 moritur : m.oriente : mori : moriente 462-3 moriendo : raoW : moriens : mors § 2. One alliterating word in the first, three in the second verse line 62-3 oreve : breve : oret'i : brevibus 231-2 p'raps : parai : properat : prior 340-1 prolixa : perlargum : pes : pectus 359-60 prelia : pariter : prelia : plaga 422-3 probus : pedites : pectora : pi7is AND IN THE CARMEN 45 § 3. Three alliterating words in the first, one in the second verse line 13-4 regnum : regni : rex : rogo 35-6 da : die : do : die 54-5 furore : furor : furens : furor 102-3 visis : visa : videre : visa 138-9 feri : ferire : ferum : feritas 219-20 scit : seduci : sedicione : solvere 321-2 prius : passa : passim : passum 358-9 viros : viduata : viris : vix 396-7 fugiens : fugiente : fugit : feriens 399-400 leditur : Jecfa'i : ledentem : lesus 409-10 hostem : hostis : hostem : hostis § 4. Two alliterating words in the first, three in the second verse line 33-4 breve : breve : brevis : brevis : brevis 67-8 regna : rege : regnis : regna : remota 179-80 reditum : redeuntem : rear : redire : ratus 221-2 mde£ : msa : viros : vast's : viris 311-2 gravis : gravatur : gravior : gemitus : gravi 315-6 viso : msa : victor : vindice : victus 322-3 cede : cecfo's : corde : cede : cruentus 363-4 Martem : mortem : Marte : mestus : magis 413-4 voce : veniam : veniam : veniale : videt § 5. Three alliterating words in the first, two in the second verse line 111-2 probus : reprobus : improbitate : improbitas : probum 127-8 parere : pares : parcet : parcat : parcere 237-8 vincere : vmc* : virtus : vmci : vincere 256-7 regum : remanet : rege : Rollandum : rm£ 263-4 feriente : /eros : ferit : ferum : ferox 331-2 ia?)x : £o£ : tanta : tarn : tot 46 ALLITERATION IN THE ROLAND § 6. Three alliterating words in the first, three in the second verse line 143-4 probus : probitas : probanda : probitas : probitate probat 208-9 ducit : ducere : duces : ducum : duce : redwcta § 7. Three alliterating words in the first, four in the second verse line 303-4 viri : vires : viriles : viris : vires : vincere : viros § 8. Four alliterating words in the first, three in the second verse line 95-6 satis : satis : satis : satis : satis : satis : satis § 9. Four alliterating words in the first, one in the second verse line 93-4 micat : mage : mane : micante : magnificat 295-6 tota : tuta : Turpini : tota : timens § 10. One alliterating word in the first, four in the second verse line 326-7 poterit : pauci : pugnant : pauci : pauci § 11. Four alliterating words in the first, two in the second verse line 243-4 animus : animos : animavit : arma : armis : arma § 12. Two alliterations of more than three alliterating words each 195-6 Gallis : visis : Gallia : visa : videri : Gallia : visis : Gallia : visa III. Alliterations extending over three verse lines. A. Alliterations of four alliterating words. 1. model: a/a : a/a 190-2 refert : rex : regnum : retro 264-6 turba : turma : turmas : timent AND IN THE CARMEN 47 391-3 patris : perdens : perditur : preveniens 419-21 tercius : tercius : tercius : Turpinus B. Alliterations of five alliterating words. § 1. model: a : a / a : a / a 97-9 veste : vestitur : vestem : vestis : vicem 233-5 Martem : minatur : minans : minis : manet § 2. model : a / a : a : a / a 6-8 mente : magnum : mirum : mirificabant : maxima 87-9 regis : regem : reperit : retro : regem 213-5 terrent : terribiles : terribiles : timor : timet 248-50 ratus : reges : regna : regentes : regibus 454-6 mori : moriens : magis : mortem : mori 459-61 credunt : comu : capere : curat : cornu § 3. model : a : a / a / a : a 50-2 furor : feritas : /e^Te : /wror : /erai 366-8 pudent : pudor : perpes : obprobrium : pudor § 4. model : a / a : a / a : a 335-7 equum : equus : eques : equum : egm C. Alliterations of six alliterating words. 76-8 pa/ra£ : pavor / parum / primum : procul : prope 79-81 timore : timor / timet : timidum : timor / terret 84-6 instimidant : instimidatus / incessanter : abire / ira : ire 223-5 hostis : hostes / haut : hostes / hunc : hostes 260-2 pereunt : pariter : pereunte / perire / pariter : pari 286-8 gentis : gens / gens : gens : gens / gens 350-2 mille / Marti : Mars : Marte / Martem : Mars 415-7 pedibus / proh : pudor : proh : perit / plus D. Alliterations of more than six alliterating words. 119-21 potis : potestas / possit : perdere : posse / plures : plurima 48 ALLITERATION IN THE ROLAND 121-3 reges : regna / reges : regna / rex : regibus : regno, 158-60 minis : minans / magis : magis : ?nage : magis / mwnera : magis : mens : mota : magis 209-12 pars / pars : pars : pedes / pars : parat : pars / pars IV. Alliterations extending over more than three verse lines. 103-6 milia / millia : mille : metus / miratur : mira : mira / mirans : Marsilium 234-7 omnia / omnibus : omnes / opus : ope / optima 401-4 vulneribus : vulnera / vulnerat : viros / velut / visus : vigor Let us now determine the number of alliterations in both poems, the percentage of the verses containing alliteration, and the number of each kind of alliteration according to the divisions and subdivisions followed in the tables given above. Some preliminary remarks will be necessary to avoid er- rors. When speaking of the number of alliterations, I mean the number of groups of alliterating words, each group rep- resenting a unit, no matter whether it consists of two or more alliterating words or whether it extends over one or more verse lines. Each alliteration has been mentioned only once in the above tables. Thus the alliteration arbres : amont (Rol. 2874) oc- curring with another in the same verse line, has not been quoted in the table containing alliterations of two alliterating words (within one verse line), under the heading §§ 4-9, but only in the table of more than one alliteration in one verse line. It will, however, be counted in the total number of allitera- tions belonging to the subdivision in question. The case is different with alliterations extending over two or more verse lines, such alliterations not counting as simple alliterations, that is, as alliterations of two alliterating words. Alliteration extending over more than one verse line has been considered intentional when out of two verse lines at least one contained two alliterating words, or when out of three at least two AND IN THE CARMEN 49 contained two alliterating words each. In a few instances it was found necessary to mention the same alliteration twice : when two alliterating words making part of an alliterative group of three or more, extending over more than one verse line, occurred with another alliteration in the same line, the example has been quoted twice, first in the table of more than one alliteration in one verse line, second in the table of allit- erations consisting of more than two alliterating words. Such is the case with the alliteration mediocres : mediocri {Carmen, 477). In the following summary R stands for the Chanson, C for the Carmen. R Total number of alliterations 731 Alliterations within one verse line 488 A. Alliterations of two alliterating words 466 297 139 112 I 1-3 | 24 | 1-4 22 | 1-5 | 17 | 1-6 | 16 1-7 | 13 | 1-8 | 12 | 1-9 | 14 1-10 10 | | 2-3 | 3 | 2-4 16 | 2-5 6 | 2-6 | 14 2-7 8 | 2-8 | 8 | 2-9 13 2-10 14 | 1 1 3-4 4 1 3-5 10 | 3-6 1 13 3-7 11 ! 