Studies in Arthurian Illustration, Vol. I 1904597378, 9781904597377

Citation preview

Studies in Arthurian Illustration Vol. I

Studies in Arthurian Illustration Vol. I

Alison Stones

The Pindar Press London 2018

Published by T eh P i nda r P r e s s 30 W e nt w ro t h D r i ve L ond H A 5 2P U · U K

B r itis hL A c a t a l ogue

i br a r y C

r e c or d f or

a t a l ogui

t hi s bok i

ng i

nP

ubl

i c a t i on D

s a va i l a bl e f r om

t he

I S B N 978 1 904597 37 7

Printed by S hor t R un P E xe t e r , D e vo T hi s bok i

nE

r e s s L t d, X 2 7L W

s pr i nt e d on a

c i d- f r e e pa pe r

a ta B r itis hL

ib ra ry

Contents Foreword

i

A

Approaches to French Secular Illustration

I

Secular Manuscript Illumination in France

II

Sacred and Profane Art: Secular and Liturgical Book-Illumination in the Thirteenth Century

29

III

Arthurian Art Since Loomis

57

IV

Indications écrites et modèles picturaux, guides aux peintres de manuscrits enluminés aux environs de 1300

93

1

V

Fabrication et illustration des manuscrits arthuriens

121

VI

Text and Image in Arthurian Manuscripts

155

B

The Lancelot-Grail Romance: Methods and Subjects

VII

The ‘Lancelot-Graal’ Project

165

VIII

Teaching and Research on Medieval Art on the Web: Three Sites

180

IX

Towards a Comparative Approach to Manuscript Study on the Web: the Case of the Lancelot-Grail Romance

192

X

Stories in Pictures and their Transmission: A Comparative Approach to the Manuscripts of the Lancelot-Grail Romance

201

‘Mise en page’ in the French Lancelot-Grail: the First One Hundred and Fifty Years of the Illustrative Tradition

221

XII

Short Note on Manuscripts Rylands French 1 and Douce 215

239

XIII

Another Short Note on Rylands French 1

247

XIV

Les débuts de l’héraldique dans l’illustration des romans arthuriens

267

XV

Lancelot and Identity

289

XVI

A Note on the Heraldry of a Very Special Gauvain

316

XVII

Le merveilleux dans le Lancelot-Graal: l’exemple du cerf accompagné de quatre lions

328

C

Estoire and Queste Illustration

XVIII

The Earliest Illustrated Prose Lancelot Manuscript?

343

XIX

Two French Manuscripts: WLC/LM/6 and WLC/LM/7

370

Un schéma d’emplacement pour l’illustration de l’Estoire del saint Graal et les débuts de la tradition manuscrite

403

XXI

Seeing the Grail. Prolegomena to a Study of Grail Imagery in Arthurian Manuscripts

432

XXII

The Grail in Rylands MS French 1 and its Sister Manuscripts

485

XI

XX

XXIII

Signs and symbols in the Estoire del saint Graal and the Queste del saint Graal

524

XXIV

The Illustrations of the Queste del saint Graal in Yale 229 and Other Queste Manuscripts

556

XXV

The Illustrations of BN, fr 95 and Yale 229. Prolegomena to a Comparative Analysis

607

XXVI

L’Estoire del saint Graal dans la version adaptée par Guillaume de la Pierre pour Jean-Louis de Savoie, évêque de Genève : sources et traitement pictural

678

Foreword

T

hese essays span close to 50 years of work in the area of French secular illustration and, in particular, of the Lancelot-Grail romance, the romances of Chrétien de Troyes, Tristan, and Wace. I was first drawn to medieval literature through my undergraduate studies in French and German, followed by graduate work in medieval art and illuminated manuscripts in particular. The two disciplines combined seamlessly into research initially on the Lancelot-Grail romance, which has remained a Leitmotif throughout my career, taking an important place alongside work on other texts and pictures both sacred and secular in French and Latin. I was challenged by the enormity of the illustrative tradition of the LancelotGrail: close to 200 manuscripts, most of them illustrated, the number of pictures ranging from a single illustration to the astounding total of 748, found in the Additional 10292–4 copy housed in the British Library, made in Flanders c. 1317 and dated on the famous Tombs of Judgement commissioned by Flegentine in the Estoire del saint Graal. This is still a lot of material to deal with: how does one handle such a huge iconographical tradition ? The essays that follow demonstrate various approaches to the problem and show that there is still a great deal to do to resolve all the issues this vast corpus of illustration raises: questions of geographical and chronological distribution, of relations to other works commissioned and made by and for the same patrons and makers, of the options and choices about what to illustrate and what pictorial emphasis to present in each illustrated copy, and of comparisons in the layout and presentation of the page in illustrated and in unillustrated copies. Mapping these disparate dimensions in tandem is just beginning to be possible in an approach based on Geographic Information Systems (GIS) which allows a comparative three-dimensional view of page and book as opposed to traditional linear methodology and goes beyond the synchronic (within one manuscript) and diachronic (across clusters of manuscripts) approaches to manuscript studies.

ii

I welcome this chance to update what I have written over these past years and to correct numerous errors, though no doubt many remain. I have combined the references into a single bibliography and added pointers where appropriate to images now available on the web, while realizing that URLs are often volatile. Finally I express my gratitude to the many friends, mentors, collaborators, and keepers of manuscripts who have generously encouraged my research over the years and whose help is gratefully acknowleged in the articles that follow and on the Lancelot-Grail website (http://www.lancelotproject.pitt.edu). This work has benefitted from grants, leaves and fellowships from the Universities of Minnesota and Pittsburgh; All Souls and Magdalen Colleges Oxford; Corpus Christi College Cambridge; the École pratique des Hautes Études Université de Paris-IV Sorbonne; the Institut national d’Histoire de l’art, Paris; the National Endowment for the Humanities and the Andrew W. Mellon Foundation, the American Council of Learned Societies, the American Philosophical Society, the Fulbright Foundation. I also thank Liam Gallagher for his enthusiasm and for bringing this set of essays to fruition. I dedicate these essays to my grandchildren: Eva, Thomas, Arthur and Elodie, hoping they will one day enjoy these stories and pictures as much as I have. Alison Stones

A Approaches to French Secular Illustration

I Secular Manuscript Illumination in France

T

he flowering of French medieval literature may be said to begin around 1100 with the composition of the Chanson de Roland. This epic theme and those that emerge later in the century, whether classical, like the trilogy Thèbes, Troie and Enéas, or Arthurian like the Tristan of Thomas and Béroul or the works of Chrétien written in the third quarter of the century, constitute a core of material whose popularity continues throughout the Middle Ages, not only in France but in the rest of western Europe. By the thirteenth century the prose versions of these earlier verse texts continue an established tradition, enriched with new themes both in verse and prose through the fourteenth and fifteenth centuries. There is evidence from the twelfth century and before, that literary themes provided inspiration for artists.1 The illumination of French literary manuscripts, however, does not begin until shortly before the middle of the thirteenth century, with the emergence of the lay craftsman. It flourishes between the thirteenth and sixteenth centuries, when illumination gives place on the one hand to panel painting, murals and tapestries, and on the First published in Medieval Manuscripts and Textual Criticism, ed. C. Kleinhenz (North Carolina Studies in the Romance Languages and Literatures, Symposia, 4), Chapel Hill, NC, 1976, pp. 83–102. 1 In Western Europe one of the earliest examples is the Franks Casket (London, BM 1867,0120.1), early 8th century, which contains some scenes from Norse mythology; one of the earliest Arthurian examples is the lintel of Modena cathedral north door, which contains scenes based on a text that is now lost. See Loomis, Arthurian Legends, figs. 4–8; see now G. Allaire, “Arthurian Art in Italy,” in ead. and F.R. Psaki, eds. The Arthur of the Italians: the Arthurian Legend in Medieval Italian Literature and Culture, Cardiff, 2014, ch. 13. The lintel is to be dated in the early twelfth century. Themes from Roland also appear in the twelfth century in sculpture, at Cremona cathedral, for example; see Lejeune and Stiennon, Roland, and Ross, Roland.

2

other to the printed book with its mass-produced illustrations. The purpose of this essay is to present an outline of the main developments that take place during that period in terms of texts and patrons, in the layout and style of the miniatures and their relation to liturgical illumination, and the influence of French secular illustration on production elsewhere in Europe. Literary fashion and secular illumination The sequence of secular illumination in France, as elsewhere, is governed to a large extent by literary fashions. The MSS that survive are only a fraction of what the total output must originally have been, but they nevertheless give some idea of the relative popularity of the texts they contain. Some of these texts were popular only for a limited period of time, falling out of favour as new compositions took their place; others seem to have withstood changes in literary taste, remaining in demand from the time of their composition to the end of our period and beyond. Among the latter are the Roland and the Tristan mentioned above, although in both cases it is the later variants on the earliest versions that were more widely read and illustrated, just as the prose Lancelot, composed in the early thirteenth century, has survived in far more MSS than have the texts of Chrétien. The prose Tristan and the prose Lancelot were two of the most popular texts of all, if one may judge on the basis of the numbers of MSS still extant.2 They also contain two of the longest cycles of miniatures; the complete Lancelot cycle, including Estoire, Merlin, Lancelot, Queste, and Mort Artu, was often contained in a single volume of over 400 pages of large folio, with a cycle of over 200 miniatures, and the Tristan cycle is almost as lengthy.3 The Alexander romance, its various French versions deriving from the classical texts,4 and the Roman de la Rose by Guillaume de Lorris and Jean de Meung,5 were also most prolifically reproduced. The MSS of prose texts are listed in Woledge, Bibliographie. For example, the Bonn Lancelot, Bonn, Landes-und-Universitätsbibliothek, MS 526, which contains all five branches of the cycle, is 477 pages long and contains over 230 miniatures, while the Vienna Tristan, ÖNB 2542 has 500 ff. and 198 miniatures and historiated initials. 4 See Ross, Alexander Historiatus. 5 See E. Langlois, Les Manuscrits du Roman de la Rose, description et classement, Paris/ Lille, 1910; A. Kuhn, Die Illustration des Rosenromans, Freiburg/Breisgau, 1911; J. V. Fleming, The Roman de la Rose, A Study in Allegory and Iconography, Princeton, 1969. The earliest illustrated MS of the Roman de la Rose I now think is perhaps Ambrosiana I 78 Sup, last quarter 13th c.; BnF fr 1559, is datable c. 1300, while Cologny-Genève, Fondation Bodmer 2 3

