Robert Francis Cook

Translators and Traducers: Some English Versions of the Song of , Stanzas 83-851

IKE ALL SCHOLARS AND TEACHERS whose fields are aspects of medieval culture, those of us who work with the history and literature of Lmedieval , or who teach and study the literary masterpieces of that civilization as elements of comparative literature, are obliged to remember constantly that our work is affected, in greater or lesser degree, by the nature of the translations available to us, to our colleagues, and to our students. There is nothing surprising or profound about this notion as a general matter, but it should not, perhaps, go without saying in every instance. It is especially when we see incontrovertible evidence of a failure larger than the slip of the pen, of an error of translation which is not strictly lexical, that we begin to recall how complex an endeavor it can be to interpret a lost language and a lost civilization. The technical problems of translating are not the only ones involved. Our training as specialists is designed to make us keenly aware of morphological and stylistic traps, of the uncertainty of lexical boundaries in medieval language (when is a glaive in fact a , despite our expec- tations to the contrary?) and of rhetorical patterns—periphrastic cors, doublets like bele et gent (or is that a doublet?)—especially troubling to an audience familiar with Pound or William Carlos Williams. But the somewhat different problem I want to treat here is a problem peculiar to specialists and the experienced, and it is one not often raised in connection with the translation of medieval literature. It is a matter not so much of an inability to read as of over-reading; it is the difficulty raised by the fact that the translator inhabits a particular academic universe, and hence a particular hermeneutic world, and it can have far-reaching conse- quences, as a consideration of , stanzas 83-85, will show. The examples I will give show nothing new about the art of trans- lation, therefore, except as they represent not a misunderstanding of the brute linguistic fact nor even a misunderstanding of some particular

1An abbreviated version of this paper was read at the 1978 meeting of the South- eastern Medieval Association in Lexington. Kentucky.

327 328 / Vol. 7, No. 4 / Summer 1980 cliché of the author's culture, but rather a sort of induced mental block, a beclouding of the act of translation by the action of an outside force whose influence, here at least, can hardly be said to have served the transmission of the known Oxford text to non-specialists. The lion in the path for all students of Roland in English2 is surely the Penguin Roland by Dorothy Sayers.3 Through it, countless students and some professors have had their first contact with the famous text, so often cited as a paradigm of the knightly condition and knightly ideals in the twelfth century (if not, indeed, in the eighth). For some time, this paperback was the only inexpensive Roland in English, and although Roland translations are now more numerous and more diverse than formerly, it remains one of the most widely available, with all the prestige and marketing power of the series, and twenty years' habits of citation, behind it. Its most obvious defect is its forced archaism, and that slippery coating may have made it too easy for us to swallow the interpretive prejudice it expresses. More of the latter in a moment. In order to avoid later confusion, I would like first to digress briefly, and say a word or two about the effect of archaism itself. Deliberate archaism is almost too easy for us to deal with in these times. The error of historical perspective that it represents is very simple; all it does is put ancient and unfamiliar terms into the mouths of charac- ters whose discourse in fact contains few such terms or none. It is misleading because its effect is to give the reader, not the "feel" of the text itself, but the "feel" of what it is like to be an antiquarian of limited experience, to whom the original text is attractive because it is still a bit queer. Poetic diction is another matter, not of necessity compounded of unusual terms and contorted syntax. The artificial introduction of those things into a medieval poem implies—wrongly for most purposes—that the patina of age is one of the work's more important features. To the