3-8 | 9 | 3-9 18 3-10 9 1 t 1 1 |4-5 8 | 4-6 | 13 4-7 16 | 4-8 | 7 | 4-9 23 5 4-10 5-10 16 | 1 1 1 | | | 5-6 | 4 5-7 9 ! 5-8 1 8 | 5-9 2 1 1 1 1 1 | 1 1 | 6-7 2I6-8| 5 | 6-9 14 8 9 6-10 7-10 8-10" ~9-10~ 15 | 1 1 1 | | I 1 |7-8| 2 | 7-9 6| 1 1 1 | | 1 1 1 118-9 81 1 1 1 I 1 1 1 1 1 1 1 1 1 2| II. III. IV. R B. Alliterations of three alliterating words 20 C. Alliterations of four alliterating words 2 Alliterations extending over two verse lines 216 A. Alliterations of three alliterating words 193 §1. Two alliterating words in the first, one in the second verse line 106 §2. One alliterating word in the first, two in the second verse line 87 B. Alliterations of more than three alliterating words 23 Alliterations extending over three verse lines 24 A. Alliterations of four alliterating words 21 B. Alliterations of more than four alliterating words Alliterations extending over more than three verse lines 3 24 3 128 75 40 53 27 4 23 50 ALLITERATION IN THE ROLAND How are these figures to be interpreted? The total of all alliterations in R is 731 in 4002 verses, in C it is 297 in 482 verses. The percentage, then, of the verses containing allitera- tion is 18.3 for R, 63.7 for C, which means that in the Latin poem the percentage is about 3.5 times as large as in the Chanson. It may be of interest to know the proportion which each of the different kinds of alliterations takes in the total number in both poems: I shall give the percentages in tabu- lated form, indicating the divisions and subdivisions by their symbols. Symbol Percer itage I, A B C R 63.7 2.7 .3 66.7 C 37.7 8.3 1.0 Total 47.0 II, A B 26.5 3.1 25.0 17.8 Total 29.4 42.8 III IV 3.3 .4 9.1 1.0 Considering first the alliterations within one verse line, we see clearly from the table that the percentage of alliterations of more than two alliterating words is larger by far in the Carmen than in the Roland (about 3.1 times as large), where- as the percentage of alliterations of two alliterating words is larger in the Chanson than in the Carmen (about 1.7 times as large). Considering the alliterations extending over two verse lines, we find that the percentage of the alliterations which consist of three alliterating words in the Roland exceeds somewhat the percentage of the corresponding group in the Carmen. On the other hand, the percentage of alliterations consisting of more than three alliterating words is 5.7 times as large in the Latin poem as in the Chanson. As for the total of the al- literations within one verse line, its percentage is 1.4 times as large in the Roland as in the Carmen, while on the other hand the percentage of the alliterations extending over two AND IN THE CARMEN 51 verse lines is 1.4 times as large in the Carmen as in the Ro- land, which means that the proportion of the two correspond- ing groups is just the reverse. As for the alliterations extending over three and more than three verse lines, we see that the percentage of the former is 2.8 times, that of the latter 2.5 times as large in the Carmen as in the Roland. Summing up our conclusions, we may say : 1. Practically all types of alliterations which are found in the Roland are found also in the Carmen, 2. The percentage of the total of the alliterations as well as that of the different types is different in both poems. C prefers alliterations of more than two alliterating words and such as extend over more than two verse lines, whereas R favors alliterations of two words only and within one verse line. In other words, the author of the Carmen used more and more complex allitera- tions than the author of the Old French poem. The reason for this is not difficult to find. The genius of the Latin language undoubtedly favors the use of complex alliterations, owing to its synthetical character. On the other hand, the unit of the Latin verse line was not so strongly felt as was the French ten-syllable verse, hence the desirability of introducing another factor which might add to the structural beauty of the poem. The figures give us no information as to the relation- ship of the two poems. We see clearly that alliteration is used as a poetic device in both ; but it is impossible to conclude from them whether the poet of the Carmen, noticing the many al- literations in the Roland, imitated this device, overdoing it, favored as he was by the character of the Latin language, or whether Turoldus had before him the Carmen and introduced the same device he saw used there into his own work, his own good taste and the genius of the vernacular tongue preventing him from falling into the artificiality which is so prominent in the Latin poem. 52 ALLITERATION IN THE ROLAND It will be necessary, then, to examine more closely the allit- erations of both poems and try to find some common charac- teristics. From the tables given above we see that the large majority of the alliterations contained in the Carmen are rhetorical figures, which more than the alliterations themselves con- tribute to confer upon the poem this mark of artificiality. As a matter of fact, the total of all rhetorical figures is 236 in 297 alliterations, that is, a percentage of '79.5. These rhetorical figures may be subdivided into four groups : 1. a word alliterates with itself in the same form 2. a word alliterates with a form of different inflexion 3. a word alliterates with one of its derivatives or a de- rivative of its root 4. only the prefix of the two alliterating words is the same, the roots differing. In cases where the alliterating group consists of more than two alliterating words we may have a mixture of these four classes. In order to avoid counting one example twice, cases representing a mixture of classes 1 and 2 have been counted in 2 only, as the more general class; such as show a mixture of 1 and 3, or of 2 and 3 have been counted in 3 as the most general class of the three. There is only one case where 4 coincides with another group in the same alliterative unit; the example in question has been reported under the heading of 4 and omitted in the other class. A computation gives the following results: Class 1 44 examples Class 2 118 Class 3 : 72 Class 4 : 2 Total : 236 AND IN THE CARMEN 53 The percentages