SECULAR MANUSCRIPT ILLUMINATION IN FRANCE

3

Little is known of the circumstances under which these works were composed, who were the original patrons of the poets, and indeed who the poets themselves were, for even when their names are known, there is little further information about their lives. Information is also lacking as to the original patrons of most of the extant MSS of these texts. Jean, Duke of Berry, owned a Lancelot, BnF fr 117–20, and a Roman de la Rose, BnF fr 380; both were made for him between about 1380 and 1400; they contain his name and appear in his inventories.6 One may cite earlier cases of documentary evidence for the ownership of MSS, like the inventory of Jean d’Avesnes, count of Hainaut (d. 1304), at the head of which appears “uns grans roumans a rouges couvertures ki parolle de Naschen, de Mellin et de Lancelot du Lach.” Robert de Béthune, count of Flanders (d. 1322) also owned a Lancelot, which appears in his inventory, while among the books confiscated by Robert of Artois from Mahaut in 1316 there are three Tristan MSS. There is also evidence that ecclesiastics owned secular books; the will of Guillaume d’Avesnes, bishop of Cambrai (d. 1296) and brother of Jean d’Avesnes, count of Hainaut, requests that his “livre de gestes” be given to the monks of the monastery of Saint-Sépulchre, Cambrai, since one of them had had it made for him.7 The wording of the will is not precise enough for one to tell whether the book had actually been made at the monastery, but it would seem unlikely as the available information indicates that at this date both scribes and illuminators were predominantly laymen.8 There is no proof that any of the MSS of Guillaume or Jean d’Avesnes, Robert de Béthune, or Robert or Mahaut of Artois are among those extant 79 was written in 1308; see Stones, Gothic Manuscripts, Part II vol. 2, pp. 235–43, with reference to further literature. 6 M. Meiss, French Painting in the Time of Jean de Berry, vol. 1, The Patronage of the Duke, London/New York, 1967, pp. 312, 313. 7 These documents and many more are published by C. Dehaisnes, Documents et extraits divers concernant l’histoire de l’art dans la Flandre, l’Artois et le Hainaut avant le XVe siècle, Lille, 1886. For Géraud Mahaut of Artois see J. M. Richard, Une petite-nièce de Saint-Louis, Mahaut, comtesse d’Artois et de Bourgogne, Paris, 1887; C. Balouzat-Loubet, Mahaut d’Artois; and note 8a below. 8 There is more information on this for Paris than for the provinces. The main source of information is the tax rolls, which generally refer to the taxpayers by name and by profession. Scribes and illuminators are mentioned from the middle of the thirteenth century, see R. Branner, “Manuscript Makers in Mid-Thirteenth Century Paris,” Art Bulletin, 1966 and F. Avril, “A quand remontent les premiers ateliers d’enlumineurs laïcs à Paris ?” Les Dossiers de l’Archéologie, 1976, pp. 36–44. See also H. Géraud, Paris sous Philippe le Bel, d’après des documents originaux, et notamment d’après un MS contenant le rôle de la taille imposée sur les

4

today.8a The library of Jean de Berry on the other hand is not only exceptional in itself, but also later in date, and consequently better preserved and better documented. Many of the most splendid secular MSS contain no precise clues that identify their original owner: such are the Harley Rose BL Harl. 4425 (c. 1500) (fig. 1), the Bodleian Alexander, Oxford, Bodleian, Bod. 264 (1338–44), and the Pierpont Morgan Lancelot, New York, Pierpont Morgan Library M.805–6 (c. 1300–20) (fig. 2). Where there are clues, they are often of doubtful interpretation. Shields of arms sometimes appear in the margins, but these may be later additions,9 or purely decorative, disobeying fundamental rules of heraldry by employing metal on metal or colour on colour,10 or they may simply be unidentifiable.11 The Yale Lancelot, Yale University Library 229, ex-Phillipps Collection 130, is one of the few major MSS whose owner may perhaps be traced on the basis of the heraldry, A marginal knight appears on the opening page of the Queste section of the text (fig. 3), with a horse whose housing contains the arms or a lion sable, a bend gules. The representation is small in scale, and inconspicuously placed on an inner page, but the arms are identifiable from Wijnbergen12 as those of Guillaume de Termonde, second son of Gui de habitants de Paris en 1292, Paris, 1897; J.-A. Buchon, Le Livre de la taille de Paris en l’an mil trois cent treize, Paris, 1827; K. Michaëlsson, Le Livre de la taille de Paris, Acta Universitatis Gothoburgensis, LXIV, 1958, 4; LXVII, 1961, 3. See now Rouse and Rouse, Manuscripts and their Makers. There appear to be no documents for Cambrai itself owing to losses in WWII. 8a P. Stirnemann has proposed that BnF fr 12467, Enfances Ogier, was owned by Mahaut d’Artois (“Bibliothèques princières,” p. 190, n. 67). Since its illumination is attributable to the Méliacin Master, it is likely in my view that she inherited it from her father. See Stones, Gothic Manuscripts, Part I, vol. 1, p. 117. 9 BL Roy.20 D.IV, Lancelot, was owned by a member of the Bohun family, probably Humphrey of Bohun, 7th Earl of Essex in the late fourteenth century. The book was made in Flanders c. 1310–20. Humphrey added his coat of arms and had some of the miniatures, including the opening one, overpainted. See M. Rickert, Painting in Britain in the Middle Ages, Harmondsworth, 1966, for Humphrey’s library; and now L.F. Sandler, Illuminators and Patrons in Fourteenth-Century England: The Psalter and Hours of Humphrey de Bohun and the Manuscripts of the Bohun Family, London, 2014, and the BL web site. 10 BL Roy.14 E.III, Lancelot, has on its opening page some arms that are later additions and others that have colour on colour. Now reproduced on the BL website. 11 The basis for identifying shields used in this way is the contemporary rolls of arms, of which the most important for France are Wijnbergen, Berry, Navarre, and Gelre. The oldest is Wijnbergen, the date of which is between 1265 and 70. See P. Adam-Even and L. Jéquier, “L’Armorial Wijnbergen,” Archives Héraldiques Suisses, 1951–4. See now L’Armorial le Breton, ed. E. de Boos et al., Paris, 2004. 12 Wijnberghen, p. 72, no. 1235.

SECULAR MANUSCRIPT ILLUMINATION IN FRANCE

5

Dampierre, count of Flanders. Guillaume was born in 1248 or 49 and died in 1312; there is no precise date for the MS, but c. 1280–90 would seem reasonable on the basis of the style. The documentary evidence for the books owned by the counts of Flanders is scant, and there is no inventory of the possessions of Guillaume. His name does come to mind again in connection with books of the second main category, those that enjoyed limited popularity and exist in only very few copies. Most of these MSS are presentation copies, intended for the patron who commissioned not only the manufacture of the MS but also the composition of the text itself. Guillaume’s name is connected with MS BnF fr 15104, La noble chevalerie du Judas Machabée et de ses nobles frères (fig. 4). The text was composed in 1285 for “mon seigneur Guillaume, de Flandre.”13 There are no arms or other marks of ownership on the MS, and there is no absolute proof that BnF fr 15104 is in fact the original MS, but the fact that this is the only MS containing this version of the text would support the suggestion. It is true that the style of the illumination is rather different from that of the Yale Lancelot, but a date of 1285 seems convincing for it, and there is no reason to suppose that Guillaume might not have acquired MSS from different sources, illuminated in different styles, just as Mahaut of Artois did in the early years of the fourteenth century.14 The relationship between the commissioning of a work of literature and the presentation manuscript are rather more explicit in Parisian products of the years around 1300. This was an important moment for activity under the patronage of the French royal court. The works of the poets Girart d’Amiens and Adenet le Roi figure large among the texts illustrated at the court workshops. It was probably for Marguerite, daughter of Philip III and Marie de Brabant that Girart composed Méliacin. She is included in a royal portrait miniature on the opening page of the text of the earliest MS, BnF fr 1633, and she and the other figures may be identified on the basis of the heraldic garments they are wearing.15 The MS can be dated between 1285, since Philip IV is included, crowned, and 1291, date of the death of Jeanne de Châtillon, and it is almost certainly the presentation copy intended for the patroness, Marguerite. 13

65.

