- 'Men speke of romances of prys,
Of Horn child and of Ypotys,
Of Bevis and Sir Gy.'
Since the time of Chaucer's 'Rime of Sir Thopas', and earlier, the romantic heroes Sir Bevis of Hampton and Sir Guy of Warwick have been familiarly associated in English literature. It is not surprising, then, that the lives of the two should be found side by side in an Irish manuscript, and it is not inappropriate that they should appear together in the first printed edition of the Irish texts.
The only1 existing copy of these texts, so far as I know, is that preserved in MS. H. 2. 7 in Trinity College Library, a vellum folio in various hands, probably of the fifteenth century.2 A few passages from both romances were printed by Nettlau in the Revue Celtique 10, 187-191. The language, which was long ago characterised by O'Donovan as 'pure and of great value to
p.10the Irish scholar',3 can doubtless be dated with some definiteness when the verbal forms are fully tabulated and compared with those in other late Middle Irish texts. The Stair Fortibrais, a translation in similar style of which a copy exists in the same manuscript, is vaguely assigned by Dr. Stokes, its editor, to the fourteenth or fifteenth centuries.4 The translation of John Mandeville, on the other hand, is distinctly stated in the manuscript to have been made by Fingin O'Mahoney in 1475,5 and a comparison of the grammatical forms of all these pieces with it and with the translation of Marco Polo6 ought to help in establishing a more precise date for them, and perhaps also to shed some light upon the question of their authorship. But the investigation of these matters cannot be satisfactorily completed while the greater part of the foreign romantic material in Irish, to which Nettlau called attention in his articles in the tenth volume of the Revue Celtique, still remains unpublished.
The exact sources of both the 'Guy' and the 'Bevis' are unknown, though there is good ground for believing that they go back to English originals, as was assumed long ago by O'Donovan7 and O'Curry.8 The principal evidence for this opinion is to be found in the proper names. Zimmer,9 arguing from those in Nettlau's extracts, pointed this out, and an examination of the complete list practically places the matter beyond doubt. To be sure, many of the names are indecisive and might go back equally well to French or to English.
Others are so distorted like Aimistir Amundae from Amis de la Mountagne that it is difficult to draw conclusions from them. But a number of forms remain which it is easiest to explain by assuming an English intermediary between the Irish and the French. Thus Heront (Eront), from French Heraut (Heralt) is very likely to have got its n as a result of the errors of English scribes. Compare the way in which Rohand or Roband was made out of the French Rohaut (Rohalt) in some English versions of the story.10 The Irish Uront shows the same development in the last syllable and apparently corresponds to Yorauld, a name which I have found in Copland's version alone. (The other English versions have Torold, and the Wolfenbüttel French text Corraud.) Pani (for French Pauie) and Gincadh (for French Guichard) both show the same transformation of u into n, and in these instances Copland's 'Guy' has forms with n (Pani and Gincharde). The Irish form Sision probably rests upon an English modification of Sessoigne.11 Finally the constant use of 'Sir' in titles (Sir Gyi, Sir Heront) is plainly modelled on the English, and there are several instances where the English word 'kin'g ('Cing O Niubie', 'Cing Herrneis',' Cing Caulog') has been taken over intact into the Irish text. All these indications, the last of them practically decisive, point to an English source for the 'Guy'. In the Bevis fragment, which is much shorter, the evidence is not so clear. There is very little difference between the French and the English forms of the names, but where these disagree the Irish stands in every case nearer to the English unless it departs from both alike. The Irish name 'Babilon', too, for the country of Ybor's brother, may be due to the English 'Dabilent' (itself a corruption of French 'd'Abilen't).12 So far as it goes, then, the testimony of the names in the 'Bevis' is consistent with that of the 'Guy'.
I have not attempted to draw any conclusion from the presence in both texts of a considerable number of loan-words, apparently from English. I have no doubt that words of English origin are more numerous because the author was working with an English romance. But it is obvious that they prove nothing decisively, for the Irish writer need not have taken them from his source. In fact all, or nearly all, of them occur in other texts. Sometimes, moreover, it is not easy to decide whether a word is of English or French origin. A critical study of the foreign elements in the Middle Irish vocabulary, ascertaining the sources of loan-words and the date of their introduction into the language, yet remains to be made.