of each are : Class 1: 18.6 Class 2: 50.0 Class 3: 33.5 Class 4: 0.9 The Roland likewise contains a certain number of rhetori- cal figures, which will be grouped below and classified from the same point of view as those of the Carmen: Class 1: 57-8 testes : testes 1483-4 molt : molt 349-50 tant : tant 2093 alquanz : alquanz 409-11 tint : tint 2185 cerchet : cerchet 411 mot : mot 2229 at : at 958-9 veit : veit 2235 guardet : guardet 1002 fut : jut 2271 halt : sont : halt : sont 1011 granz : granz 2306-7 tantes : tantes 1015 ont : ont 2308-10 at : at 1022 tanz : tanz 2641 laissent : laissent 1030-1 sont : sont 3038-9 sont : sont : sont 1190-1 Franceis : Franceis 3585 nut : nut 1439-40 milliers : milliers 3701 mandet : mandet Class 2 : 143 (List : dit 1105 fait : ferat 227 mort : moerjuns 1285-6 met : mist 240 fereit : fesist 1399-1401 tante : tant : 439-40 fut : fust tante : tant 480 mul : mule 1417-8 moerent : mort 496 dist : dit 1539-40 met : mist 540-1 tantes : tanz 1715-6 dist : dis 553-5 tantes : tant : tanz 1744 venget : vengier 757 mul : m/ide 2332-4 teneit : tient 794-8 vint : vint : vint : 2419-20 plort : plorent vint : vint : venuz 2713 rei : re'ine 969 Franceis : France 2744 fil : fille 983-4 dient : dist 2824 sire : seignor 1023-4 feront : faite 3980-1 Deu : Deus 54 ALLITERATION IN THE ROLAND Class 3: 33 charre : chareier 1466 49-50 Franceis : Franc : 1486 France 1594 107 Gerins : G criers 1604 131 charre : charreier 1881 163-4 matin : matines 1895 174 Geriers : Gerins 2147 177 Frans : France 2186 253 Sarrazin : Sarragoce 1 2404 669-70 matin : matines 2641 701-2 Franc : France 2670 804 Frans : France 3123 808 Franceis : France 3226 938 Franceis : France 3254 989 Franceis : France 3408 1210-1 France : Franc 3833 1438 Franceis : France chancon : chantede codart : codardie Malquidanz : Malcud Franceis : Frans monies : mostiers Yvorie : I von jorz : ajornez Gerin : Gerier Gerins : Geriers Marbrise : Marhrose Clarifan : Clariien altretel : altre Sorbres : Sorz Hums : Hongres regnes : rets fel : felonie The total number of all rhetorical figures in the Roland is 82, which are distributed over the three classes in the follow- ing way : Class 1: 24 Class 2: 26 Class 3: 32 The examples which might have been counted under the last heading have not been considered, since in the Roland it is extremely doubtful whether a rhetorical figure was intended by the author, the combinations in question being perfectly natural, as for instance: aval : amont (Rol. 2235). There 1. Sarrazin and Sarragoce, from Caesaris Augusta, would, of course, not be considered as a rhetorical figure in our sense of the word. Still, when seeing the combination one cannot help thinking that the author of the poem believed them derived from the same etymon and related to each other, that the combination was meant by him to be a rhetorical figure. AND IN THE CARMEN 55 are, moreover, very few cases belonging- to this class. The per- centages of the other three are: Class 1: 29.3 Class 2: 31.7 Class 3: 39.0 The percentage of the total of all rhetorical figures is 11.2 for the Roland, 79.5 for the Carmen. We see, then, that the percentage of all rhetorical figures is for the Carmen seven times the percentage of those in the Roland. As for their distribution over the different subdi- visions in both poems, we may say that in the Carmen class 2 shows the largest number of examples, in the Roland class 3. The reason for this is doubtless that the synthetical char- acter of the Latin language favored the formation of rhetor- ical figures of class 2, which is not the case in the modern languages. Coming to our main conclusion, we must say that it is im- possible to see any influence of the Carmen on the Roland as far as rhetorical figures are concerned. For the moderate use of this poetic device, Turoldus may have found precedent in almost any work of Classical or Mediaeval Latin literature. There is nothing which favors the supposition that the Car- men was his model, unless we assume that this model did not exercise any influence upon his own work as far as the form is concerned. On the other hand, the Chanson hardly exer- cised any influence on the Carmen with regard to the form. To explain this we must bear in mind that the latter undoubt- edly represented a higher genre, being written in Latin ; the author would probably not have sought his model among works of the vernacular, but would rather have looked to specimens of Latin erudite literature, and there he would have found examples enough of the exaggerated and artificial use of rhetorical figures. It will be of interest to see to what an extent the author of the Carmen drew on Classical and Mediaeval Latin litera- ture for the alliterations contained in his work. The groups 56 ALLITERATION IN THE ROLAND of syntactically coordinated alliterating words found in Clas- sical and Mediaeval Latin writers having been collected by Wolfflin and Ranninger, it will be of value to quote those which occur in the Carmen and to note those among them which have been found in the works of previous writers. In the following table they will be given in alphabetical order 1 . micat : magnificat obest : obit parat : properat patet : procedit patricios : proceres : pedites perdens : preveniens pes : pectus (presignis : prestans:) pins : pot ens R 34 reges : regna R 36, S 80 remanet : ruit stat : secat timet : terret timidus : tutus urbes : urbanos We see that the large majority of the alliterating groups found in the Carmen have not been quoted by the two authors who have treated this subject in Classical and Mediaeval Latin literature. Whether or not these combinations of alliterating words are the inventions of the author of the Carmen cannot be decided before an exhaustive study of the whole field down to the twelfth century will have been undertaken. 178 accipit : adit 93-4 376 audit : agnoscit 306 245 cassis : clipeus 232 3-4 clipeus : contemptor 223 94 decus : decor 473 466 dux : decus 348 ferit : fugat 392-3 282-3 fluens : fsetinans 341 240 fugando : fugiendo R 27, S 59 5-6 305 fugat : ferit 121 241 Gero : Gerinus; also: 256-7 267 427 7 magnum : mirum R 29, 80-1 — — _ — S 68 III, 449 280 229-30 Mars : mors 126 363 Martem : mortem 1. The figures indicate the page of Wolfflin's and Ranninger's ar- ticles on which the examples in question are quoted: III referring to Wolfflin's study in the Archiv fur lateinsche Lexikographie und Gram- matik, III, S to his article in the Sitzungsberichte der Kgl. Bayerischen Akademie der Wissenschaften, Philosophisch-philologische Classe, Mun- chen, 1882, R to the study of Ranninger, entitled Ueber die Alliteration bei den Gallolateinern. AND IN THE CARMEN 57 The next question to be answered touches the position of the alliterations in the epic verse of the Roland. Did the poet scatter them all along the verse line or did he follow a certain scheme in using them? Examining all alliterations of two alliterating words with regard to their position in the ten-syllable verse and grouping them in the order of their totals, we arrive at the following result : 24 | 23 | 22 | 18 | 17 | 16 | 15 | 14 1-3 I 4-9 | 1-4 | 3-9 | 1-5 | 1-6 | 2-4 | 4-7 | 4-10 | 6-10 | 1-9 | 2-6 | 2-10 | 6-9 13 I 12 I 11 I 10 I 9 1-7 | 2-9 | 3-6 | 4-6 | 1-8 | 3-7 | 1-10 | 3-5 | 3-8 | 3-10 | 5-7 | 8-9 8 | 7 ! 