Edited by J. R. Smeets, Assen, 1955. The patron is mentioned in the text, lines 7944–

She purchased books locally at Hesdin, also in Arras, and from two booksellers in Paris. See Dehaisnes and Richard, op. cit. 15 Vitzthum, Pariser Miniaturmalerei, pl. III, pp. 24–32. 14

6

A similar royal gathering is to be found in the opening miniature of Ars. 3142, a compendium of poetry featuring largely the writings of Adenet le Roi. This time the poet himself, wearing a crown as his nickname suggests, is represented amidst the royal group that includes Marie of Brabant, second wife of Philip III (married 1274), Blanche of France, widow of Ferdinand de la Cerda (died 1275), and Mahaut of Artois. It has been suggested that the MS was dedicated to Mahaut,16 but there is no evidence for this other than her inclusion in the miniature; it could have been intended for Marie.16a There is no element of uncertainty about the recipient of BnF lat 8504. It is a Latin translation of Dina et Kalila, from the Spanish, made in 1313 for Philip IV. On the opening page is a royal group with the king seated in the center, and an inscription beneath the miniature names the patron and gives the date of the translation. This is undoubtedly the presentation copy given to the king17 (fig. 5). The Dina et Kalila MS stands isolated; the texts of Girart and Adenet, on the other hand did enjoy a certain popularity in the years around 1300. One of the latest MSS contains both texts: BR IV 319. Stylistically it is similar to BnF fr 22495, written in 1337, and probably produced in Paris. From earlier in the thirteenth century there are a few isolated MSS that most probably were presentation copies, though the identity of the patron is obscure. One outstanding example is the lavishly illuminated Roman de la Poire, BnF fr 2186 (fig. 6), made perhaps for the lady represented on f. 10v, who receives a book from a kneeling male figure. Both wear garments with the same heraldic device, which has so far eluded identification.18 This is a particularly sumptuous MS containing a cycle of full-page miniatures (fig. 9), which is very unusual in French secular illumination, and this also suggests that some important person was to be the recipient. It is also the only surviving illustrated copy of this text. A special case of isolation is met with in the books associated with René, duke of Anjou and king of Sicily. Two texts are involved, both composed by Ibid., p. 56, pl. X. I have suggested that the intended recipient may have been Marie’s young nephew, the future Jean II de Brabant, see Stones, Gothic Manuscripts, Part I, vol. 2, p. 76. 17 Ibid., p. 170 ff. See now N.F. Regalado, “Le Kalila et Dimna de Paris, BnF, MS Fonds latin 8504 (1313): Raymond de Béziers enseigne la fable orientale aux princes français,” in D’Orient en Occident: Les recueils de fables enchâssées avant les Mille et une Nuits de Galland, ed. M. Uhlig and J. Foehr-Janssens, CELAMA 16, Turnhout, 2014, pp. 283–308. 18 The arms are France charged with a cross or lozengy gules. 16

16a

SECULAR MANUSCRIPT ILLUMINATION IN FRANCE

7

René himself. One is the Livre du cœur d’amour épris, ÖNB 2595, and the other the Livre des tournois. Several copies of the latter are extant, all dating, as does the Cœur MS, from c. 1460–65. Two original MSS of the Tournois text are extant; BnF fr 2695–6, illuminated by the same hand as that of Cœur. Indeed it has been suggested that René himself was responsible for the execution of the miniatures as well as the composition of the texts, and this constitutes a unique case of the patron fulfilling the additional roles of both author and artist.19 Secular book production in relation to liturgical MSS Developments in both the layout and the style of secular illustrations are closely related to those of liturgical or devotional books, and in many cases there is evidence that both were produced by the same workshops and artists. While this is generally the rule, many of the outstanding and sumptuous literary MSS have no parallel in liturgical books, while the opposite is also true, and there are splendid liturgical books illuminated by artists who appear not to have participated in secular MS illustration. No secular MS is comparable, for instance, with the elaborate cycle of fullpage miniatures illustrating scenes from the Old Testament in the Pierpont Morgan Library, MS M.638, dating from the middle of the thirteenth century,20 although certain moduli from this are taken up again much later in secular illustration.21 Similarly there is no secular MS illuminated by the hand of Honoré as it appears in the frontispiece to the Breviary of Philippe le Bel, BnF lat 1023, f. 7v. or in the Somme le Roi BL Add 54180,22

See O. Pächt, René d’Anjou et les Van Eyck, Cahiers de l’Association internationale des études françaises, 1956; O. Smital and E. Winkler, Le Livre du Cuer d’Amours espris, Vienna, 1927; E. Pognon, Verve vol. IV, no. 16, 1946; see now Splendeur de l’enluminure, le roi René et ses livres, ed. M.-E. Gauthier with F. Avril, Angers, 2009. 20 Facsimile ed. J. Plummer, London, 1969; reproduced on corsair.themorgan.org. 21 For instance, the motif of the falling knight in BL Add. 10294, f. 81v [fig. 7], may be compared with f. 9 of the Picture-Bible. 22 Honoré’s name is in the Paris tax rolls of 1292, 1296, 1297, 1299, 1300, and the Breviary was paid for in 1296. The quality of the works attributed to him varies considerably and it is probable that he was head of a shop and worked with assistants. See E. G. Millar, The Parisian Miniaturist Honoré, London, 1959, and D.H. Turner, “The Development of Maître Honoré,” British Museum Quarterly, XXXIII, 1–2, pp. 53–64. There is now a huge literature on Honoré. For a summary see Stones, Gothic Manuscripts, Part I, vol. 2, Cat. no. I–43. 19

8

although the Girart and Adenet MSS referred to above were undoubtedly produced in the circle of Honoré’s influence. No secular MS illuminated by the hand of Pucelle has come to light,23 while the secular MSS made later in the fourteenth century for Jean de Berry and around 1400 are rarely the work of the well-known illustrators of his noted devotional books, such as Jacquemart, the Boucicaut Master, and the Limburg brothers.24 Cases of outstanding secular MSS without parallel in liturgical books are less frequent, but the Bodleian Alexander, MS 264,25 and the René MSS find no exact counterparts, and the same could be said of the Harley Rose, BL Harl. 4425 (fig. 1). Close connections between secular and liturgical MSS do exist from the beginnings of secular illumination in France, shortly before the middle of the thirteenth century. The earliest MS of the crusading epic, Histoire de la guerre sainte, by Guillaume de Tyr26 (fig. 8), is very similar stylistically to the Toledo MS of the Bible moralisée, whose frontispiece with its portrait of Louis IX is now in the Pierpont Morgan Library, MS M.240. It is likely that both were made for Louis before his departure to the Holy Land in 1248, and it is probable that the Guillaume de Tyr MS was made for Louis himself, although it contains no specific marks of ownership. The style and layout of the Roman de la Poire, BnF fr 2186 (fig. 9) may be closely paralleled in the fragments containing scenes from the Old Testament, Oxford, Bodleian Library, Douce 381 (fig. 10), and in the Psalter Douce 50. Later in the thirteenth century the Yale Lancelot (fig. 11), probably made for Guillaume de Termonde, may be related to liturgical books. This time the relationship is rather more complex than in the case of Poire because two hands are involved in the illumination. One may be related to the illustration See K. Morand, Jean Pucelle, Oxford, 1962. On Pucelle see most recently Jean Pucelle: Innovation and Collaboration in Manuscript Painting, ed. K. Puyn and A. Russakoff, Turnhout, 2013. 24 Meiss, op. cit. and vol. 2, The Boucicaut Master, London, 1968. Two of Jean de Berry’s Boccaccio MSS are attributed to the Boucicaut Master; his Rose, BnF fr 380 is attributed to the circle of Jacquemart, while his Lancelot BnF fr 117–20 is attributed to the Master of the Cleres Femmes. 25 Facsimile edited by M.R. James, Oxford, 1933. Now beautifully reproduced on the Bodleian Library web site. 26 A thorough study of the Guillaume de Tyr MSS has been made by J. Folda, The Illustrations of William of Tyre’s History of Outremer, thesis, Johns Hopkins University, Baltimore, 1968; id., Crusader Manuscript Illumination at Saint-Jean d’Acre, 1275–1291, Princeton, 1976. 23

SECULAR MANUSCRIPT ILLUMINATION IN FRANCE

9

in the Psalter made for Gui de Dampierre, father of Guillaume, BR 10607 27 (fig. 12), while the other (fig. 13) reappears in BnF fr 95,28 Estoire, Merlin (fig. 14–15), Sept Sages, Pénitence Adam, and a group of liturgical MSS that include BnF lat 1076 (fig. 16), Psalter, and Marseille BM 111, Book of Hours.29 Since the work of the two hands is interspersed throughout the Yale MS, it is likely that they are to be considered as two artists of different training employed in the same workshop. Where this workshop was located is uncertain; the only clue is provided by the liturgical use of the Book of Hours and by the calendar of the Psalter, both of which indicate the diocese of Thérouanne. This included parts of Flanders. There is perhaps a sharper division between liturgical and secular books in the fourteenth and fifteenth centuries than in the earlier period, and the examples that do show work by the same illuminators in both types of books are frequently among the less well-known products of the period. Jean de Berry’s Lancelot, BnF fr 117–20, while not the work of his most noted illuminators, is related to the Bible, Ars 5058,30 and the Lancelot BnF fr 9631 finds a close stylistic parallel with the Book of Hours, Chantilly lat. 13621.32 Similarly the name of Jean Colombe, who completed the Très Riches Heures left unfinished at the death of the brothers Limburg in 1416, has been linked with the illustrator of the Estoire BR 9246.33 As far as the layout of the illumination is concerned, the main difference between liturgical and secular books is in the use of the full-page miniature. This was employed far more frequently in liturgical than in secular books. It is the regular choice for the Crucifixion and Christ in Majesty miniatures This MS is very problematical, see L.M.J. Délaissé, Miniatures médiévales, Brussels, 1959, no. 9, p. 50, pl. l49v–50. The heraldry has been repainted. 28 Loomis, op. cit., p. 95, suggests that BnF fr 95 and the Yale Lancelot were intended as two of a three-volume set of the complete prose Lancelot cycle. (The first part of the Lancelot proper, now missing, would be the third volume.) Although the illumination is in part very close indeed, fr 95 also includes two non-Arthurian texts, Sept Sages and Pénitence Adam, while none of the other complete Lancelot compendia have any other texts. It is probable that this workshop made several Lancelots, some complete, others containing only part of the cycle. See Woledge, op. cit., for a list of which MSS contain which parts of the cycle. 29 See J. Billioud, “Très anciennes heures de Therouanne,” Trésors des bibliothèques de France, 1935, no. 5, pp. 165–85, pl. LXI–IV. 30 Loomis, op. cit., p. 106. 31 Ibid., p. 108. 32 J. Meurgey, Les principaux manuscrits à Chantilly, Paris, 1930, pp. 102–3, pl. LXVI. 33 Loomis, op. cit., p. 111. See Stones, “Jean-Louis de Savoie,” in this volume. 27