2French translations pose problems different enough in some respects to be beyond the scope of this discussion. Chief among these is surely the false cognate; thus Moignet, for example, sometimes translates the technical feudal terms aimer, amur,feid, onur, etc., by transposition, using their Modern French cognates, and sometimes he translates them lexically; cf. his vv. 45, 86, 315, and others. Thus soldeiers is 'mercenaries' in 34 as well as in 133. 3The Song of Roland (Harmondsworth: Penguin Books, 1957). Cook / Translations of the Roland 329 reader of the translation, of course, the archaism becomes an intrinsic feature of the work, for the reader has no immediate way of circumscribing it or separating it from the other features. Such trafficking, then, masks the poet's art by making it appear, especially to those students who themselves have limited historical perspective, that the Song of Roland must have been, from the moment of its conception, verbally strange, possibly even laughable. We should not forget that this sort of tampering can assume consid- erable importance in the surface texture of a translation. A few examples will suffice to remind us of what it looks like, though it may be years since we have read through one of the translations here quoted. Both Charles Scott-Moncrieff and Frederick Bliss Luquiens indulged in atmospherics before Sayers. Scott-Moncrieff's stanzas 74 and 55 were enough to give any budding medievalist pause (the full effect is obtained by reading them aloud): From the other part, Turgis of Turtelose, He was a count, that city was his own; Christians he would them massacre, every one. Before Marsile among the rest is gone, Says to the King: "Let no dismay be shewn! Mahum's more worth than Saint Peter of ; Serve we him well, then fame in field we'll own. To Rencesvals, to meet Rollanz I'll go, From death he'll find his warranty in none. See here my sword, that is both good and long [:] With I'll lay it well across; Ye'll hear betimes to which the prize is gone. Franks shall be slain, whom we descend upon, Charles the old will suffer grief and wrong, No more on earth his crown will he put on. (The Song of Roland [New York: Dutton, 1920], p. 31.) or this:

Charles the Great that land of Spain had wasted, Her castles ta'en, her cities violated. Then said the King, his war was now abated. Toward Douce France that Empereur has hasted. Upon a lance Rollant his ensign raisèd, High on a cliff against the sky 'twas placèd; The Franks in camp all through that country baited. 330 Olifant / Vol. 7, No. 4 / Summer 1980

Cantered pagans, through those wide valleys racèd, Hauberks they wore, their sarks were doubly plated. to their sides were girt, their helms were lacèd, Lances made sharp, escutcheons newly painted: There in the mists beyond the peaks remainèd, The day of doom four hundred thousand waited. God! What a grief. Franks know not what is fated. (p. 24) But while such verse may bore the professor of literature and bemuse the undergraduate, it does not attain the pinnacle of awkwardness reached by such lines as these, from Sayers's Penguin translation, stanzas 18 and 154:

"Barons, my lords, whom shall we send of you To Saragossa, the Sarsen king unto?" "Myself," quoth Roland, "may well this errand do." "That you shall not," Count let loose; "You're high of heart and stubborn of your mood, You'd land yourself, I warrant, in some feud. By the King's leave this errand I will do." The King replies: "Be silent there, you two! Nor you nor he shall on that road set foot. By this my beard that's silver to the view, He that names any of the Twelve Peers shall rue!" The Franks say nothing; they stand abashed and mute. (p. 61) or this: The County Roland is mighty of his mood, Walter de Hum well-famed for knightlihood, And the Archbishop a warrior tried and proved; Betwixt their valours there's not a pin to choose. In the thick press they smite the Moorish crew. A thousand Paynims dismount to fight on foot, And forty thousand horsemen they have, to boot, Yet 'gainst these three, my troth! they fear to move. They hurl against them their lances from aloof, Javelins, jereeds, darts, shafts and spears they loose. In the first shock brave Walter meets his doom. Turpin of Rheims has his shield split in two, His helm is broken, his head has ta'en a wound, His hauberk's pierced, the mail-rings burst and strewn, By four sharp spears his breast is stricken through, Killed under him his horse rolls neck and croup; Th'Archbishop's down, woe worth the bitter dule. (p. 131) Cook / Translations of the Roland 331