6 | 5 | 4 | 2-7 | 2-8 | 4-5 | 5-8 | 7-9 | 8-10 | 4-8 | 2-5 | 7-10 | 5-9 | 6-8 | 3-4 | 3 | 2 2-3 | 5-6 I 5-10 | 6-7" f 7-8 | 9-10 Let us try to interpret the figures of this table. Allitera- tion group 1-4 represents a linking of the first and the last word of the first hemistich, if the last word be a monosyllable or a dissyllabic word with a feminine e in the second syllable. 1-3 represents the same kind of linking in cases where the last word of the first hemistich consists of two syllables, the last being masculine, or of three syllables, the last being fem- inine. It follows, then, that alliteration serves in the first place to unify the first hemistich of the ten-syllable verse and to mark the csesura more strongly than would have been the case without alliteration. 4-9 means a linking of the last word of the first with the last word of the second hemistich, just as 4-10, 3-9, and 3-10, according to the number and char- acter of the last syllables of the alliterating words. All of these, with the exception of 3-10, are represented by large numbers of examples. 1-5 means a linking of the first word of the first hemistich with the first word of the second. A modification of these cases is seen when the first word of the hemistich is an unstressed monosyllable: article, preposition, pronoun, etc. Then the second word takes the place of the 58 ALLITERATION IN THE ROLAND first, thus explaining the frequency of alliterations belonging to groups 2-4, 1-6, 2-6, 6-9, and 6-10. In the second and third place, alliteration serves to link the two hemistichs of the epic verse, or, in the case of 6-9 and 6-10, to unify the second hemistich. 1-9 and 1-10 mean a linking of the first and the last word of the verse, thus marking it off from the pre- ceding and following one. As we see from the table, the ex- amples of 1-10 are not quite so numerous, but there are many examples for 2-9 and 2-10, which represent verses where the first word is an unstressed monosyllable, the stress falling therefore on the second syllable. The explanation for the rather frequent instances belonging to group 1-8 is found in the fact that out of the twelve examples there are ten the last word of which consists of three syllables, the third being mas- culine, or of four, the fourth being feminine with regard to the ending. This group serves, then, the same purpose as 1-9, 1-10, 2-9, and 2-10, that is, to mark off the verse line. The only group having numerous examples and the purpose of which cannot be seen is 4-7 with sixteen examples ; it may be that this comparatively large number is due to chance. The total amount of alliterations which can be proved to serve a definite purpose in the structure of the verse is about 270 (con- sidering only alliterations of two alliterating words) , that is, 57.7 per cent of the whole number of alliterations belonging to this subdivision. It is therefore certain that we are here real- ly in the presence of an example of the use of alliteration as a poetic device 1 . As for the alliterations consisting of three or four alliterat- ing words and extending over one verse line, no conclusion can be drawn from the examples found in the Roland, the total amount being too small. The same must be said with regard 1. The fact that the alliterating words are preferably placed at prom- inent parts of the verse line, was noticed already by Kohler, Ueber al- literierende Verbindungen, pp. 100-1: Auch scheinen es manche Dichter zu lieben, die Caesur in ihren Versen dadruch recht deutlich hervor- treten zu lassen, dass sie vor die Caesur und an den Versschluss allite- rierende Worter stellten. But it has never been worked out fully, nor in AND IN THE CARMEN 59 to the examples of more than one alliteration occurring in one verse line. Examining the alliterations extending over more than one verse line, we find that out of 243 examples there are 59 in which the alliteration falls on the same syllable in two suc- cessive verse lines, the percentage being then 24.0. In five cases the syllable in question is the tenth, so that in 2.0 per cent of all cases alliteration and assonance fall on the same syllable in two consecutive verse lines. Considering the alliterations of both poems with regard to the alliterating letters, and grouping them in the order of their totals we get the following tables : 1. Roland Total | 113 | 101 | 85 | 62 | 43 | 54 | 52 | 41 | 38 | 33 Symbol p I m | f | d | k | r | t | v | b | s Percentage 15.4 | 13.8 | 11.6 | 8.5 | 5.9 | 7.4 | 7.1 | 5.6 | 5.2 | 4.5 Total | 27 | 20 | 14 | 9 | 7 | 6 | 4 | 3 | 1 Symbol ts | a | e | j | g | 1 | i | n | o | u | ts Percentage | 3.7 | 2.7 | 1.9 | 1.0 | 1.2 | .8| .5 | .4 | .4 | .1 | .1 2. Carmen : Total | 41 | 37 | 26 | 25 | 23 | 22 | 19 | 15 | 13 | Symbol |m|p |v|k|r|s|f|t|d|a| Percentage | 13.8 | 12.5 | 8.8 | 8.4 | 8.4 | 7.7 | 7.4 | 6.4 | 5.1 | 4.4 | Total | 11 | 9 | 7 | 6 | 4 | 3 | 2 Symbol | g | e | i | o | h | 1 | j | n | b | u "Percentage \ 3.7 | 3.0 | 2.3 | 2.3 | 2.0 | 2.0 | 1.3 | 1.3 | 1.0 | .7 We notice first that in the Roland there are 42, in the Car- men 38 vowel alliterations, the percentage being 5.7 and 12.8 ; that is, the proportion is more than twice as large in the Car- detail for any definite poem of Old French literature. On the other hand, G. Korting's statement in his Encyklopaedie und Methodologie der Romanischen Philologie, Teil II, Heilbronn, 1884, p. 424: "Die Ver- bindung der Vershalften durch die Alliteration kennt das Romanische nicht, es verwendet vielmehr die Alliteration nur gelegentlich in rein onomatopoietischer Weise," must be modified accordingly. 