10

standard in Missals, and for the cycle of the Life and Passion of Christ that are frequent in the illustration of Psalters and Books of Hours (fig. 10). Cycles of full-page miniatures are rare in French secular illustration, and this is particularly interesting in relation to German secular manuscripts.34 There are some isolated examples in the thirteenth century in France, notably the Roman de la Poire, BnF fr 2186 and the Roman de Troie, BnF fr 1610.35 The former contains either a single scene or two scenes contained within roundels or quatrefoils, while the latter has scenes arranged in two or three registers. The Bodleian Alexander has several full-page miniatures subdivided to include various scenes; the Tournois manuscript of René has the exceptional arrangement of full-page miniature over two pages, a verso and the following recto, containing a single scene. The nearest approximation to the full-page miniature in most secular books is found on the opening page, which is usually treated differently from the rest of the decoration of the book. In the fifteenth century the opening page usually contains a miniature taking up three quarters of the page, with a single scene or divided into several scenes, with the first few lines of the text at the bottom. The rest of the illumination of the book generally takes the form of small miniatures in one column of the text. The idea of emphasizing the opening page is already present at the beginning of the thirteenth century; some of the earliest secular MSS contain only one miniature at the beginning of the text. By the middle of the thirteenth century when cycles of secular illumination have developed it is usual to find some distinction in the opening illustration. Either it is larger in scale than the rest, or it is different in type; for instance it may be a historiated initial (fig. 15) if the rest of the illustration consists of miniatures, or vice-versa. By 1300 it is the established practice to mark the opening page by a series of miniatures, two or four, three or six, depending on whether the text is written in two or three columns. An alternative is a single miniature divided into compartments, and this is the arrangement that leads on to the fifteenth-century practice of including a large miniature with a single scene or a composite one. See below, p. 00. This was written in 1264. See F. Saxl, Lectures, London, 1957, “The Troy Romance in French and Italian Art,” pp. 125–78, pl. 74 a, b; 75 a, b; 78 a; 79 a. See now E. Morrison, Illuminations of the Roman de Troie and French Royal Dynastic Ambition (1260–1340), Ph.D., Cornell University, 2002, 83–105; for another view see Stones, Gothic Manuscripts, Part II vol. 1, Cat. no. V–4. 34 35

SECULAR MANUSCRIPT ILLUMINATION IN FRANCE

11

Decorative borders are employed from the last quarter of the thirteenth century as part of the opening page layout. This represents the adaptation, on a much reduced scale, of an element that forms an important part of northern French and Flemish liturgical illustration from the late 1260s to the middle of the fourteenth century. The main element in the border is a leaf stalk which emanates from the initial at the top of the page and runs down the left-hand side of the page and into the bottom margin. An integral part of such borders is the inclusion not only of leaves, but also of hybrids, grotesques, apes, and human figures playing games, hunting, and indulging in a wide variety of activities. Paradoxically, this kind of marginal decoration plays a major part in the illustration of the small-scale Psalters and Books of Hours that were the popular aid to private devotion in the period 1280– 1340. Marginal decoration of this sort, usually totally unrelated to the text on the page, appears on most pages of these small-scale books. In the MSS containing secular texts, by contrast, such decoration is almost always (the Yale Lancelot and the Bodleian Alexander are exceptions) confined to the opening page only, or for the opening page of each new text in a volume. The far greater size of the pages in the secular books as well as their great length probably explain why border decoration fills only a minor rôle. In the thirteenth and early fourteenth centuries the scheme for the cyclical pictures scattered in the body of the text may be either miniatures or historiated initials. Both are to be found in Bible illustration, where they mark the beginning of each of the books of the Bible, and historiated initials are used to mark the main divisions in the text of Psalters.36 After the second quarter of the fourteenth century the historiated initial becomes obsolete. The reason is doubtless inherent in the form, which allows little scope for the battle scenes, banquets and the like, which make up the standard repertoire of secular illumination. From the last third of the thirteenth century the miniature may take a variety of forms. Mention has already been made of the composite miniatures of opening pages. There are two main possibilities for the miniatures of the main illustrative cycle. Either they may be small, square or rectangular in one text column, usually representing a single scene; very occasionally (the Yale Lancelot and BnF fr 95) (figs. 11, 13, 14) they may be divided 36 G. Haseloff, Die Psalterillustration, Kiel, 1938, provides an analysis of the subjectmatter of the historiated initials. See now F.O. Büttner, ed., The Illuminated Psalter: Studies in the Content, Purpose and Placement of its Images, Turnhout, 2005.

12

into two registers, an idea probably derivative from Psalter illustration in which such division is frequent in full-page miniatures. The alternative is long rectangular miniatures extending horizontally over the whole of the written space, generally two or three columns of the text. The earliest MS in which these long rectangular miniatures appear is Ars 3139, Chevalier au cygne, written in 1268. The arrangement is taken up in the well-known Lancelot BnF fr 342,37 written in 1274, the Pierpont Morgan Lancelot (fig. 2), and occasionally in the fifteenth century, for instance in BnF fr 117–20, the duke of Berry’s Lancelot. This layout appears to be without parallel in liturgical illumination. One of its main advantages is that it enables scenes of continued action to be represented, showing one scene following directly on another, or the continued actions of one episode. The stylistic developments that one can observe in secular MS painting between the thirteenth and sixteenth centuries parallel those in liturgical illumination, and indeed, in the later part of the period, those in the other two-dimensional media like panel painting and tapestry. The introduction of marginal illustration in both liturgical and secular books in the 1260s bears witness to a growing interest in levels of reality (or of fantasy) beyond those of the picture-plane of the main miniature. This is just one aspect of the growing awareness of space and pictorial depth and the ability to render them in two-dimensional media that is one of the most important developments in this period. It is accompanied by the increasing interest shown in natural forms used in border illustration around 1300 and later employed within the miniatures themselves, where gold and diaper “sky” backgrounds are replaced by natural colours, landscapes are better adapted to the scale of the figures in them, and architecture is rendered in convincing perspective. Techniques and Transmission The number of scenes included in the longest secular cycles is so vast that there is necessarily some degree of repetition. The stock repertoire consists to a large extent of general scenes that can be used regardless of specific context, such as scenes of joust battle and tournament, banquets, encounters with hermits, conversations between knights and ladies, kings receiving letters, characters in bed, and so on. Some such scenes may be equally applicable to 37

Loomis, op. cit., pl. 213–6.

SECULAR MANUSCRIPT ILLUMINATION IN FRANCE

1. Bel-Acueil shows the lover the rose. London, BL Harley 4425, Roman de la Rose, fol. 36 (© British Library Board)

13

14

2. Initial: Galehot leaves an abbbey where he has rested to cure a wound. New York, Pierpont Morgan Library, M.806, Lancelot, fol. 155 (© Morgan Library and Museum)

SECULAR MANUSCRIPT ILLUMINATION IN FRANCE

3. Detail: A knight, probably Guillaume de Termonde, kneels before a lady. New Haven, The Beinecke Rare Book and Manuscript Library, Yale University, MS 229, Queste del saint Graal, fol. 187 (© Yale University)

15

16

4. Four scenes of combat. Paris, BnF fr 15104, La noble chevalerie de Judas Machabée et de ses nobles frères, fol. 26 (© BnF)

SECULAR MANUSCRIPT ILLUMINATION IN FRANCE

5. Miniature: Philip IV of France, the future Philip V (or Charles de Valois) and Charles IV (or Philippe VI), with their sister Isabella (or Margaret, widow of Edward I), Louis of Navarre (the future Louis X) and Charles, comte de la Marche (or Louis d’Évreux or Charles of Valois). Paris, BnF lat. 8504, Dina et Kalila, fol. 1v (© BnF)

6. An unknown lady receiving a book from a kneeling man. Paris, BnF fr 2186, Roman de la Poire, fol. 10v (© BnF)

17

18

7. Miniature: Mordret and his men besiege the Tower of London. London, BL Add 10294, Mort Artu, fol. 81v (© British Library Board)

8. Initial: The Siege of Jerusalem. Paris, BnF fr 9081, Guillaume de Tyr, Histoire de la guerre sainte, fol. 77 (© BnF)

SECULAR MANUSCRIPT ILLUMINATION IN FRANCE

9. The god of love and two lovers. Paris, BnF fr 2186, Roman de la Poire, fol. 1v (© BnF)

19

20

10. Last Judgement scenes. Oxford, Bodleian Library, MS Douce 381, fol. 124 (© Bodleian Library)

SECULAR MANUSCRIPT ILLUMINATION IN FRANCE

21

11. Miniature: top, Lancelot and Mordret see a white stag surrounded by lions; below, they encounter two knights. New Haven, The Beinecke Rare Book and Manuscript Library, Yale University, MS 229, Lancelot, fol. 126 (© Beinecke Library)

12. Border: Jonah and the whale; ape astride a peacock. Brussels, BR 10607, Psalter of Gui de Dampierre, fol. 177r (© BR)

22 13. Miniature: top, Bohort, Gauvain and Lancelot return to Arthur’s castle, Guinevere watches their arrival; below, Lancelot speaks to Guinevere, knights stand in the doorways. New Haven, The Beinecke Rare Book and Manuscript Library, Yale University, MS 229, Mort Artu, fol. 290v (© Beinecke Library)

14. Miniature: top, King Brangoire assembles men and supplies in the city of Estragoire; below, a combat. Paris, BnF fr 95, Merlin, fol. 205v (© BnF)

SECULAR MANUSCRIPT ILLUMINATION IN FRANCE

15. Historiated initial: The Trinity enthroned. Paris, BnF fr 95, Estoire del saint Graal, fol. 1 (© BnF)

23

24

16. Beatus initial, Psalm 1: top, David, enthroned, plays the harp; below, David kills Goliath. Paris, BnF lat 1076, Psalter, fol. 17 (© BnF)