An analysis of the contents of the Irish 'Guy' and 'Bevis' might be expected to lead much farther toward the determination of the sources. But it does little more than confirm the results already derived from the study of the proper names. Both romances differ in so many features from all the other versions I have seen that I must assume their immediate sources to be unknown. A brief statement, however, of their relations to their respective cycles is of interest, particularly in the case of the 'Guy'.
I have been unable to compare in detail the Irish 'Guy' with the French versions of the story, since none of these has been published except in summaries or extracts.13 But it is clear that none of the French texts of which I have suceeded in finding a description stands in any close relation to the Irish, and I have already shown it to be probable that the source of the latter was English. Of the English versions the most important are easily accessible. Zupitza has published metrical texts of the fourteenth and fifteenth centuries,14 and I have
p.13been able to compare with them the rare edition by William Copland of which the Harvard University Library possesses a copy. I have also examined Rowlands's poem15 and several later chap-book versions.16 The Middle English metrical versions, which agree with each other essentially in plot, stand closest to the Irish and I have made them the basis of my comparison. A brief statement with regard to the proper names and the principal incidents will show the relation they bear to the Irish.
More than two-thirds of the Irish names are either the natural equivalents of the English, or can be explained without difficulty as transformations of them.17 There are seven substitutions,18 and six names of new persons and places19 occur without any equivalent in the English. These additions and substitutions are hardly to be regarded as the invention of the Irish author, but probably stood in his English source. Some of them are of special interest. 'Richard' in the place of 'Rohaut', the name of Guy's father, may have chronological significance, as I shall point out below. 'Cing Caulog',20 who appears once in
p.14the place of King Athelston, is probably King Havelok, the Danish leader (better known as Anlaf Cuaran), whose name became somehow confused with that of his English opponent. John de Alcino belongs in an episode which will be discussed a little later.
With respect to the narrative itself the Irish translation shows considerable independence. It contains every episode of importance in the English and has several additional incidents besides. Such are the fight between Guy and the duke of Lombardy (Chapter 4); the three days' tourney in Brittany (Chapter 5); and the tournament in Normandy (Chapter 7). In all these cases the English has nothing to correspond except general statements that Guy fought in Normandy, Brittany, France and Spain. In Chapter 8 the Irish relates a fight in the market-place at Bruidis, instead of which the English and French versions seem to have a tournament at Benevento. In Chapter 29 the Irish gives an account of a fight with a Turk, not paralleled in English. And in Chapter 34 there is a long discourse on Christian doctrine, not found in the English, concerning which I shall speak more particularly below. These chapters, I should add, are lacking not only in the Middle English romances but also in every other version of the story I have been able to consult.
With the few exceptions mentioned, six chapters out of forty-three, the general plot of the Irish romance agrees, incident for incident, with the Middle English. But there is hardly a paragraph in which there are not differences of detail. In chapter 1, for example, the account of Felice's skill in embroidery is peculiar to the Irish. The description of Siccard's rule is much fuller in the English. Nothing is said in the Irish of Guy's early training by Heront; and much is made of his piety and of the religious ceremonies at his knighting, both unmentioned in the English. In Guy's interviews with Felice
p.15the English, which relates them much more fully, suggests that his proposals were improper, whereas the Irish makes no mention of 'folye'. And the conditions in Chapter 1 are by no means peculiar. I have noted similar variations in thirty-nine out of forty-five chapters. Sometimes they concern unimportant details; sometimes the plot in the Irish is manifestly improved; occasionally the Irish redaction confuses the story; and in a few cases appears to adjust it to the conventions of native tales. It is impossible to say how many of these modifications are deliberate changes on the part of the translator, but when all due allowance is made for his independence I think that many of the variations in plot as well as in the proper names must be attributed to his source.