60 ALLITERATION IN THE ROLAND men as in the Roland. This result is not surprising since the number of vowels is much smaller in Latin than it is in Old French. Comparing the single sounds, we note that the per- centage is the same in both poems for m. We also see that the eight sounds showing the highest totals are the same in the Chanson and in the Carmen, with the exception of a in the Roland and s in the Carmen. The order of the different sounds agrees in a general way with the result deduced by Ranninger 1 , except for the sounds m and r, for both of which there are numerous examples in the Roland and in the Car- men, but which the German scholar did not find to be promi- nent in the Gallo-Latin authors. To sum up the results obtained in the course of this study, we may say that three main conclusions must be drawn from the facts as they have been presented : 1. With regard to their origin, the alliterations of the Roland must be divided into three classes: the first, contain- ing syntactically coordinated alliterating words, is in large part of popular and traditional origin. The second, repre- senting the alliterating personal proper names, comes from a Teutonic source, though it is impossible to determine wheth- er this source must be sought in continental Germanic, Anglo- Saxon or Old Norse literature, or whether it is literary at all, or merely due to Norman tradition. The third, by far the largest because it comprises all the rest, goes back to Mediae- val Latin literature, which is certain to have exercised the most powerful influence on the Chanson. In this point the result confirms the conclusions arrived at by Wilhelm Taver- nier. 2 2. The Carmen de prodicione Guenonis exercised hardly any influence upon the Roland with regard to the form. The percentage of its verses containing alliteration is 3.5 times as large as that of the verses of the Roland. What is still more important is the fact that the large majority of the allitera- 1. Ueber die Alliteration bei den Gallolateinern, p. 99. 2. cf. Zeitsch. f. franz. Spr. u. Lit., XXXVI 1 , 1910, 71, XXXVII 1 , 1911, 83. AND IN THE CARMEN 61 tions found in the Carmen are rhetorical figures, which play an entirely subordinate part in the Roland. It is impossible to draw any conclusions from this comparison as to the prior- ity of the Roland or the Carmen; for Turoldus may have in- tentionally avoided imitating the doubtful art of the Latin poem, even if the latter were his model as far as the contents are concerned. 3. Alliteration in the Roland serves a very definite pur- pose in the structure of the epic verse. 62 ALLITERATION IN THE ROLAND APPENDIX Alphabetical Index of the Alliterations of the Chanson de Roland 1 afubler : altre 3941 Aide : aler 3723 altre : avison : Ais 2555-6 amont : Aspre 1103 arbre : amont 2874 arme : avenant 1154 asez : aveir : abez 2955 aveir : ainz : aler : arpent : avant 2229-31 aveir : altretant 3021 avefr : amor : altretel : altre 3122-3 ait : aZer 1184 aiire : aveir : atendre 3740-1 altre haltement 2597 aproismer : amont 2692 arcevesque : aZer : aidier 2169 asembler : Ais 3744 avant : aler 2858 areir : aZ£re 2309 arefr : a££re : angoissos 822-3 baisier : boche 1530 baroe : 6Zemc 2334, 2930, 3618, 3712 barbe : blancheier 261 baston : brief 341 bel : Baliganz bel : bon bien : baron : blanc 3201 3047 barbe 3172-3 Baliganz : baron : baston 2686-7 barbe : blanc : bronie 3087-8 baston : batre 2588 bataille : blecier : blesmir 589-90 bel : Baliganz : bon 3184-5 bien : baisier : boche 632-3 bien : batre 3739 1. Nouns and adjectives have been given in the oblique of the singu- lar, adjectives moreover in their masculine form, verbs in the infinitive, if the alliteration did not disappear thereby. In the latter case the form occurring in the text has been chosen. Proper names of persons have been put in the nominative. The rhetorical figures are not included in this list. AND IN THE CARMEN 63 Men : batre : baston 1825 blanc : barbe 117, 2943, 4001 blanc : bloi 999 boche : buisine 3523 brandir : abatre 1957, 3929 brochier : Men 1536, 1573, 1944, 2128, 3877 bronie : barbe 3122 Men : brandir 1249 bfcmc : freZ 2250 blanc : bloi : baron 1800-1 bon : bel 3064 brief : Babilonie : Baliganz 2613-4 brochier : Men : Boeve : Belne 1891-2 bronie : barbe : blanc : baron 1843-4 Capuels : Capadoce : clamer 1614-5 chameil : chargier 645 champ : chevalchier 1338 Chanelius : chevalchier 3269 chargier : charre : chareier 32-3 Charles : chanceler : chadir 3608 Charles : chapele 52 Charles : chier charre : charreier 3031 131 chevalchier : champ 3463 chevalchier : chemin 405 chadir : Charlemagnes 3720 champ : Charles 1928 chancon : chanter 1466 charette : chemin 2972 Charlemagnes : chanut 538, 551 Charles : chanut 2308 Charles : chevalchier 2454, 3234, 3695 Charles : chier : chevalier 1560-1 cheval : Chernubles : charboncle 1325-6 chevalchier : Charlemagnes 1842 chevalchier : chief 3018, 3195 chevalier : chadir 3486 chevel : chaitif 2596 chrestien : reclamer : crider chrestien : conoissance 3987 3998 clamer : co?p^ clamer : quite 3800 1132, 2239, 2383 clamer : quite : conte 3809 esclargier : coer 3628 Clarifans : Clariiens 2670 codart : codardie • • 1486 64 ALLITERATION IN THE ROLAND code : crignete 1494 descolorer : cler : cors 1979-80 colp : cuidier : cors : costet 1505-6 coer : codarder colp : conoistre 1107 2875 comander : conte contender : conseil compaignon : cors : contrevaleir conduire : cors : col conte : codardise conte : comander 3057 179 1983-4 3370-1 3531 con 1816-7 conte : conquerre : clamer : colpe 2363-4 cors : cler : contenant 3115-6 cors : contenance 3086 cors : costet 1506 corteis : conte : creidre 576-7 recredant : recomencier : crit 2063-4 colper : cort : quisse : crope : costet : code : crignete 1491-4 combatre : compaignon 878 compaignon : corn 1051 conduire : cors 892 conquerre : Commibles 198 cewie : codart 2134 conie : combatre : cors 2099-2100 corre : crote 2580 cors : cler : compaignon 1159-60 cors : color 2895, 3763 cors : contenant 118 cors : quatre 2080 costet : conduire : cors 1315-6 creidre : chrestientet 3980 damage : doZor : departir 3479-80 demener : dolor : deu 2695-6 derompier : cfows : desevrer 3466-7 desevrer : c£os 1201 destre : Deu : descendre 2373-4 destre : (Zire : demander 3845-6 Dezt : doner 289, 2241, 2938 demander : doicement 1999 dementir : Dew : ^'re : dre{£ 3890-1 descendre : De% : benedire 1136-7 desrengier : destreindre : descendre 809-10 destre : Deu : dolor 1195-6 Deu : demander 3980 Deu : doner : benedire : dote: 2016-7 AND IN THE CARMEN 65 Deu : dons : dreit 3898 deveir : dire : Deu 333-4 deveir : dire : dus 242-3 devenir : doloros 2722 dire : dame 2724 dire : Damnedeu 3906 dire : Damnedeu : doner 1898 dire : dechedeir 1628 dire : dehet 1047, 1938 dire : deignier 1716 dire : demander 3846 dire : desfaire 450 dire : destre 47 dire : destre : deit 508-9 dire : Deu dire : Deu : demander 3557-8 123, 428, 676, 1608, 3993 dire : Deu : desmentir 787-8 dire : Deu : dolor 1579 dire : diable 746, 983 dire : deignier : damage 1715-7 dire : dolent 2823 dire : dolor 3627 dire : dolor : demener 2945-6 dire : dreit 496-7 dire : dux 2944 dis : dolor 2301 dolor : damage 2983 dolor : degret : descendre 2839-40 dolor : desevrer 1977 dolor : diable : doner 