17. Miniature: the lover sees his reflection in a stream. Paris, BnF fr 802, Roman de la Rose, fol. 12 (© BnF)

SECULAR MANUSCRIPT ILLUMINATION IN FRANCE

25

religious subjects: royal banquets with a king in the center and a servant on the nearside of the table fits the standard iconography for the Last Supper with Christ in the center and Judas on the near side of the table; battle scenes may be purely secular, or they may illustrate the books of Kings or Macchabees from the Bible; and there are many more such examples. In the same way “moduli” — figures in certain positions, or parts of scenes rather than the scene in its entirety — may be reused from scene to scene, manuscript to manuscript, and from secular book to liturgical as well as vice-versa. It is clear from an analysis of the cycles of MSS containing any one text that the position of the illustrations in the text and the content of the illustrations were not invented afresh for each manuscript. They tend to remain fairly standard, and it is possible to work out recensions for the iconographical cycles in much the same way as it may be done with texts.38 Artists relied both on visual models and on written instructions. The visual models could be derived either from another MS of the same text with the same illustrative cycle, or from a model-book containing representations of stock scenes.39 They could be transferred directly to the space they were intended to occupy in the MS in preparation, or they could be sketched roughly in the margin first, to act as guides. Such a system as this presupposes a large workshop in which various individuals were involved in the production of the book. The marginal sketches probably were intended to be rubbed out or removed by trimming, but a few have been preserved40 (fig. 17). The written instructions were doubtless also intended to be removed once the miniatures were completed. They are usually generalized in type, not mentioning the characters in question by name, but they often include precise information about such things as the tinctures and devices to be represented on shields (fig. 2). It is possible too that in some MSS where there are rubrics, or 38 For Psalters see Haseloff, op. cit. and Büttner, op. cit. An archive of Bible illustration in the 12th and 13th centuries by P. Brieger is housed at the University of Toronto. For Alexander see Ross, op. cit., and for Guillaume de Tyr, see Folda, op. cit. For the Queste and Mort Artu sections of the Lancelot, see M.A. Stones, The Illustration of the French prose Lancelot, thesis, University of London, 1970 and now http://www.lancelot-project.pitt.edu. 39 See R.W. Scheller, A Survey of Medieval Model Books, Haarlem, 1963; and id., Exemplum: Model-Book Drawings and the Practice of Artistic Transmission in the Middle Ages, Amsterdam, 1995. 40 See D. J. A. Ross, “Methods of Book-Production in a Fourteenth-Century Miscellany,” Scriptorium, VI, 1952, pp. 63–75. See now Alexander, Medieval Illuminators.

26

captions above the miniatures explaining their content, that these may have been used also as guides to the illuminators.41 Secular manuscript illustration outside France The beginnings of secular illumination are better documented in Germany than they are in France. Many of the earliest examples are MSS containing texts that are based on French works, like the Ruolantesliet, Heidelberg University Library Pal.Germ. 112, dated between 1180 and 1195.42 Its decorative scheme consists of line-drawings interspersed with text, a form rare in Fiance, though found in England and Italy. Almost as rare in French secular illumination is the full-page layout, which is standard in much of German work, notably in Heinrich von Veldecke’s Eneit, Berlin, Mgf 282,43 dating from c. 1220, and Wolfram’s Parzival, Munich cgm 19, of c. 1250.44 There is no parallel for such a scheme in French illumination of this date, since both Poire and the Troie BnF fr 1610 are both at least a decade later than 1250. Whether one may assume lost French prototypes for the MSS with full-page decoration, to parallel the French source of the texts, is uncertain. What has been convincingly suggested, however, is that the Ruolantesliet may depend in part at least on a lost French Chanson de Roland MS, perhaps of English manufacture.45 In Italy secular manuscript illumination is dominated by Dante’s Divine Comedy, of which detailed study has been made.46 What is interesting from the point of view of French texts is that many MSS were written in See C.E. Pickford, L’Évolution du roman arthurien en prose, Paris, 1959, Deuxième partie, la technique, pp. 129–75, and F. Wormald, “A Medieval Description of Two Illuminated Psalters,” Scriptorium, VI, 1952, pp. 18–25. 42 See Lejeune and Stiennon, Roland, chapter XIV, pp. 111–138, pl. 84– 90, 92, 94–124. 43 Saxl, op. cit., pl. 73. 44 Loomis, op. cit., pl. 355–358. 45 Ross, Iconography of Roland, pp. 55–59. For German manuscripts see now D. Green, Medieval Listening and Reading: The Primary Reception of German Literature 800–1300, Cambridge, 1994; M. Curschmann, Wort-Bild-Text: Studien zur Medialität des Literarischen in Hochmittelalter und früher Neuzeit, 2 vols., Baden-Baden, 2007; J. Wolf, Buch und Text: Literatur-und kulturhistorische Untersuchungen zur volkssprachigen Schriftlichkeit im 12. und 13. Jahrhundert, Tübingen, 2008. 46 P. Brieger, M. Meiss, C.S. Singleton, Illuminated Manuscripts of Dante’s Divine Comedy, Princeton, 1969. See also D. Delcorno Branca, Boccaccio e le storie di re Artù Bologna, 1991, and Allaire and Psaki, op. cit. 41

SECULAR MANUSCRIPT ILLUMINATION IN FRANCE

27

French by Italian scribes and illustrated in Italy. The most popular texts were the prose Lancelot and Tristan, and Guillaume de Tyr’s Histoire de la guerre sainte.47 The best known MSS from the Iberian peninsula are the songbooks, notably the Cantigas of Alfonso X, but it seems that a few French MSS may have been illuminated in Spain, among which is BnF fr 750, a Tristan written in France but probably illuminated in Spain.48 An interesting situation arises where the Crusading Kingdom of Jerusalem is concerned. Manuscripts were written and illuminated there by the Acre scriptorium on the one hand, and there is also evidence that a French illuminator went there and illuminated books, two of which are Guillaume de Tyr MSS,49 a third a Lancelot, Tours 951. England seems to have produced very little secular illumination in the Middle Ages by comparison with Italy and Germany. There is no great tradition of Chaucer illustration to correspond to that of Dante in Italy. In fact none of the MSS containing Chaucer’s texts have narrative cycles to illustrate them, and even the most splendid, the Ellesmere MS of the Canterbury Tales, now in the Huntington Library, San Marino, California, has only small marginal pictures showing each of the pilgrims next to the opening lines of his tale. The portraits of Chaucer in manuscripts of his and Hoccleve’s work add a new dimension to secular illustration50 which makes the lack of illustration of the texts themselves the more surprising. Against such paucity of illustrated English texts, one may set the few illustrated French texts made in England. Matthew Paris’ workshop seems to have given rise around the middle of the thirteenth century to a small group of secular manuscripts illustrated in line drawing like the Chanson d’Aspremont, BL Landsdowne 782, and the Roman de toute chevalerie by Thomas of Kent, Cambridge, Trinity College, 0.9.34 and BnF fr 24364.51 47 Loomis, op. cit., pl. 305–337; see also S. Parkinson, ed. Cantigas de Santa Maria: An Anthology, Cambridge, 2015. 48 Ibid., pl. 210–212; now thought to have been made either in southern Italy or the Holy Land, see T. Delcourt, ed., La Légende du roi Arthur, Paris, 2009, no. 36, p. 118, by M. Bessyre. 49 H. Buchthal, Miniature Painting in the Latin Kingdom of Jerusalem, Oxford, 1957, and Folda, op. cit., Part I. Folda names the French illuminator the ‘Master of the Knights Hospitaller.’ 50 Rickert, op. cit., pp. 185–6. Cf. also the portrait of Adenet in Ars 3142. 51 D.J.A. Ross, “A thirteenth century Anglo-Norman workshop illustrating secular literary manuscripts?,” Mélanges Lejeune, Liège, 1968, pp. 689–94; see now F. Avril and P.D.

28

In general, then, French manuscripts did play a part in the development of secular illumination outside France, while in France itself the exceptionally large number of them produced from the thirteenth century onwards bears witness to the degree of public interest of which they were the object in the late middle ages. The different libraries of Europe and the United States, in which they are now located, testify also to the tastes of the bibliophiles of the succeeding centuries, and it is fitting that so much recent scholarship has turned its attention to their illumination.

Stirnemann, Manuscrits enluminés d’origine insulaire VIIe–XXe siècle, Paris, 1987, no. 171 (BnF fr 24364), and M. Meuwese, “The Exploits of Alexander the Great in Trinity College,” in The Cambridge Illuminations: The Conference Papers, ed. S. Panayotova, London and Turnhout, 2007, pp. 129–37.

II Sacred and Profane Art: Secular and Liturgical Book-Illumination in the Thirteenth Century

T

he eleventh to thirteenth centuries mark the emergence of new directions in medieval thought which found their reflection in the development of a secular tradition in literature and in art. The flowering of new literary genres — epic, romance, and poetry — in France and Germany in the late eleventh and twelfth centuries bears witness to this and the popularity of the new material continued into the thirteenth century and beyond, in the translations and prose versions of poems relating the deeds of the great heroes like Roland, William, and Arthur. There is much in the stories and legends of these figures that reflects the Christian ethic; the cult of Charlemagne arose at the end of the twelfth century,1 Roland became a popular saint although he was never officially canonised2 as was William,3 while the quest for the Holy Grail is permeated with christological symbolism.4 The illustration of this new textual material likewise marks a new development in medieval book-production, and in many ways the illuminators, like the authors of First published in The Epic in Medieval Society: Aesthetic and Moral Values, ed. Harald Scholler, Tübingen, 1977, 100–12. 1 It seems that popular devotion to Charlemagne began in Germany at the time of the quarrel between Frederick Barbarossa and the pope. The title of “Blessed” was conferred upon Charlemagne by Benedict XIV. See Acta Sanctorum, III, 3, pp. 490–507, January 28. 2 Despite the fact he was never canonised, Roland is represented among a group of martyrs in the sculpture of the south door jambs of the cathedral of Chartres (see Lejeune and Stiennon, Roland, II, pp. 191–203), and in the glass of Chartres he is shown with a halo (ibid. pp. 203–07). 3 Canonised in 1066. See Acta Sanctorum, XIX,6, pp. 798–817, May 28. 4 See in particular the description of the Grail liturgy in La Queste del saint Graal, ed. A. Pauphilet, Paris, 1923, pp. 269–70.