I must confess my inability to understand how the author of Lord Peter Wimsey's tripping speeches, a poet in her own right, with an Honors degree from Oxford in Medieval French, can have perpetrated such crimes not only against the Song of Roland but against English poetry. The habits induced by exposure to a tradition of scholarship and of trans- lation seem the most likely explanation. Let us grant Savers an eleven- syllable line, found virtually nowhere in English outside the work of translators hoping against hope to reproduce the very different effect of the epic caesura. Let us recall—and this caveat applies to the trans- lations to be discussed further on—that there are not all that many rhyme words in English (mood : feud may be comically misread) and that a rhymed or assonanced translation of an assonanced epic is in any case the worse part of valor, or démesure. This does not explain facile inversions, or pleonasms such as "silver to the view," or recondite contractions such as Sarsen, or solecisms such as the false metaphor "let loose" for 'say' or the intransitive use of the verb rue. And all of this pales alongside the pervasive and misleading archaisms, the paynims, the quoths, the County Roland, begone or shift me for the Eng. 3p. sing, 'go', and dozens more. As more recent translators (Terry, Robertson, Merwin, Harrison, Owen) have apparently recognized, such ornaments do not serve the original text thus putatively mediated. But while Dorothy Savers may have been alone among recent trans- lators in the striking extent to which she invests the Song of Roland with the rhetoric of her grandfather's time (or of Malory's), she is not entirely alone in the most important and least obvious of her infidelities, and it is to those that I would like to turn now. The problem of hypertranslation—of presenting as part of the text meanings and events that are not there, but whose presence we expect—is every bit as real a cultural difficulty for the translator as is the tempta- tion of Tennysonian diction, and its causes are partly analogous. Let us remind ourselves that we approach every act of mediation—not just translation but the teaching of those texts we cannot often read in detail— with a preconceived notion of what the text most probably contains.4

4For general discussions of interpretation as a logical procedure subject to the ana- lytical instruments applied to most other kinds of reasoning, see E. D. Hirsch, Validity in Interpretation (New Haven: Yale University Press, 1967), esp. pp. 164-207, and The Aims of Interpretation (Chicago: University of Chicago Press, 1976). 332 Olifant / Vol 7, No. 4 / Summer 1980 Let me suggest as a corollary that our prejudices become more precise and firmer the greater our training and experience, the more times we have viewed the text through a filter initially imposed upon it by what we once heard in the classroom or read in a textbook ourselves. This basic problem of hermeneutics becomes particularly acute in the case of the Song of Roland, for here we are dealing with a poem which, for over one hundred years, has had a fixed, traditional interpretation in academic practice, one which has only lately come to be tested by scholarly dialogue and debate. The ideal translation would not, perhaps, be a product of our readings, but raw material for them, indistinguishable from the untranslated text in matters of vocabulary, tone, order, inclusion, exclusion and so forth. In practice, translations are a form of applied hermeneutics. The English translations of the Song of Roland furnish some particularly clear examples of that process of application, and those examples are not simply matters of style but involve the trans- formation of the poem's substance. It may well be that, as much recent work (familiar to the readers of Olifant) has suggested, the standard interpretation of the Song is a scholarly convention, one of those notions that are much better examined than repeated. Obviously I cannot debate that question in detail here, but on the other hand I can point out the sort of thing that happens when the text is mediated by someone to whom that question has never occurred, and that may be enough to give us pause whatever our convictions in the matter. The stakes are, professionally speaking, fairly high. Outside the small world of the active specialist in Old French epic, the Song of Roland is almost uniformly assumed to be what Gaston Paris and Léon Gautier said it is: the drama of a tragic defeat, caused by Roland's blind and exag- gerated confidence in himself, and by nothing else. It is one thing to insist, with the original text in hand, that such a reading is adequate, satisfactory or necessary; it is quite another to produce (under influences similar to those evident in the discussion of archaism just given) a translation which can be read in that way and in no other. Yet the power of scholarly tradition is great. The same training that makes it possible for us to avoid simple technical errors of the type first discussed may, under certain circumstances, favor a more tenuous but equally real strain of misconstruction. It should come as no surprise that Cook / Translations of the Roland 333 more than one modern English translator, sure of his or her ground, has injected elements of the traditional reading—learned, no doubt, at the time of first contact with the Song of Roland— into the translation, into places where the Song itself admits of no such clear and compelling inter- pretation. The most flagrant examples are surely Roland's promises. All of these deserve close attention, which I hope to give them elsewhere. I would like to confine myself, for the moment, to the most important ones, those in the "First Scene." Let us assume the common sort of reading in which the text guides the reader rather than the opposite. If, then, our hero is a consummate fool, he will likely expect to defeat the entire Saracen army with his twenty-thousand-man rear guard; and if he is supremely blind to the situation, then his words will reveal his blindness and condemn him. That notion is in fact implicit in classic Roland criticism and occurs in much the same form, for example, in Pierre Le Gentil's standard study of the poem: "Roland aime la guerre pour la guerre. Il savoure l'âpre joie des combats. Il rêve de conquêtes. Le sentiment de l'honneur dicte toutes ses paroles et inspire tous ses actes" (La Chanson de Roland [Paris: Hatier, 1955], p. 103). Such ideas may proceed, nonetheless, from our certainties and not directly from the Old French text, as an examination of parallel stanzas 83-85 and of their translations will show. I give first the Oxford version in Gérard Moignet's 1969 edition; the italics are mine.