3646-7 dolor : dire : Deu 1631-2 pardoner : Deu 2007 doner : diable 1502 dous : desaffrer 3426 doze : deveir 1346 doze : dis : dous 1308-9 drecier : dolor 2234 drodmont : dire 2730 duire : detordre 772 Durendal : doner : destre 2780-1 dus : don 247 eschiele : establir 3217 eschiele : establir : estre 3068-9, 3237-8 escut : espiet 1799, 3114 esf order : eschange 3714 Espaigne : espleitier 2165 Estramariz : Eudropins 64 estreit : estre 356 evesque : eue 3667 faire : felonie : faillir 2600-1 faire : fier faire : fol faire : fort : fier faire : Franc faire : Franceis faire : fust 1053 faire : forsfaire 2125 faire : fosset 561 faire : Franc 807-8 faire : fui 1779 faldestoel : /ci2re 897, 1111 2029 3166 700-1 2371 115 66 ALLITERATION IN THE ROLAND felon : faire felon : ferir fer : fuM ferir : Valferrede 844, 3833 2144 1602, 3154 1370 ferir : forcele 1294 ferir : fraindre : fendre 3603-4, 3926-7 ferir : fraindre : flor 1353-4 ferir : France : fraindre : reflamber 3615-6 ferir : Franceis : froissier : forbir 3481-2 afichier : fut : fort : fin 3393-5 2420 2744-5 fil : fredre fil : fut Flameng : Frisie 3069 fo'ir : Franc 2844 fort : fier 1879 fou : flambe 2535 fraindre : desconfire 1247, 1305 France : ferir 1210-1 Franceis : faire 278 Franceis : felon 1190-1 Franceis : ferir : fol 1438-9 Franceis : fo'ir : faire 1255-6 Franceis : fus fredre : faillir front : faire fut : ferir fut : fort 1604 1866 3919 439-40 2278, 3382 felon : fer 3735 felon : ferir : forbir 1924-5 ferir : faire 1185 ferir : force : fraindre 1618-9 ferir : fou : faire 3916-7 ferir : fraindre : desconfire : fort 1304-6 ferir : Franc 1234 ferir : Franceis 1259, 1937 ferir : fut : froissier 2288-9 fier : faillir 3133 fil : fredre : furent 3411-2 flambeios : faire : Franceis : felon 1022-4 flor : faire 1856 force : Franc 1660 fort : freit 1118 fraindre : fendre 3604, 3927 Franc : felon : ferir 3558-9 France : flor 2431 Franceis : desfaire 49 Franceis : ferir 454, 1080, 1347, 1416, 1652, 1654, 1662, 1939, 3203 Franceis : France : florir Franceis : florir 969-70 Franceis : fo'ir : faillir 1047-8 Franceis : fustes 2475 fredre : faire : comfaitement 1698-9 fustes : fil 2208 fut : reflamber 1002-3 AND IN THE CARMEN 07 Gabriels : guarder 2847 Gefreiz : Gerins : Geriers 106-7 Gerins : Geriers Gefreiz : Jozerans 3535 174, 2186, 2404 grant : guerre 242 gualt : grant 2549 riedreguarde : grant 883 guarder : gloton 1230, 1251 Guenles : grant 3969 Hum : Hongre 3254 irat : itant jorn : ajorner lancier : laissier loder : legerie lune : luire Machiners : Maheus Malbiens magne : dementir maile : menut major : metre mat : molt : Marie Malprimes : molt malvaisement : mielz mander : Marsilies : mander : mercit mar : Marsilies Marbrise : Marbrose Margariz : molt Marsilies : marchiet Marsilies : mi 2734 isnelement : issir 2766 2147 2154 leon : leupart 1111 206 lone : larc 2852 2512 • « magne : mar 1949 66-7 1404 magne : muder 841 1329 main : comander 696 1532 major : molt 1784 1472-3 mal : mot 1190 3176 Malquidanz : Malcuz 1594 : morir mander : Marsilies 680 1517-8 nain mander : Marsilies : molt 222-3 125-6 239 comander : Miles : march is 2970-1 196 marbre : metre 2966 2641 Val-Marchis : Malprimes : 1131 1150 1449 mercit Marsilies Marsilies Marsilies 3208-9 mander : morir 2770-1 martirie 1467 molt 201 68 ALLITERATION IN THE ROLAND Marsilies : molt : mer cider _ 908 Marsilies : muder 441 Mars one : morir : Malpalins : molt 2994-6 meillor : mener 502 mer : Marbrise : Marbrose : amont 2640-2 message : mander 3699 message : moZ£ : Marsilies : moZ£ 143-4 messe : molt : metre : mostier 3860-1 metre : main 443 1285, 1539 moZ£ 2934-5 : morir 1725-6 1570 3085 1115 metre : more metre : morir : mzete: : mesure mil : mang6n milie : meillor milie : metre moillier : Bramimonde : moW 2576-7 moZ£ : malement 2106 mo££ : merveillos 2466 moZ£ : Monjoie 3300 amonter : mwZe 3943 monter : mul 2811 more : mercit : manovrer 2505-6 morir : Marsilies : moW 437-8 Marsilies : trametre : mes- sage : masse 181-2 matin : messe 163-4, 669-70 meillor : mil : mangon 620-1 merveillos : menor : m^h'e 3218-9 message : mam : mo%< 418-9 message : mantel message : monter monter 2707-8 2765 mesure : medisme monter morir molt 1035-6 : message : 91-3 mi? 1306-7 2635 metre : main metre : mie : mer mielz : morir 359, 1091, 1701, 2336, 2738 31 1454 410-1 miZ : mudier : mul 31-2 milie milie menede mot molt : mal molt : medisme molt : merveillier monie : mostier monter : Marcules mordre morir : : mal demander morir : Marsilies 886 1483-4 537 1881 : mer 3155-6 727 molt 3713-4 mar 1730-1 AND IN THE CARMEN 69 morir : merveillos : amener morir : mercit 3721 3963-4 morir : message : Marsilies morir : milie 3530 3772-3 morir : molt 1690, 2010, 2232 morir : millier 1439 morir : mot : marchis 2030-1 morir : monter 227-8 mot : mie 2087 mortel : mielz : morir 3908-9 mot : monter 2457 ?>io£ : molt 2597 muder : medisme : metre : mo£ : monter : magne 3621-2 mercit 2381-3 mw£ : malvais 480-1 mul : Marsilies : mander • 847-8 mitZe : amener : Marsilies 89 nevoid : nom ont : orgoeil 1188 Nigre : noefme 1941 0i;nV : oei7 paien : pasmer 1348 pcwe?i : plaisir : porter 3537-8 paien paien paien paien pais : : poder 61, 537 : por£ : plevir 1057-8 : Preciose : perte 3298-9 : prodome 3264 paien : apres enter 3596-7 parenz : pro£ : p£em 2905-6 parler : Pinabels : plaisir : parentet 3905-7 par£ : Pinabels : per 361-2 pasmer : plusor : pro£ 2422-3 pasmer : pres 2270 passer : porfe : po?i£ 2690 pecchiet : poroffrir 2365 empeindre : passer : plein : place 1272-3 3229 2285 paien : plenier : perdre : plus 3400-2 paien : poi 1940 paien : Preciose 3471 paien : presse : prodome 1543-4 ovn'r : oez7 2285 parler : padrastre : pw£ 762-3 part : peweii 940, 1562, 1613 part : poder 2034 pasmer : pitiet : plorer : prendre 3724-6 pasmer : poder 2891 passer : port 790 passer : pui : plus : par font 3125-6 pendre : parenz : plaidier 70 ALLITERATION IN THE ROLAND per : parenz 1421 3932-3 per : prendre : pesance per : perdre 937 3710-1 per : preier 2261 perdre : peil : empleir 1012-3 perdre : piet 3720 pers : pale 1979 petit : pas : poder 2227-8 piet : poder 1218 piet : poder : paien 3167-8 piet : presenter 3792 pitet : poder : plorer 2873 pitiet : plorer 2419 plaie : pan : piz 2173-4 plait : preier 3799 plein : piet : paien 2400-1 plorer : plaindre 2915, 3722 plus : Piedres 921 plus : poder : pui 2868-9 poder : paien : parfondement 973-4 poder : plorer 773, 841, 2193, 2381 pooler : poldros : paienor 2426-7 poder : preisier 1683 podestet : prendre 417 poign : depecier 837 Poillain : Polerne 2923 porprendre : pafi 3332 por£ : passer 944, 1766, 2772 porter : paredis 2396 prametre : prendre 1519 paien : perie 568 perdre : per : parenz : por£ 1420-2 perdre : p2ai£ : pendre : parenz 1408-10 