30

the texts, relied to a large degree on patterns already familiar from liturgical illumination. At the same time it becomes apparent in thirteenth-century bookillumination that cross-fertilisation is at work; in the first half of the century the illumination of secular texts5 is still rare and their quality is often inferior in scope and in execution to their liturgical counterparts, while the illuminators adapt liturgical themes to meet the requirements of their secular texts. By the end of the century, however, the best workshops not only made splendid MSS of secular content as well as religious books, but the decorative layout of the latter reflect the secular content of the former, while the motifs themselves take on an independence of the context of the text they illustrate, whether it be sacred or profane in nature. It is the purpose of this paper to examine the kinds of change that take place during the thirteenth century, drawing examples from the illustration of epic, romance, and poetry, and setting them in relation to religious illumination. Since it is in France that the sequence of secular illumination may best be followed, most of the material here selected for discussion is French. The relations between picture, text, and model will be considered first, followed by an analysis of secular and liturgical products of the same shop. The earliest illustrations in secular texts tend to rely heavily on generalised types without rendering the complexities of the situations their texts narrate. One of the earliest secular MSS to contain a cycle of pictures is the Ruolantesliet Heidelberg, Universitätsbibliothek, MS cod. pal. germ. 112.6 (fig. 1) In combat scenes, pagans are not visually distinct from Christians; the illuminators use topoi such as could have been adapted from such sources as, for instance, Psychomachia illustration, of which there are 5 In art, as in literature, the distinction between the “secular” and the “liturgical” is not an easy one to make, nor would the difference between what we would now term the “sacred” and the “profane” necessarily have occurred to the medieval mind. In terms of bookproduction one may reasonably distinguish between different types of book on the basis of the need for which the book was created. Thus “liturgical” is here understood to refer to books created for use in the liturgy of the church, interpreted in the broadest sense to include the Bible as well as Psalters, Breviaries, Missals etc., while “secular” refers to books made to be read and looked at for general enjoyment and entertainment. We are concerned in the present inquiry with the degree to which “sacred” and “profane” elements overlap from a “liturgical” to a “secular” context or vice versa. 6 Facsimile Wiesbaden, 1970. Lejeune and Stiennon, op. cit. 1, pp. 111—138.

SACRED AND PROFANE ART

31

eleventh and twelfth-century illuminated copies,7 or a scene drawn from biblical illumination from a book such as Kings or Maccabees. This famous late twelfth-century Ruolantesliet in many ways stands at the beginning of secular illumination used for texts written in the vernacular; its illustrations form a sharp contrast to later representations of the Roland story such as the Chartres window of c. 12208 or the Karl der Große MS of c. 1300, St. Gall, Biblioteca Vadiana 302 (figs. 2 and 3).9 Both MSS contain scenes of Roland’s death, showing the hero breaking his sword on the rock and blowing his oliphant. In each case, the representation conveys a great deal of the narrative detail of the text version they illustrate, showing that at some stage in the planning of the illumination, either the artist himself, or the person in charge of the illustrative layout of the book or window, must have read the text and based pictorial compositions directly on its verbal description. Another method of text-illustration found in thirteenth-century secular MSS shows that the artist’s imagination might also be aided by familiar prototypes in a manner somewhat akin to the use of generalised scenes in the Ruolantesliet MS dicussed above. An example of this is the illustration of an episode from the Queste section of the French prose Lancelot.10 The story describes how Lancelot falls asleep beneath a cross, whereupon the Grail appears. Lancelot, since he is asleep, does not see the Grail, but the latter is seen by a lame man on a bier, whom Lancelot has just rescued. Only six Queste MSS contain an illustration of this episode.11 In BnF fr 342 (fig. 4) the Grail is not represented at all in the miniature; but in the other MSS it is represented as a chalice (figs. 5–7). The illuminators make use of a well-established liturgical object in their depiction of this important 7 See H. Woodruff, The Illustrated Manuscripts of Prudentius, Cambridge, MA, 1930. Cf. also the Psychomachia illustrations in the Hortus Deliciarum of Herrad of Landsberg (A. Straub, Hortus Deliciarum, Straßburg, 1879–99; now ed. and tr. by A.D. Caratzas, New Rochelle, NY, 1977; and R.B. Green et al., eds., Hortus deliciarum of Herrad of Hohenbourg, London, 1979). 8 See above, note 2. The iconography of this window is assumed to represent scenes from a now lost version of the Roland story. See, in addition to Lejeune and Stiennon, Ross, Iconography of Roland, and now C. Maines, “The Charlemagne Window at Chartres Cathedral: New Considerations on Text and Image,” Speculum, 52, 1977, pp. 801–23 and M.J. Schenck, “Taking a Second Look: Roland in the Charlemagne Window at Chartres,” Olifant, 25, 2006, pp. 371–85. 9 Lejeune and Stiennon, op. cit. I, pp. 226–238, pl. XXIV–VI. 10 Queste, ed. Pauphilet, pp. 57–62. 11 This episode is illustrated in BnF fr 342, written in 1274; Bonn 526, written in 1286; BL Add 10294 and Roy. 14 E. III; Manchester, Rylands French 1. The last three MSS all

32

life-giving symbol, which the French prose version of the text describes as a vessel. The liturgy of the Grail as described in Queste refers to the presence in the Grail of a bleeding figure, a concept no doubt inspired by the concept of the real presence of Christ in the Eucharistic Host, which, as the doctrine of Transubstantiation, was in place by the fourth Lateran Council in 1215. Two of the Queste illustrations reflect this concept, as manifested not only in the text which they illustrate, but also in the doctrine of the church and in liturgical illumination. In Bonn 526, the chalice contains a cross; this motif is not used in the text in the verbal description of the Grail, either at the point in the text when it appears before Lancelot, or later, in the Grail liturgy. That the motif in the Queste illustration is derived from liturgical sources is suggested by the appearance of the same motif in the illustration of the Moralised Bibles12 (fig. 8) in representations concerning the consecration of the sacred elements. The representation of the Grail liturgy in Ars 5218 shows the Grail as a chalice containing a human figure that is bleeding. come from the same workshop and were made c. 1315–25; Ars 5218, although it contains no miniature for this particular episode, does include a unique representation of the Grail liturgy whose iconography is of great interest to the present discussion. It was written and illuminated in 1351 by Piérart dou Tielt. See H. Martin, “Un caricaturiste au temps du roi Jean, Piérart dou Tielt”, Gazette des Beaux-Arts, 2, 1909, pp. 89–102 and now F. Avril in Fastes du Gothique, le siècle de Charles V (exhibition catalogue), Paris, 1981, no. 300; L.J. Walters, “Wonders and Illuminations: Pierart dou Thielt and the Queste del saint Graal,” in K. Busby, ed., Word and Image in Arthurian Literature, New York, 1996, pp. 339–72. 12 The Moralised Bibles were royal commissions, initiated in all probability by Blanche of Castille for her husband Louis VIII and son Louis IX, kings of France; the portrait of the commissioning king in Vienna 1179 is likely to be Louis VIII and the portrait of the young Louis IX along with that of his mother, Blanche of Castille, appears in the frontispiece illustration in the New York section of the Toledo Bible. The thirteenth century Moralised Bible MSS are: 1) ÖNB 2554, 2) ÖNB 1179, 3) New York, Morgan Library MS 280 and Toledo Cathedral, Tesoro, Biblia de San Luis), 4) BnF lat 11560; BL Harley 1526–7; Bodleian Library Bod. 270b. For the Oxford/Paris/London MS see A. Laborde, La bible moralisée illustrée, Paris, 1911–27. A facsimile of the Vienna 2554 MS and a book on the Moralised Bibles are being prepared by Prof. Dr. R. Haussherr (see now Bible moralisée: FaksimileAusgabe im Originalformat des Codex Vindobonensis 2554, Graz and Paris, 1973). See also R. Haussherr, “Tempel Salomonis und Ecclesia Christi”, Zeitschrift für Kunstgeschichte, 1968, pp. 101–121, and “Sensus litteralis und sensus spiritualis in der Bible moralisée”, Frühmittelalterliche Studien, 6, 1972, pp. 356–80. See now Bible moralisée: Prachthandschriften des Hohen Mittelalters: gesammelte Schriften von Reiner Haussherr, ed. E. König, C.T.Seifert, G. Seifert, Petersberg, 2009; J. Lowden, The Making of the Bibles moralisées, 2 vols., University Park, PA, 2000; and the facsimile of the Toledo-Morgan Bible moralisée, Biblia de San Luis, 5 vols., Barcelona, 2000–2002.

SACRED AND PROFANE ART

33

While it is possible to point to parallels in liturgical illumination for the consecrated host with a human figure emerging from it, there appear to be no parallels for a chalice containing a figure, far less a figure that is bleeding (Paris, Musée de Cluny, S. Martin embroidery). It would therefore appear that we witness in this mid-fourteenth-century MS a return to the direct method of deriving picture from text that is evident more than a century earlier in the Chartres window. Secular text illustration, then, employs no single method in selecting models, but may equally well adapt a model from a liturgical context or invent one on the basis of the words of the text concerned. A consideration of secular MSS in relation to the workshops that produced them leads to other kinds of parallels between secular and liturgical illumination. There is evidence that both types of book were illuminated by the same illuminators, although in the thirteenth century we are still in a period in which comparatively little is known about the exact circumstances of book-production because of the scarcity of documents, unlike the fifteenth century, when it is possible to follow in detail the work of different scribes, decorators, and illuminators in different books and to trace their movements as they shifted from one workshop to another.13 For the first half of the thirteenth century our concept of the artistic workshop is based largely on similarity of style and motif; there is also some evidence from tax records, in particular for Paris, to show that bookproduction was by and large a commercial enterprise operating in towns rather than in monasteries, and that the work was on the whole done by lay craftsmen.14 This applies not only to secular books but also to liturgical ones including the Bible. As far as documentary evidence is concerned, there is far more material extant from the late thirteenth century, both from the Paris tax rolls of the decade of the 1290s and also from the tax records and 13 See in particular the exhibition catalogue by the late L. M. J. Délaissé, Le Mécénat de Philippe le Bon, Brussels, 1959, which pioneered study in this area. 14 R. Branner, “Manuscript makers in thirteenth-century Paris”, Art Bulletin, 1966, pp. 65–7; H. Géraud, Paris sous Philippe le Bel d’après des documents originaux et d’après un document contenant le rôle de la taille imposée sur les habitants de Paris en 1292, Paris, 1837; K. Michaëlsson, “Le Livre de la taille de Paris, 1313”, Acta Universitatis Gothoburgensis, Göteborg, 195I, also for 1296 (ibid. 1958) and 1297 (ibid. 1961). This tax roll also includes records for 1298, 1299 and 1300 which arc still incompletely edited. See, however, F. Baron, “Les peintres, enlumineurs et imagiers parisiens”, Bulletin de la Société française d’archéologie, 1968, pp. 1–49, which includes material from all the rolls. See now Rouse and Rouse, Manuscripts and their Makers.