LXXXIII Dist Oliver: "Paien unt grant esforz; De noz Franceis m'i semblet aveir mult poi! 1050 Cumpaign Rollant, kar sunez vostre corn, Si l'orrat Carles, si returnerat l'ost." Respunt Rollant: "Jo fereie que fols! En dulce France en perdreie mun los. Sempres ferrai de Durendal granz colps; 1055 Sanglant en ert li branz entresqu'a 1'or. Felun paien mar i vindrent as porz: Jo vos plevis, tuz sunt jugez a mort.'" AOI. LXXXIV "Cumpainz Rotlant, l'olifant car sunez. Si l'orrat Carles, feral l'ost returner, 1060 Succurrat nos li reis od sun barnet." 334 Olifant / Vol. 7. No. 4 / Summer 1980

Respont Rollant: "Ne placet Damnedeu Que mi parent pur mei seient blasmet Ne France dulce ja cheet en viltet! Einz i ferrai de Durendal asez, 1065 Ma bone espee que ai ceint al costet; Tut en verrez le brant ensanglentet. Felun paien mar i sunt asemblez: Jo vos plevis, tuz sunt a mort livrez." LXXXV "Cumpainz Rollant, sunez vostre olifant, 1070 Si l'orrat Carles, ki est as porz passant. Je vos plevis, ja returnerunt Franc. — Ne placet Deu, ço li respunt Rollant, Que ço seit dit de nul hume vivant Ne pur paien, que ja seie cornant! 1075 Ja n'en avrunt reproece mi parent. Quant jo serai en la bataille grant E jo ferrai e mil colps e .VII. cenz, De Durendal verrez l'acer sanglent. Franceis sunt bon, si ferrunt vassalment; 1080 Ja cil d'Espaigne n'avrunt de mort guarant."