Guy of Warwick was a mediaeval hero of the type of St. Alexis, and a principal feature of his story in all its forms is the desertion of his bride. All the versions, therefore, make a plea for religion and asceticism. But the Irish, as compared with the English, is particularly insistent on works of piety and charity. This has already been pointed out for Chapter 1. Again in Chapter 39 the pious deeds of Felice are described in Irish, but not in the corresponding portion of the English. In Chapters 19 and 35 the Irish makes special mention of prayers of which the English says nothing. But the most conspicuous addition of a religious nature is Chapter 34, which is otherwise of special interest. When Guy is overcome by remorse for his sins and decides to abandon Felice, the Irish romance alone represents him as seeking spiritual counsel and obtaining instruction in Christian doctrine. He sends for a holy father, John de Alcino, to whom he confesses his sins and by whom he is exhorted to keep the commandments, to avoid the eight21 mortal sins, to emulate the sufferings of the saints, and to believe in all the articles of the Apostles' Creed. The name of the confessor, John de Alcino, furnishes a clue to the source of this theological chapter. It is a condensation of part of the material
p.16found in the Middle English 'Speculum Gy de Warewyke',22 which rests in turn upon the 'Liber de Virtutibus et Vitiis'23 of Alcuin. This moral treatise was originally written by Alcuin for a different Guy, Count Guido of Tours, a celebrated military leader under Charlemagne. But as early as the beginning of the thirteenth century it had become attached in England to Guy of Warwick, who is named as the recipient of the advice in the Auchinleck MS., the earliest copy of the 'Speculum'. On the other hand, in the romance of Guy contained in the same manuscript there is no reference to Alcuin or to the sermon, and I have not found the episode in any version except the Irish. The Irish redactor either made the combination himself, or had before him a romance into which the substance of the 'Speculum' had been woven. The latter of these suppositions appears to me the more probable. There is nothing else in the Irish text to indicate that the author compiled his work from different sources, and the combination in question would have been more naturally made by an Englishman than by a foreigner.
A number of lost versions may intervene between the Irish 'Guy' and the known Middle English texts, and the 'Speculum' may have been several times abridged in the course of transmission. Or the source of the Irish chapter may have been derived in some other way from the 'Liber' of Alcuin. As it stands, it is much shorter than the 'Speculum' and does not agree with that closely in the arrangement of material. But the three principal elements in the Irish are to be found in the English poem. For the list of deadly sins see the 'Speculum' ll. 107 ff.; for a description of the sufferings of the saints, ll. 176 ff.; and for an exposition of portions of the Creed, ll. 200 ff.
Thus the Irish life of Guy makes probable the existence of an English romance which differed in one important feature, and may have departed in many details, from the known English versions of the story. As to the date of the assumed English original, a lower limit can perhaps be established by the grammatical analysis of the Irish text. Beyond this the Irish supplies another bit of possible evidence. The name of Guy's
p.17father-in-law, as I have already pointed out, is changed from Rohalt (Rohaut, Rohand) to Risderd. The reasons for the substitution are entirely unknown, but it might have arisen from confusion with the name of a living Richard, Earl of Warwick, or from a deliberate purpose of complimenting him. There were two Earls of Warwick of that name in the fifteenth century.24 Richard de Beauchamp, who was born in 1382, was Earl from 1410 till his death in 1439, and Richard Neville (the kingmaker), born in 1428, obtained the title by marriage in 1449, and died in 1471. As between the two, I think the general probabilities of date are in favor of the earlier. Moreover Beauchamp, we are told,25 travelling in the Holy Land in 1410, was feasted and given presents by the Lieutenant of the Soldan because of his supposed descent from Guy. In 1422 he endowed the chantry at Guy's Cliff. In view of his active interest in the romantic tradition of the house of Warwick it is quite conceivable that his Christian name may have got into some contemporary version of the story.