pesme : perdre : per 2147 pie£ : pfein : prendre 2878-9 2ne£ : p 2306-8 £em£ : tere : trenchier 553-4 tant : £o£ 285, 349-50 tel : trover 2092 AND IN THE CARMEN 73 atendre : treiit : tere 665-6 tendre : trover : tres _ 780-1 tenir : tere : tant 2332-3 tenir : trenchier : taillier 1339 tere : trestorner 2587 Valterne : tere 199 teste : torner : tot 2011-2 torner : tenebros 2896 torner : £o£ 2184 £o£ : tabor 3137 tot : tenir 409 tot : £es£e : trenchier 1371-2 tot : trametre 1608 tot : trover 2953 trenchier : teste 1586, 1956 trenchier : tres 1506 trestot : tere 2020 ^ro2> : maltalant 327 trover : tenir : £o£ 2445-6 vaillant : vertuos : vassal 1636-7 mt£ : Valerie : Viviers 1625-6 wwt : vasselage 3447 wJor : voleir : vaignet 1090-1 vanter : vaillant 2861 vassal : vestir : vedeir 3532-3 vedeir : vasselage 2981 vedeir : venir 1019, 1021, 2124, 2558 vedeir : vespre : vert 2447-8 vedeir : vont 3687 veintre : vassal : vif 2087-8 tendre : £re/ 159 temV : £ere 956 temV : Maltet : £me£ 3152-3 tens : femr : te£ 2310-1 Tervagant : toldre 2589 Tiedris : temV : £o£ 3896-7 torner : teste 2360 Torgis : Tortelose 1282 £o£ : talant : £ere 400-1 tot : tews 1858, 2244 tot : toWre 236 tot : trenchier 1299 travers : Turc : tierce 3239-40 trenchier : teste : tot 3617-8 trente : tost 3217 trofs : teste 935 trop : tens 1841 voir : visage 283 vait : FaZterwe 1291 Valfo'it : -yera'r 3239 yaZ-Teweore : vont 2461 vassal : vedeir 1694, 1857 vedeir : vent 2533 vedeir : voleir 3404 i>eie : vedeir : veirement : visage 1638-40 veintre : voleir : va£ 2042-3 vemr : vedeir 2746 74 ALLITERATION IN THE ROLAND venir : : valede 1449 venir : veltre 2563 venir : vedeir : veirement 953 venir : vieil 797-8 venir : vengier 1744 vert : vedeir 2876 venir : voleir 2974 vint : vaillant 789 vertut : vait 1551, 3878 voleir : vaignet 1091 voleir : voleir : vertuose- voleir : veintre voleir : vengier voleir : devenir vont : vol 3609 vait 1873-4 155 2472 ment 1643-4 voleir : veirement 2361 voleir : venir 2103 voleir : vivre 2936 ymage : fdele 3664 Yvorie : Ivon 1895 AND IN THE CARMEN 75 Alphabetical Index of the Alliterations of the Carmen de prodicione Guenonis 1 accipere : adire 178 agere : audire : agnoscere 375-6 Anseus : acer 317 audire : arma 293 adniti : asper 457 animus : arma 243 aspectus : auris : arduus 339 caput : cassis : clipeus 245 caro : cogere 318 cawsa : comperire 376-7 cedes : cor : cruentus 322-3 clavus : celer : cursus 343-4 corpus : cedes 452-3 credere : cornu : capere curare 459-60 crws : circumquaque : corpus 448-9 caput : conamen : cruor 373-4 CGwtsa : ci£o 382 cectis : cupido : cunctus 165-6 ceZer : certamen 431 cogere : condicio 262 costo : crws : cavus 340-1 credere : Karolus : clavis 172-3 cwra : colligere : consulere 199-200 dampnus : redire : diffugere dare : dampnus 391 : ducere 430-1 dare : dicere 35 decies : duo : dare 103-4 decus : dignus : decens 8 dimidiare : decern 291-2 dolus : dicere 28-9 donare : dare 153 dux : decws 466 equus : equare 344 eripere : equus 480 extasis : exhinc 439 1. Nouns and adjectives have been quoted in the nominative singular, adjectives in their masculine form, verbs in the infinitive, if the alliter- ation did not disappear thereby. The rhetorical figures have, as a rule, not been considered in this list. 76 ALLITERATION IN THE ROLAND fama : fuisse 465 ferire : ferus 138 ferire : ferus : ferox 263-4 ferire : fugare : ferus 348 fluere : festinare : fremere fors : ferus : fugare : 282-3 ferire 305 fugere : /enre 396-7 /tmse : ferus 324 /wror : feritas : ferre 50-1 /wror : flectere 59-60 oens : gloria gravis : o^ens 472 Gero : Gerinus 241, 267 274-5 gravis : gemitus 312 Josh's : /iaw£ : toic 223-5 tarn : tacere instare : ingens iudex : iubere 259 inquit : i£a 235 271-2 instimulare : incessans : afore : z"ra : intrare 84-7 39 Karolus : cants : cor Karolus : clipeus : contemptor latitare : latens legatus : lex 145 Karolus : clavis 3-4 173 247 ledere : Mere : locare 306-7 31 letificare : legatus 175 magis : minito.re 10-11 remanere : manus 256 Mars : Marsilius : mori 383-4 Mars : mors : mestus : magis 363-4 Marsilius : Mars : wens 329-30 mediocer : minimus 477-8 metus : mors mille : Mars 426 350-1 minari : magis : munus : major : minari 289 manus : munire 246 Mars : minari : manere 233-5 Mars : mors : mi^e 229-30 meare : mirari 178-9 niens : magnus : mints : mirificare : maximus 6-8 micare : magis : mane : magnificare 93-4 mt7£e : metus : mirari : Marsiliits 104-6 minor : magis 270-1 AND IN THE CARMEN 77 mens : mover e 158-60 minor : mox 38-9 mirus : mittere : modo 285-6 monere : remanere : Minerva 56-7 mori : metuere : remanere 434-5 mors : minor : remanere 394-5 moo; : Mton 309 multus : minari 43 mirari : mors minis : movere mori : magis mors : Mars mos : mox : Mars multus : magis multus : mori 469-70 Mars 302-3 454-5 313 273-4 99-100 239-40 nunc : nullus 128 obesse : obire : Oliverus obliquus : Oliverus 397 306-7 omnis : opus : optimus 234-7 optare : omnis 231-2 parare : pavor : parimi : primus : procul : prope 76-8 parere : parare 61, 64 pars : pedes : parare 209-11 pater : perdere : prevenire 391-3 patricius : procer : pedes 473 perdere : pax 27-8 pes : pro/i : pudor : perire : plus 415-7 pwws : placere : pars 90-1 posse : pemci : pugnare 326-7 posse : precipere 32-3 prostremus : plenus : post 400-1 prelium : posse : pafo' 388 presignum : prestare : pius : potens 5-6 parcere : prosternere 349-50 parere : parcere 127 passus : precipitare 310 patere : procedere : reputare : posse 223-4 pena- : proximus : primordium 346-7 perire : par 260 petere : passus 457 plurimi : pravus : posse : plangere 356-7 posse : perdere : plures 120-1 2^osse : proma 470-1 prelium : par : pZao/a 359-60 premium : polliceri primus : putare 220 432 78 ALLITERATION IN THE ROLAND prius : par are : proper are prius : passus : passim 321 231-2 probare : prius : posse 239-40 probus : pedes : pectus : pilum 422-3 prolixus : perlargus : pes : pudere : perpes : obprobrium pectus 340-1 366-8 ratus : rex 248-9 redire : rex : ratus 179-80 referre : rex : retro 190-2 r ^x : ratio : robor 284-5 rex : redire rex : remotus rex : retinere Rollandus : referre 19-20 rex ' remanere : Rollandus ruere 67-8 r e% ' reperire : retro 292 rex : rogus 453 Rollandus : rex 256-7 87-8 13-4 151, 205 Rollandus : ruere 265, 278 satis : sohts scelus : sedicio scopulus : supplex seducer e : sedicio simul : summus Siria : insidiae species : solus stare : studere 471 scandere : scopulus 377 scire : submonere 457-8 sedere : sinister solvere sexaginta : superesse 219-20 476-7 singulus : sufficere 73 soZ?ts : similis 140 store : secare 82 summa : superstes tentorium : turma 191 tantus : tergum : timere : talis 300-1 tercius timere : terrere 80-1 terrere tot : tantus 331 timidus totus : £w£ws : Turpinus : fwrfra : timere 295-6 Turpinus : htizts 335-6 fatmere : Turpinus : timor : tutus turma : timere 264-6 212 382 91 359 337-8 350-1 427 328 45-6 420-1 213-5 280 urgere : wr&s venia : vox 125-6 458 vestis : vicis 98-9 AND IN THE CARMEN 79 vexare : vallum 213 videre : victor : vindex : vir 315-7 videre : vir : vallum 221-2 vincere : velle 59-60 vincere : virtus 237 vir : viduare : vix 358-9 vir : vincere 304 vir : revocare : victoria 297 vivere : videre 455-6 vix : volvere 446-7 votum : videre 52-3 vox : venia : venialis 413-4 vulnus : vir 384-5 vulnus : vir : velut : videre : vigor 401-4 vulnus : vir : advenire 428-9 80 ALLITERATION IN THE ROLAND BIBLIOGRAPHY I. Texts : T. Atkinson Jenkins, The Song of Roland (Oxford Version), D. C. Heath & Co., forthcoming. G. Paris, Le Carmen de prodicione Guenonis, Rom. XI, 1882, p. 465. F. J. Mone, Lateinische Hymnen des Mittelalters, Freiburg i. Br., 1853-5. H. A. Daniel, Thesaurus hymnologicus, Lipsiae, 1855-6. P. Gall Morel, Lateinische Hymnen des Mittelalters, Einsiedeln, 1868. Joseph Kehrein, Lateinische Sequenzen des Mittelalters, Mainz, 1873. Sidonius Apollinaris, Carmina, Mon. Germ. Hist. Auct. antiqu, VIII. Sancti Aldhelmi Opera quae extant, ed. J. A. Giles, Oxford, 1844. Walafried Strabo, Vita S. Mammae Monachi, Migne, Patrol, lat., t. CXIV, 1051. Milo of St. Amand, Vita S. Amandi, Acta Sanct. Bolland, ed. Renschen, Februarii Tomus I, p. 884. Heiric of Auxerre, Vita S. Germani, Acta Sanct. Bolland., Julii Tomus VII, p. 238. Arnulfus, Delicie cleri, Romanische Forschungen, II, 243. Roger of Bee, De contemptu mundi, Migne, Patrol, lat., t. CLVIII, c. 690. II. Critical Works: A. On Latin Alliteration: A. F. Naeke, De allitteratione sermonis latini, Rheinisches Museum, 1829, 324. J. Maehly, Ueber Alliteration, Neues schiveizerisches Museum, 1864, 207. E. Loch, De usu allitterationis apud poetas Latinos, Halle, 1865. Karl Bartsch, Die lateinischen Sequenzen des Mittelalters, Rostock, 1868. R. Klotz, Alliteration und Symmetric bei Plautus, Zittau, 1876. J. Huemer, Ueber die altesten lateinisch-christlichen Rhythmen, Wien, 1879. H. Jordan, Kritische Beitrage zur Geschichte der lateinischen Sprache, Berlin, 1879. C. Ziwsa, Die eurythmische Technik des Catullus, Wien, 1879. A. Ebert, Allgemeine Geschichte der Liter atur des Mittelalters im Abend- lande, Band I 2 , Leipzig, 1889, II, 1880, III, 1887. E. Wblfflin, Ueber die alliterierenden Verbindungen der lateinischen Sprache, Sitzungsberichte der konigl. bayer. Akademie der Wissen- schaften, philosophisch-philologische Classe, 1881, II, 1. E. Wolfflin, Die Gemination im Lateinischen, ibid. 1882, 422. W. Ebrard, Die Alliteration in der lateinischen Sprache, Bayreuth, 1882. L. Buchold, De paramoeoseos (adlitterationis) apud veteres Romanorum poetas usu, Leipzig, 1883. AND IN THE CARMEN 81 C. Boetticher, De alliterationis apud Romanos vi et usu, Berlin, 1884. J. Bintz, Beitrage zum gebrauche der alliteration bei den romischen pro- saikern, Philologus, XXXXIIII, 1885, 262. E. Wolfflin, Zur Alliteration und zum Reine, Archiv f. lat. Lexikogra- phie u. Graminatik, III, 1886, 443. M. Kawczynski, Essai comparatif sur I'origine et Vhistoire des rythmes, Paris, 1889. E. Wolfflin, Eine Alliteration Caesars, Archiv f. lat. Lexikographie u. Grammatik, VII, 1892, 568. G. Grbber, Uebersicht iiber die lateinische Literatur, Grobers Grundriss, II, 1, Strassburg, 1902, p. 97. F. Ranninger, Ueber die Alliteration bei den Gallolateinern, Gymnas. Progr., Landau, 1895. cf. Zeitschr. f. franz. Spr. u. Lit. XVIII 2 , 1896, 20 \Zeitsch. f. rom. Phil. XX, 1896, 538; Rom. XXVI, 1897, 143. E. Wolfflin, Zur Alliteration, Archiv f. lat. Lexikographie IX, 1896, 567. E. Wolfflin, Zur Alliteration, Melanges Boissier, Paris, 1903, 461. Wilhelm Meyer, Gesammelte Abhandlungen zur mittellateinischen Rhyth- mik, Berlin, 1905. G. M. Dreves, Hymnologische Studien zu Venentius Fortunatus und Rabanus Maurus, Miinchen, 1908. B. On Romance Alliteration: A. Fuchs, Die romanischen Sprachen in ihrem Verhaltnis zum Latei- nischen, Halle, 1849. Goldbeck, Beitrage zur Altfranzosischen Lexikographie, Progr. d. Lu- isenschule, Berlin, 1872. C. Michaelis, Studien zur Romanischen Wortschopfung, Leipzig, 1875, 25, Anm. 1. Becq de Fouquieres, Traite general de versification franqaise, Paris, 1879. Paul Meyer, De I' alliteration en roman de France, Rom. XI, 1882, 572. G. Grober, Altfranzbsische Alliterationen, Zeitschr. f. rom. Phil. VI, 1882, 467. E. J. Groth, Vergleich zivischen der Rhetorik im altfranzosischen Ro- landslied und in Karls Pilgerfahrt, Herrigs Archiv, LXIX, 1883, 391. Pio Rajna, Le origini dell' epopea francese, Firenze, 1884. K. Nyrop, Storia dell'epopea francese, transl. E. Gorra, Torino, 1888. W. Riese, Alliterierender Gleichklang in der franzbsischen Sprache alter und neuer Zeit, Halle, 1888; cf. A. Mussafia, Lit. bl. f. germ. u. rom. Phil., X, 1889, 171. M. Kbhler, Ueber alliterierende Verbindungen in der altfranzosischen Literatur, Zeitschr. f. franz. Spr. u. Lit. XII 1 , 1890, 90. F. Kriete, Die Alliteration in der italienischen Sprache, Halle, 1893. E. Stengel, Romanische Verslehre, Grobers Grundiss, II, 1, Strassburg, 1902, p. 1. 82 ALLITERATION IN THE ROLAND R. L. Taylor, Alliteration in Italian, New Haven, 1900; cf. Giornale storico della let. ital. XXXIX, 366. F. Koehler, Die Alliteration bei Ronsard, Erlangen u. Leipzig, 1901. C.-M. Savarit, Les lois de I'allitteration et de Vassonance, semi-allitte- ration et semi-assonance, Mercure de France, Fevr. 1908, 430. M. Scholz, Die Alliteration in der altprovenzalischen Lyrik, Zeitschr. f. rom. Phil. XXXVII, 1913, 385, XXXVIII, 1914-5, 76, 193, 311. C. On Teutonic Alliteration: K. Weinhold, Altnordisches Leben, Berlin, 1856. Ludwig Uhland, Schriften zur Geschichte der Dichtung und Sage, Band I, Stuttgart, 1865. H. Oertel, Hildebrand's Theory on Alliteration, Mod. Lang. Notes, VII, 1892, 287. E. B. Setzler, Anglo-Saxon Versification, Baltimore, 1904. E. Sievers, Altgermanische Metrik, Pauls Grundiss, II, 2, Strassburg, 1905. W. Kalbow, Die germanischen Personennamen des altfranzosischen Hel- denepos, Halle, 1913. D. Critical Works on the Chanson de Roland and the Carmen de prodicione Guenonis in general: L. Jordan, Roland und Olivier, Romanische Forschungen, XVI, 1904, 637. Wilhelm Tavernier, Beitrage zur Rolands f or schung, Zeitschr. f. franz. Spr- u. Lit. XXXVI 1 , 1910, 71; XXXVII 1 , 1911, 83; XXXVIII 1 , 1911-2, 117; XXXIX 1 , 1912, 133; XLI 1 , 1913, 49; XLII 2 , 1914, 41. Wilhelm Tavernier, Zeitschr. f. rom. Phil. XXXVIII, 1914-5, 99, 226, 412. ■>■■ DATE DUE -J , AF 4 4jTj — - — j = — | — — _ . L 1 — 1 GAVLORD r pi I1NTEO IN U S A 3 1970 00544 7245 UC SOUTHERN REGIONAL LIBRARY FACILITY A A 000 200 043 8