34

1. Heidelberg, cod. pal. germ. 112, Ruolantesliet, fol. 63, combat between pagans and Christians

SACRED AND PROFANE ART

2. St Gall, Vadiana 302, Der Stricker, Karl der Große, fol. 52v, death of Roland

35

36

3. Chartres Cathedral, Charlemagne window, death of Roland (© Philip Maye)

4. Paris, BnF fr 342, Lancelot, fol. 77, Lancelot asleep as the Grail appears (© BnF)

SACRED AND PROFANE ART

5. Bonn Landes-und Universitätsbibliothek 526, Queste del saint Graal, fol. 426, Lancelot asleep as the Grail appears (© Landes-und-Universitätsbibliothek Bonn)

37

38

6. Manchester, John Rylands University Library, French 1, Queste del saint Graal, fol. 195v, Lancelot asleep as the Grail appears (© Lancelot-Grail Project)

7. Paris, Ars 5218, Queste del saint Graal, fol. 88, the Grail liturgy (© BnF)

SACRED AND PROFANE ART

8. Oxford, Bodleian Library, Bod. 270b, fol. 50, Moralised Bible

39

40

9. New York, Pierpont Morgan Library, M.638, Old Testament Picture Bible, fol. 28v, David and Goliath (photo after facsimile)

10. New York, Pierpont Morgan Library, M.638, Old Testament Picture Bible, fol. 3v, The rescue of Lot (photo after facsimile)

SACRED AND PROFANE ART

11. Paris, Ars. 3139, Chevalier au cygne, fol. 109, siege of Nicea (photo: author)

41

42

12. Bonn, Landes-und-Universitätsbibliothek 526, Estoire del saint Graal, fol. 1, scenes from the life of the hermit-author and Joseph of Arimathea (© Landes-und-Universitätsbibliothek Bonn)

SACRED AND PROFANE ART

13. Bonn, Landes-und-Universitätsbibliothek 526, Estoire del saint Graal, fol. 97v, battle between the King of Nantes and the Saxons (© Landes-und-Universitätsbibliothek Bonn)

14. Boulogne-sur-Mer, BM 192, Cycle de Guillaume d’Orange, fol. 310, Guillaume fights a giant (photo: author)

43

44

15. Saint-Omer, BM 5, Bible, fol. 145v, Illustrations to I Maccabees (photo: author)

16. Douai, BM 193, Psalter, f. 143r, Psalm 80

45

SACRED AND PROFANE ART

17. Saint-Omer, BM 5, Bible, fol. 62, Jeremiah (photo: author)

18. Montpellier, Bibl. interuniversitaire, Section médecine, H 196, Chansonnier (Motet Book), fol. 87v, Trinity and game (photo: author)

46

19. Montpellier, Bibl. interuniversitaire, Section médecine, H 196, Chansonnier (Motet Book), fol. 88 Trinity and game (photo: author)

SACRED AND PROFANE ART

20. Verdun, BM 107, Breviary, fol. 1, Psalm 1, border: knight pursues Saracen (photo: author)

47

48

town plans of the north-eastern French and Belgian commercial centres like Arras, Cambrai and Tournai, showing that scribes, illuminators, parchment makers, bookbinders and book-dealers all had their shops in neighbouring streets and that their operations were closely interrelated.15 A comparison between the secular and liturgical products of some workshops from the mid-thirteenth to the early fourteenth centuries indicates the kinds of changes that occur in the course of that period in relation to the layout or type of illustration selected for each kind of text and the relative quality of each, as well as shedding further light on the question of dependence or independence of illustration on text. Few of the secular MSS made around the middle of the thirteenth century in France contain illumination that is comparable either in quality of execution or in complexity of layout and design to that of the major Bibles of the period. Apart from the Moralised Bibles mentioned above, the most spectacular collection of illustrative bible material surviving from this period must surely be the Old Testament Picture Bible, New York, Pierpont Morgan Library MS 638. It contains a full-page sequence of illustrations starting with Genesis and ending with the end of the book of Kings.16 The number of pictures per biblical episode, the emphasis on dramatic narrative, and the attention to detail make this sequence of Old Testament illustrations unique17 while puzzles also remain as to the original patron, the provenance, and the date of the MS. There is certainly no secular epic or romance that contains so full a cycle of pictures depicting battles, cruelty or heroism, nor indeed does any secular illustration of the period rival the wealth of narrative C. Dehaisnes, Documents et extraits divers concernant l’histoire de l’art dans la Flandre, l’Artois et le Hainaut avant le XVe siècle, Lille, 1886; id., L’Art à Douai dans la vie privée des bourgeois, Paris, 1864; J. Lestocquoy, Études d’histoire urbaine, villes et abbayes: Arras au moyen âge, Arras, 1966. See now Le rentier d’Artois, 1298–1299. Le rentier d’Aire, 1292, ed. R. Berger et al., 2 vols., Arras, 2006. 16 Facsimile New York/London, 1969; see now D. Weiss et al., The Morgan Crusader Bible Commentary, Lucerne, 1999, suggesting ownership by Louis IX; for another view see A. Stones, “Questions of Style and Provenance in the Morgan Picture Bible,” in Between the Picture and the Word: Manuscript Studies from the Index of Christian Art, ed. C. Hourihane, University Park, PA, 2005, pp. 112–121. 17 See now H. Stahl, The Iconongraphic Sources of the Old Testament Miniatures, Pierpont Morgan Library, M.638, Ph.D. thesis, New York University, 1974. As regards its elaborate sequence of full-page miniatures, the MS stands out not only among bible illustration but also in relation to secular illumination, where the full-page miniature as such is comparatively infrequent. Despite the fact that full-page pictures are the rule in much of 15

SACRED AND PROFANE ART

49

detail that characterise the work of this illuminator. One is struck not only by the realistic brutality of the scenes of combat but also by the different types of arms and armour here represented, showing different varieties of helmet, kettle-hat and helm, mail coifs, plate greaves for Goliath (fig. 9), shield straps and sword belts. Particularly noticeable in the representations of armour are the padded gambesons and coudières (fig. 10), and in the battle equipment most striking are the elaborate mechanisms of war.18 In a broad sense this wealth of detail must have been observed from life and may be seen as a reflection of the general move towards naturalism which found other manifestations in thirteenth-century France such as the sculpted foliage of the capitals and the interior west wall of the cathedral of Reims.19 However, unlike the sculpture of Reims, for which one can find parallels at the Sainte-Chapelle20 or at the cathedral of Southwell in England,21 no parallels have so far been found for the treatment of arms, armour, and battle equipment in this MS. It is largely on the basis of the use of similar details that a secular MS may now, I believe, be attributed to the same workshop as the Bible. It is the Chevalier au cygne MS Paris, Ars 3139.22 Superficially there is very little resemblance between the illustrations of this MS and the Bible. The Cygne the German secular illumination of the thirteenth century (e.g. Eneit, Berlin SBPK Mgf 282, Tristan, Munich BSB Cgm 19; Parzival, Munich BSB Cgm 51; and the Karl der Große cited above, St. Gall, Vadiana 302) the most common form for secular illumination in France is small miniatures, or, more occasionally, historiated initials, scattered in the text (see below). Comparatively few MSS contain even a few full-page miniatures in the thirteenth and turn of the fourteenth centuries. They are BnF fr 1610 (see F. Saxl, “The Troy romance in French and Italian art”, Lectures, I, London, 1951, pp. 123–138, and H. Buchthal, Historia Troiana, London, 1971; see now E. Morrison, Illuminations of the Roman de Troie and French Royal Dynastic Ambition (1260–1340), Ph.D., Cornell University, 2002); BnF fr 12558 and 12559, Cygne; BnF fr 2186, Poire; Berlin Kupferstichkabinett 78.C.1, BR 11040 and BL Harley 4979, Alexander (to which another copy in private hands has been added); BnF fr 1433, Yvain; and BnF fr 146, Fauvel. For the French manuscripts see now Stones, Gothic Manuscripts. 18 See especially facsimile plates 74 (fol. 10v), 150 (fol. 23v) and 215–6 (fol. 35v). 19 See W. Sauerländer, Gothic Sculpture in France, Munich, 1970. 20 See L. Grodecki, La Sainte-Chapelle, Paris, 1968 and now J.-M. Leniaud and F. Perrot, La Sainte Chapelle, Paris, 1991. 21 See N. Pevsner, The Leaves of Southwell, London, 1949 and now P. Binsky, Gothic Wonder: Art, Artifice and the Decorated Style, 1290–1350, New Yaven, 2014. 22 Written in 1268. H. Martin and P. Lauer, Les principaux MSS à peintures de la Bibliothèque de l’Arsenal à Paris, Paris, 1929, pl. X.