The Irish Bevis is only a fragment, though a rather long one. The comparison of its contents with other versions of the story is made easy by Köbing's edition26 of the Middle English texts and Stimming's edition27 of the Anglo-French. Both editors discuss the relations of the French, the English, the Welsh and the Norse red actions.28 Besides these mediaeval versions I have also examined an English chap-book Bevis, probably' of the year 1680.29
When compared with the French and English romances the Irish Bevis shows less new material than the Guy. It
p.18contains no incident of importance not to be found in both the Middle English and the Anglo-French. But in the matter of minor variations it stands in about the same relation to them that the Guy bears to the texts with which I have compared it. Out of 22 names30 of persons and places, 17 are the natural equivalents of those in the Middle English, 4 are explicable31 as modifications of the English, and only two (that of Para, the son of the Emperor, and that of Biroig,32 a stream on the borders of Scotland) are new. A comparison of the narratives shows constant variation in details. According to the Irish, Bevis's mother is in love with the son of the Emperor; and according to the Middle English and the French, with the Emperor himself. (In the chap-book of 1680 it is the Emperor's brother.) In the Irish account, her determination to marry her lover is awakened by seeing her own beauty in a bath. No such situation is mentioned in the French or the English. In both the French and the English the little Bevis is set to tend sheep, not swine; and there is no conversation parallel to that by which in the Irish version he is impelled to avenge his father's murder. In chapter 8 the Irish represents Bevis as journeying to India and Rhodes, while the Middle English takes him to Jerusalem, and the French to Jerusalem and Egypt. (The chap-book has no eastern travels at this point.) The episode of Sisian and Yvor in chapter 9 is introduced considerably earlier in the English and the French (and in the chap-book as well). The dragon-fight in chapter 11 contains some vivid details about four waves of vomit which are very likely the Irish redactor's own invention. From most of these features of the Irish narrative I am led to conclude that it had its source in a lost version. That this was probably English I infer from the
p.19proper names, as already pointed out, and also from the fact that where the French and English versions differ with regard to the details of the story, the Irish, if it does not depart from both usually resembles the English.33 In a few cases where the Irish agrees with the French as against the Middle English metrical versions the English prose version of 1680 is like the Irish.
The result, then, of this comparison of both the Guy and the Bevis with the corresponding stories in other languages is to make it probable that the Irish lives are free redactions of lost English versions. The assumed original of the Bevis appears not to have differed in any important particulars from the other existing forms of the story. In the case of the Guy, on the other hand, the Irish text points to the existence in English of a combination, hitherto unknown, of the romantic material proper with the religious material, originally distinct, of the 'Speculum Gy de Warewyke'.
Stylistically regarded, the Irish texts are clearly very free renderings of their originals. Though the number of foreign words in them may be somewhat larger because of their foreign sources, the manner of the narrative is thoroughly Irish, and they read in general like the native stories in the somewhat ornate prose of the period. The accumulation of adjectives and adverbs, often in alliterating groups of three, is characteristic of late Middle Irish, and the Guy and Bevis are by no means extreme examples of the practice. In this matter, and in the general structure of sentences, I have adhered in my translation very closely to the original, though the traditions of English prose are so different from those of Irish that the
p.20resulting style will sound sometimes monotonous, and sometimes redundant and artificial.
It is now nine years since I first copied and collated these texts at Dublin. During the interval I have profited several times by the courtesy and liberality of the Librarian and staff of Trinity College, and I now desire to express my grateful acknowledgements. I am also under much obligation to both the editors of the Zeitschrift for reading my proofs and giving me the benefit of their counsel. Wherever it is possible, particular acknowledgment will be made of their suggestions and corrections.
Since only one manuscript of these romances is known to me, I have simply tried to print its readings as accurately as possible. Obvious errors or omissions are occasionally corrected in the text or in foot-notes in order that the narrative may be readable. A certain amount of normalization is also involved in the punctuation and the separation of words and the expansion of contractions. But I have made no attempt to correct the grammar or orthography of the scribe. His errors and in consistencies, for example, in initial mutations and in the general treatment of spirants have all been allowed to stand.
In the form in which my text was sent to press all expanded contractions were indicated by italics, so that the reading of the manuscript could be instantly ascertained in every case from the printed page. But out of regard for the strong preference of Professor Stern I have abandoned that plan and used italics only in cases which are in some respect doubtful or exceptional. The typographical appearance of the text is much improved by the change, and I think there has been no loss in accuracy. The work of the editor, however, has become less easy to control, and it is important for me to make an exact statement of the method I have pursued and the liberties I have allowed myself. Short specimens of the text with all the
p.21abbreviations indicated are furnished by the passages which Nettlau printed in the Revue Celtique 10, 187 ff.