Our attention shall be given principally, for the time being, to lines 1055-58, 1065-69, and 1078-81. Each time, Roland says, of course, pretty much the same two things. If we attempt to stick to the letter of the text, we may see that he promises to strike many good blows, and adds, in 1080, that his men will do the same. And then he adds, first, that the pagans are making a mistake, and then, second, with the unnoted transi- tion characteristic of paratactic constructions, that they are all jugez or livrez a mort, inevitably. Certainly the most striking thing about these three strophes, in the present context, is the nature of the promises Roland makes. In point of fact, though he personally assures Oliver that the pagans are making a serious mistake and will die, he never explicitly says how, nor when, nor where the pagans will meet their doom. He swears to his friend that the pagans are marked and promised to death, and if we continue to follow the letter of the poem, what Roland says is not an error nor an exaggera- tion but the truth. The pagans do die, to the last man, at the hands of . Many scholars will wish to add that the pagans are all going to die eschatologically as well. The Christian soldiers are holy Cook / Translations of the Roland 335 martyrs (vv. 1135-37, 1520-23) and can expect paradise as their reward for what they are about to do. On the other hand, the pagans' very souls are doomed.5 We do not habitually think of Roland as taking the larger view of things; but our habit, here at least, must impose itself upon a text which is at most ambiguous. In the Oxford Roland, the hero twice describes the Saracens' fate in the passive voice but never uses an active subject; thus, categorically, he avoids any explicit statement that he himself, or indeed any living man, will be the agent of their destruction. Nor does he ever use any localizing or temporally limiting adverb. (Ja in 1081 is about as general as an adverb can get.) Ultimately, our reading of what Roland states here, and our inferences about what he expects, will have to account for his references to his own possible death, in lines 1091: "Melz voeill murir que huntage me venget," and 1122-23: "Se jo i moerc, dire poet ki 1'avrat . . . que ele fut a noble vassal," as well as Turpin's subsequent speech to the assembled rear guard: "Pur nostre rei devum nus ben murir," v. 1128, "Se vos murez, esterez seinz martirs," v. 1134, and also by the detachment's own acquiescence in the idea of possible death; the latter is the first event in the First Horn Scene, v. 1048:"Ja pur murir ne vus en faldrat uns." These are not the words of men who are blind to approaching danger. They may be only clichés (except vv. 1122-23). Nonetheless, they do not necessarily lose all significance for that reason: there are other formulas which might have been used here in their stead, if the purpose were to express unguarded optimism.6 The principal text, however, is in the three great stanzas that contain Roland's responses to Oliver's appeals; and there, the first-person pronouns ring out in introductory clauses, jo vos plevis 'I give you my word'; the subordinates, the promises themselves, contain no first-person references of any kind. We would be wrong rhetorically and structurally to neglect the juxtaposition here with the Twelve Pagan Peers, whose

5Line 1268 gives the only clear example I have found in the Roland of the motif which shows a freshly killed pagan being carried off to hell; the topos was certainly known to the author. 6It is perhaps significant, still, that the reply of the rear guard is made not to Roland or to Turpin, but to Oliver, and that Oliver alone speaks at least obliquely of victory: "El camp estez que ne seium vencuz," v. 1046. 336 Olifant / Vol. 7, No. 4 / Summer 1980 specific promises of personal victory all contrast with Roland's passives. "Jo l'ocirai," says Marsile's nephew (v. 867); "I1 est juget [passive] que nus les ocirum [active]," says Falsarun, (v. 884). "Se trois Rollant, de mort li duins fiance"; "Se trois Rollant, n'en porterat la teste"; "En Rencesvals irai Rollant ocire" (vv. 914, 935, 963), so day the enemy. The most ironic and revealing promise of them all is that of Chernuble of Munigre, who claims: "Si cunquerrai Durendal od la meie" (v. 988); of course Chernuble is the unfortunate man who is cut in two, horse and all, by Roland's most famous and most prodigious blow. Whatever these pagans' claims, they cannot in fact kill Roland; he kills them. In such a context, only the imposed scholarly cliché allows us to claim we are sure Roland is acting the fool. Given the tradition of the gab, it may even appear that his atten- uated predictions show some restraint. Yet those English whose influence has so far been the greatest lend Roland's promises a precision and a limitation which make them seem foolhardy in the extreme. Let us start with the Sayers version (again the italics are mine): 83 Quoth Oliver: "Huge are the Paynim hordes, And of our French the numbers seem but small. Companion Roland, I pray you sound your horn, That Charles may hear and fetch back all his force." Roland replies: "Madman were I and more, And in fair France my fame would suffer scorn. I'll smite great strokes with Durendal my sword, I'll dye it red high as the hilt with gore. This pass the Paynims reached on a luckless morn; I swear to you death is their doom therefor." 84 "Companion Roland, your Olifant now sound! King Charles will hear and turn his armies round; He'll succour us with all his kingly power." Roland replies: "May never God allow That I should cast dishonour on my house Or on fair France bring any ill renown! Rather will I with Durendal strike out, With this good sword, here on my baldrick bound; From point to hilt you'll see the blood run down. Woe worth the Paynims that e'er they made this rout! I pledge my faith, we'll smite them dead on ground." Cook / Translations of the Roland 337

85 "Companion Roland, your Olifant now blow; Charles in the passes will hear it as he goes, Trust me, the French will all return right so." "Now God forbid," Roland makes answer wroth, "That living man should say he saw me go Blowing of horns for any Paynim foe! Ne'er shall my kindred be put to such reproach. When I shall stand in this great clash of hosts I'll strike a thousand and then sev'n hundred strokes, Blood-red the steel of Durendal shall flow. Stout are the French, they will do battle bold, These men of Spain shall die and have no hope." The changes are obvious. Lines 1057-8 now state, not that the pagans are marked for death, but specifically that they will die as a result of being in that place at that time; and line 1069 says, not that the pagans will be death's victims, but that Roland himself, and his men, will kill them. Finally (and Moignet's punctuation is itself influenced by this notion), line 1081 is taken to be the explanation and extension of 1080 rather than what it may very well be instead: an exact, and syntactically independent parallel to 1058 and 1069, stating only that the enemy is doomed. This is a matter of grammar and not necessarily an ambiguity contrived by the author. It would be of great help if we had the connectives that are naturally absent from these paratactic constructions. But in their absence, we may not insist on a univocal interpretation of these lines. Patricia Terry, in her 1965 translation for the Library of Liberal Arts, makes relatively few errors, and her work is free of egregious archaisms. However, this straightforwardness masks, as previously suggested, the same tendency to forced readings that we have just seen in Sayers's version. Terry's Roland answers Oliver's initial entreaty thus:

Roland replies: "You must think I've gone mad! In all sweet France I'd forfeit my good name! No! I will strike great blows with Durendal, Crimson the blade up to the hilt of gold. To those foul pagans I promise bitter woe— They are all doomed to die at Roncevaux." (The Song of Roland [Indianapolis and New York: Bobbs-Merrill, 1965], p. 43; I continue to italicize.) 338 Olifant / Vol. 7, No. 4 / Summer 1980 We may fidget when we hear los, which is not 'reputation' but 'earned reputation', or 'the community's reactions to our deeds' translated by 'good name', which could be either; and we might long for a verb in v. 1056; but it is the localizing adverb phrase that will surely bring us up short at this point. We are not especially relieved to hear Terry's Roland add, in lines 1068-69: "The Saracens will curse the evil day / They challenged us, for we will make them pay." I repeat that the translator's concern for rhyme and meter does not require this sort of thing. Our language has a fair number of monosyllabic verb-phrase elements, and in any case, the subject is not at the rhyme. Can it be doubted that the fashionable critical notion of démesure has led the translator, perhaps unawares, to settle upon the pronoun we as one of the acceptable monosyllables here? Our doubt will disappear if we consult the third stanza of the series under consideration. "We have good men; their prowess will prevail," says Roland, instead of his habitual result-neutral phraseology, as in the original "Franceis sunt bon, si ferrunt vassalment" (1080); and in Terry's version, he adds: "And not one Spaniard shall live to tell the tale." Under the pressure of the widespread idea that Roland is guilty of the sin of démesure, and secure in the knowledge that great critical concepts are expected to have been derived from the text, Dorothy Sayers and Patricia Terry have both plunged over the brink, and given, not a rendering of the Oxford text, but a gloss. It follows that their texts are not primary sources (as their titles lead us to believe) but secondary works, vehicles of a particular scholarly attitude. Which of us has not done the same thing, in the presentation of the long and foreign works that are ours to live with, by virtue of our academic standing as specialists in another culture? If démesure or courtly love or the bourgeois mentality exist, then they must be in the texts, and anywhere the texts are at least locally ambiguous, then we may tend to assume them. And where is démesure going to show up, if not in the only sustained statement Roland ever makes about anything? In fact, it should now be clear that the works I have discussed stand at the end of a tradition that has to some extent fed upon itself. More than one of the older English translations is in reality a version of the standard hypothesis rather than of the text as it stands on the page. There, as in Sayers or Terry, it is not a question of nuance or poetic licence, but of Cook / Translations of the Roland 339 substantial additions. Thus Luquiens: "Fore God! The Paynims shall not thrive in Ronceval! I pledge you, comrades, they shall die this day!" (The Song of Roland [New York: MacMillan, 1952], p. 38.) Or Scott- Moncrieff: "Felon pagans to th' pass shall not come down"; and "I pledge you now, to death they're doomed today."7 The worthy John O'Hagan, M.A. ("One of the Justices of the Supreme Court in Ireland") was aware enough of the convention to have Roland declaim, under the ubiquitous rubric "Roland's Pride," these lines: "The heathen felons shall find their fate; Their death, I swear, in the pass they wait."8 Even Joseph J. Duggan, in a careful analysis of the vantance motif as a formulaic phenomenon, was moved to translate the two passives, "tuz sunt jugez a mort" "tuz sunt a mort livrez" and their active transformation (unmarked as to agent) "Ja cil d'Espaigne n'avrunt de mort guarant" with first person singular subjects as supplied agents: "I will bring death to the pagans." The shift in question obviously has the status of a reflex. Duggan does point out the absence of a direct response (Terry's "No!") on Roland's part.9 It is immediately obvious that the translations analyzed often represent deformations of the texts. That may not be surprising, though it is certainly cautionary, and we may legitimately deplore the fact that the Song of Roland is widely available in inaccurate English versions; after all, it is taught in English almost exclusively on this continent, in depart- ments of English, History, and Comparative Literature as well as in French courses for students not majoring in French, and it seems further clear that medievalists quite conversant with Old Norse or Anglo-Saxon may not necessarily know all the other medieval languages equally well. But the problem would be much smaller than it is, if accuracy of detail were all that is involved. What we have before us is, rather, a series of misreadings of a particularly tenuous and tricky sort The interpre- tation of the Song of Roland is, once again, in dispute. Many reputable scholars do not think the Oxford text is about démesure at all. In that