I have silently expanded all the ordinary compendia scribendi unless their use appeared to be a given case irregular. The scribe freely employed the signs for acht (cht, sed), air, ar cet, con, cu, er, (ir), est, et (ed), eth (edh), m, n, nem, or, ra, re, ri, ro, ru, uath, ur, us; and certain extensions of their use are also so common in the manuscript that have adopted them without resorting to italics. Thus the sign for ur clearly means sometimes 'r' (as in 'anoir', 'senoir') and often 'uir' (as in 'docuir' 3 sg. pret.), though in a few cases the latter combination is indicated by an i with the sign for ur above it. The sign for us also sometimes stands for 'uis'. I have inserted the i in cases where its omission would be grammatically misleading (as in 'romarbuis', 2 sg. pret., or 'fochtuis', 3 sg. pret. absol.), but I have allowed spellings like 'eglus, fiadhnuse', to stand, since the scribe does not consistently observe the principle of caol le caol when he spells out words in full. In the same way I have some times expanded the sign for er as 'eir' (cf. dobeir, 3 sg. pres., of frequent occurrence), but I have left forms like derc, serc (dat. and acc.) without trying to introduce uniform indication of the i-infection. The abbreviation for eth occurs a number of times in the ending of the preterite passive where I have expanded it as edh (docuiredh).
Besides silently expanding the abbreviations which stand for definite letters, I have also made no use of italics in supplying obvious vowels before b, c, d, g, written above the line; and in cases where there could be no doubt about the construction I have added the endings of nouns and adjectives in -ach, -ech (-aigh, -igh), of preterites in -aigh, -igh, and of preterite passives and verbal nouns. All these are frequently indicated by a simple dash. In the case of verbal nouns in -dh and of preterite passives two abbreviations are usual with the scribe, a dash (rofer-), and a d above the line ('rofoerd'). For the former cases I have used the spirant 'dh', and for the latter the unaspirated 'd'. Both forms occur in words which the scribe has spelled out in full, and the distinction between them was of no importance.
In addition to the contractions thus far provided for, there are a considerable number of words habitually abbreviated by
p.22the scribe in accordance with the practice of Middle Irish manuscripts. Those which occur oftenest, and about which there can be no real doubt, I have expanded without italics, using the grammatical form required by the context. A list of them is given here. In all other words italics are used unless the manuscript abbreviations represent definite letters or the syllables provided for above.
In the matter of accents I have endeavored to follow the manuscript disregarding those, however, which obviously do not mean quantity but serve only to distinguish the letter i. Probably some of the scribe's accents have been overlooked because of their faintness, but I have not intentionally inserted any of my own. I ought to explain that Professor Stern would have preferred the consistent marking of all long vowels, but I did not wish to go quite so far in the normalization of the text. I am therefore alone responsible for the method adopted. In some other respects, too, my text follows the manuscript rather than the usual practice of modern Irish writers. The preterital prefixes do and ro, for example, I have regularly combined with their verbs, and certain enclitics which are commonly written separately I have set off by hyphens. These are not matters of importance. I cannot claim theoretic consistency in my use of hyphens, but I hope none of them will prove misleading. My general purpose has been to adhere closely to the manuscript, and at the same time to make the printed text easily intelligible.
There are of course endless opportunities for error in reproducing a text of such irregular orthography, and I regret that I cannot compare the proofs with the original. But in order to make the mistakes as few as possible I had the manuscript photographed after copying and collating it.
In the Glossary I have meant to register only such words as are not fully accounted for in Windisch's Wörterbuch. Both there and in the foot-notes references by number and letter (306a, 315b, etc.) are to the pages and columns of the manuscript, which are indicated in the Irish text. Some of the footnotes which accompany the translation will be found to contain comment of a textual nature. I expected at first to have the Irish and English printed on opposite pages, but that method proved to be too wasteful of space.