7Pp. 35-36. Mar i vindrent as porz, the Oxford reading for line 1057, is identical to that given by Petit de Julleville, whose edition Scott-Moncrieff used. 8The Song of Roland (Boston: Lothrop, Lee, Shepard, n.d.), p. 94.

9The Song of Roland: Formulaic Style and Poetic Craft (Berkeley: University of California Press, 1973), p. 24. 340 Olifant / Vol 7, No. 4 / Summer 1980

context, the changes made by translators appear in a new light. They express the prejudices of academics, and they show a lack of sensitivity for the verbe of the book, as though nothing it says could counteract the general impression we have always had of the text. (This is the antithesis, notably, of one of the attitudes we work to instill in our students.) The assumptions of secondary authors are thus made explicit in many passages of the hypertranslated text. We need only compare Sayers's lines 778-79, to the King: "The rear-guard now has been adjudged to him, And you've no baron can ever make him quit," with the original: "La rereguarde est jugee sur lui; N'avez barun ki jamais la remut." Only the first, with the extra personal pronoun, and not the original, mirrors the critics' insistence upon Roland's impetuous rush into disaster. The poem merely says that what is done cannot, for whatever reason, be undone by any of the knights. It is not too much to suggest that Sayers, Terry, and others have traduced a complex bit of text, the First Horn Scene, either by removing its characteristic syntactical ambiguity, or perhaps by making it signify something like the opposite of what it contains. This is not, of course, the only example of our failure to notice the precise features of a text we have all known a great deal about for a long time. Thus we all know that Rollant est proze Oliver est sage, but until the appearance of L.S. Crist's article in this journal (Olifant, 1[1974], 19) I think it had never been ex- plicitly stated that the same epithet is applied several times to Oliver(175-76, 546, 559, 576, 3186, 3755), and I am not sure that anyone has even yet pointed out that the treacherous is first introduced, in line 26, as a prozdom, simple though such an observation may be.10 In such a climate, Harold March could write that Roland's "impetuousity and his pride make enemies for him,"11 when we suspect that behind such a statement is our old topos of early enmity between and Roland; in fact, the text shows 100,000 of Roland's comrades-in-arms all concerned or afraid for his life (vv. 842-43).

10The term is also applied, among others, to Gangleus (v. 3509), Pinabet (v. 3915), Grandonie (prozdoem, v. 1636), and such Christian heroes as Ogier, Turpin, Acelin, and Gerer; see Joseph Duggan, A Concordance of the Chanson de Roland (Columbus: Ohio State University Press, 1969), p. 302. 11In the Introduction to the Terry translation, p. xxii. Cook / Translations of the Roland 341

In the case of a translation or a résumé, however, there is a structural fact not present in the case of secondary criticism. The function of the translation is precisely to substitute itself for the text, in aid of those who do not read the latter.12 The reader of the translation is in the translator's hands, much as the student is at first in the professor's. The fallacy of hermeneutic necessity just described is thus analogous in its unhappy effect to the fallacious archaism discussed at the outset. To the extent it can be avoided, it is avoided not simply by our knowing the language or even the culture, but by our vigilance with respect to things that remain to be proven. In this respect, the newer translations by Robert Harrison, Howard Robertson, and D.D.R. Owen, and the somewhat older but rather obscure one by W.S. Merwin, all less handy, more expensive, or less enthusiastically promoted than the Sayers Penguin, are superior translations, because (the implications of their punctuation aside) all three reproduce Roland's words very much as they stand, in Owen's case despite the strong opinions of the translator. 13 Let us hope they take their

12I will not attempt here to set this phenomenon formally into a framework of the type furnished by modern linguistics. Despite the current trend of extending semiotics to cover just about anything "readable," we may prefer to consider this invasion of the text by the readings, with its diachronic and cross-cultural aspects, to be a metalinguistic event.

13Thus Harrison: "I promise you, they all are marked for death" (vv. 1058, 1069, The Song of Roland [New York: New American Library, 1970], p. 84); Robertson:"I promise you, all are condemned to die/as good as dead," (The Song of Roland [London: Dent, 1972], pp. 31-32); Owen: "I pledge you this: they are all marked for death" (The Song of Roland [London: George Allen and Unwin, 1972], p. 53, cf. p. 10); and Merwin: "I promise you, they are all marked out for death/destruction;" but cf. vv. 1080-81: "The French are brave, they will fight hard and well, and those who have come from Spain will not be saved from death" (The Song of Roland, in Medieval Epics [New York: Modern Library, 1963], pp. 125-26). 342 Olifant / Vol. 7, No. 4 / Summer 1980 rightful place in our syllabus and our bibliographies, for it is not practical to deal simultaneously with a vital issue of historical criticism and with the side-issue of a failed translation; indeed, to a reader uninformed of the original text, the issue of démesure may be purely and simply invisible.14 Robert Francis Cook The University of Virginia

14I was fortunate to have been able to consult the indispensable introduction and commentary given by Gerard J. Brault, The Song of Roland: An Analytical Edition, 2 vols. (University Park, Pa., and London: Pennsylvania State University Press, 1978) just before submitting this paper. These points should be noted. In Moignet's line 1553 (Brault's 1510) there is a second example of the pagan warrior whose soul is taken directly to Hell; and Marsile goes to Hell upon dying (3647 in both eds.). Brault lays the question of the word proz to rest for good in a long discussion, I, pp. 181-83 (cf. pp. 12-13, and the valuable Notes, 412-14): "Proz could actually refer to a broad range of knightly virtues including most significantly, the treasured ability to give sound advice" (v. I, p. 187). Finally, Brault links lines 1057-58, 1068-69, 1080-81 by commas, in both text and trans- lation. I have also just received Frederick Goldin's translation, The Song of Roland (New York: Norton, 1978). Goldin's reading follows the recent trend toward accurate represen- tation of the parataxis in vv. 1058 ("I promise you, they are marked men, they'll die") and 1069 ("I promise you, they are all bound for death") but links the localizing adverb of 1068 to the notion of death by anticipating 1069, as the text does not: "These pagan traitors have gathered here to die."