50

MS (fig. 11) is in very poor condition and the miniatures are badly preserved; and then the scope of the illustration is modest by contrast with the Bible. Instead of a continuous sequence of full-page miniatures, the illustration is confined to a limited number of narrow rectangles the length of two textcolumns, with coloured borders and wavy-lined motifs, quite different from the gold frames of the Bible and the architectural canopies that divide the scenes. Similarly, the complex structure of the scenes in the Bible, the variety of action and movement, and the overwhelming abundance of detail, find hardly any reflection in the illuminations of the Cygne MS. Nevertheless some of the more unusual motifs found in the Bible do reappear in Cygne, in particular the padded gambesons and coudières. These are motifs that are so rare in the thirteenth century, not only in manuscript illumination but also in monumental painting and sculpture, that their appearance in the Cygne MS must indicate that the relation between it and the Bible is very close.23 I suggest that these motifs were part of the stock repertoire of the workshop to which the Bible illuminator or illuminators belonged, and that they were motifs which could therefore be reproduced by minor illuminators of the same shop — in this case the illuminator of the Cygne MS, who, I believe, was a lesser employee of the shop, set to work on a less important text than the Bible, and one whose illumination was correspondingly done on a much less lavish scale. That the artists were, however, closely related, may be seen not only in the details of battle dress but also more generally in elements like the treatment of hair, faces, and drapery, allowing of course for the fact that one illuminator is a supreme master of his craft while the other is endowed with considerably less talent. The relative value placed on each type of book and its illumination seems to me to find its reflection in the selection of the major artist to work on the Bible pictures and the minor one to work on the secular book, despite the intrinsic similarity of the illustrative materials required for both texts.24 This scale of values changes towards the end of There were many opportunities for thirteenth-century illuminators to represent scenes of battle and combat to illustrate not only in bibles the books of Kings and Maccabees, but also in the many secular epics, romances, and historical texts like William of Tyre’s crusading history (on which see J. Folda, Crusader Manuscript Illumination at Saint-Jean d’Acre, 12751291, Princeton, 1975) where such scenes were required. Despite the large body of illustrated battle scenes that do exist I have found no other examples of the use of padded gambesons or coudières. 24 Through its association with the Arsenal Cygne MS of 1268, the Old Testament Picture Bible may now be related to a large group of MSS made in the 1260s in the area of Arras, 23

SACRED AND PROFANE ART

51

the thirteenth century.25 Crucial to that change is the emerging personality of the illuminator whose imagination begins, in part at least, to assume an independence from liturgical models and also from the words of the text he illustrates. From an important workshop dating from the last quarter of the thirteenth century and located in north-eastern France there survive few liturgical books in relation to the secular output of the shop. The best- known MS of this group is the Lancelot Bonn 526, whose Grail illustration was discussed above.26 Its illumination takes the form of small square miniatures Douai, Tournai, Cambrai, or Lille. See Vitzthum, Pariser Miniaturmalerei, pp. 117–124; E. J. Beer, “Das Scriptorium des Johannes Philomena und seine Illuminatoren,” Scriptorium, 23, 1969, pp. 24–38; H. Stahl, “Le Bestiaire de Douai”, Revue de l’art, 8, 1970, pp. 6–15; M.A. Stones, “Le Missel de Tournai”, Trésors sacrés (exhibition catalogue), Tournai, 1971, pp. 51–3; R. Branner, “A cutting from a thirteenth-century French Bible”, The Bulletin of the Cleveland Museum of Art, 58, 1971, pp. 219–227; E.J. Beer, “Liller Bibelcodices, Tournai und die Scriptorien der Stadt Arras”, Aachener Kunstblätter, 1973, pp. 190–226. Further unnoticed MSS that belong to this group are the Chansonnier London, BL Egerton 274; the Bible, New York Public Library MS 4; and the Pontifical of Cambrai in the Cathedral Library at Toledo, MS 56.19. The Toledo Pontifical provides the best comparison for the figure style of the Morgan Picture Bible. Other comparisons for the Bible style have been suggested by O. Pächt and J.J. G. Alexander in their Illuminated Manuscripts in the Bodleian Library, I, Oxford, 1966, as Douce 381 fols. 122–24 (cat. no. 536), and Douce 50, Psalter (cat. no. 535). They further connect the Bible, less convincingly in my view, with the Poire MS BnF fr 2186, and with the Missal of Saint Louis at San Francesco, Assisi. There is still a great deal of work to be done on the precise stylistic relations between the MSS of this large group. It is likely that we may be dealing not only with the work of numerous illuminators of varying quality but also with the work of itinerant illuminators who contributed to the output of different ateliers. See now A. Bräm, Das Andachtsbuch der Marie de Gavre: Paris, Bibliothèque nationale, Ms nouv. acqu.fr. 16251: Buchmalerei in der Diözese von Cambrai im letzten Viertel des 13. Jahnhunderts, Wiesbaden, 1997 and for other views, A. Stones, Le livre d’images de Madame Marie (Paris, Bibliothèque nationale n.a.fr. 16251), Paris, 1997 and ead., Gothic Manuscripts, Part I, vol. 2, Cat. nos. III–41––III–62. 25 The change, it must be admitted, is not altogether absolute. Of the secular MSS produced by Master Honoré and the shop to which he belonged, for instance, there are very few which can rival the best of the liturgical books the shop produced, like the David page of the Breviary of Philippe le Bel (BnF lat 1023) or the Death of the Virgin page of the Nürnberg Hours (Nürnberg Solger 40, 4, Hours). One such, however, may have been the opening miniature of the Brunetto Latini Trésor, Florence, Laurenziana MS Ash. 122, before it was defaced. Traces of exceptionally good drapery modelling can still be seen. 26 Loomis, Arthurian Legends, pp. 94–6, figs. 217–223. This is one of the few MSS that contain the complete prose Lancelot cycle in French. For a list of Lancelot MSS see Woledge, Bibliographie.

52

set in one text column and this arrangement is followed by the other secular MSS made in the shop,27 while the liturgical books use historiated initials as is more usual for the bulk of liturgical illumination in the thirteenth century. Two complete prose Lancelots have survived from the group, Bonn 526 and Paris, BnF fr 110. Both employ a special arrangement for the opening pages (fig. 12), with two or more miniatures and a border supporting animals or figures. The use of a border of this kind is uncommon before the end of the thirteenth century in secular MSS, while in liturgical books on the other hand it was a method of illustration established since the 1260s and continued in the liturgical MSS made in the workshop under discussion.28 The illuminators of these Lancelots adapt an established pattern from a liturgical context to a secular one. Further similarities, this time of motif, between the various MSS of this group bear witness to cross-fertilisation between the secular and the liturgical. In all these MSS the battle scenes are generalised, so that the same moduli or patterns can be and are used for different episodes and between different protagonists. In the case of the secular MSS, captions explain to the reader who the protagonists are. In figs. 13–15, not only are the mail convention and helmets similarly treated, but the same motif of a falling horse appears in each. Of particular interest in the Bible Saint-Omer MS 5 is a detail of the initial terminal which shows a knight holding a severed head. In this case the detail is related in a general way to the context of the historiated initial, which illustrates the book of Maccabees, but it is also a stock motif of this workshop and reappears in the Psalter Douai 193 (fig. 16). In a context like 27 For a detailed study of the MSS of this group see Stones, French prose Lancelot, pp. 208–224 and 451–461; and now ead., Gothic Manuscripts, Part I, vol. 2, Cat. nos. 113–131. The related MSS include, in order of stylistic similarity: olim London, Sion College Arc. L. 40 2/L. 28, Bestiary (now Los Angeles, J. Paul Getty Museum Getty Ludwig XV 4; Boulognesur-Mer 192, Cycle de Guillaume d’Orange; BnF fr 110, prose Lancelot; BL Add 5474, Tristan; S. Omer 5, Bible; Douai 193, Psalter; Arras Musée Diocésain 47, Psalter-Hours; Bodl. Douce 24, Psalter; New York, Morgan M.79, Psalter; Cambrai 153–4, Missals; BnF fr 19162 and fr 24394, Lancelot; Paris Sainte-Geneviève 2200, Bestiary; BL Yates Thompson 43, Psalter. Dated MSS are: YT 43, 1277; Getty Ludwig XV 4, 1277, Bonn 526, 1286; Boulogne 192, 1295. 28 See in particular the MSS mentioned above, note 24, especially, within that group, Cambrai 189–90 and the Toledo Cathedral Pontifical, 56.19. There is also one marginal illumination without a border in Ars 3139; cf. also the Psalter of Isabella, Cambridge, Fitzwilliam 300, facsimile ed. S. C. Cockerell, 1905, and E.J. Beer, “Zum Problem der Biblia Porta”, Festschrift für Hans R. Hahnloser, Basel, 1961, pp. 271–86.

SACRED AND PROFANE ART

53

this one, the motif of a knight and severed head may originally have been invented as an allusion to David and Goliath, but, while such a theme is in general relevant to psalm illustration it does not reflect the content of this particular psalm, whose visual accompaniment is usually David playing bells.29 Furthermore, the severed head, in this instance, is female and so could not possibly refer to Goliath. Perhaps this motif was originally made up for a David and Goliath context, but once part of the repertoire of a shop could be taken out of a pattern book, adapted, and used in a context in which it had no connection with the text.30 What is still unknown about the process of using model books is who decided what motif was relevant to what context. It becomes evident in the last decade of the thirteenth and the early fourteenth centuries that clear-cut distinctions of relevance and irrelevance or secular and liturgical, sacred and profane, become blurred. Some of the border scenes in S. Omer 5 are based on the text; such a case is the killing of an elephant on the Maccabees page, whose source is an incident in the battle of Bethzechariah, in which Eleazar crawls under the elephant and kills it with his sword, whereupon he himself is crushed as the animal falls on him (fig. 15) (I Maccabees 6,45–47). Other marginal motifs, however, are used without relation to the text. In Saint-Omer 5 is a figure of a man throwing a stone. This motif, like the knight and severed head, is one of many patterns that appear several times within the products of this shop, each time used in a different, and equally irrelevant, context. The stone-throwing man is also found on the border of the Lancelot BnF fr 110, where it has nothing to do with the text; in the Bible it appears in the border of the Jeremiah page. This particular motif, again like the knight with the severed head, may have been invented to fit a particular context — in this case no doubt a scene of the martyrdom of St Stephen.31 (fig. 17) See the tables in G. Haseloff, Die Psalterillustration im 13. Jahrhundert, Kiel, 1938 and now F.O. Büttner, ed., The Illuminated Psalter: Studies in the Content, Purpose and Placement of its Images, Turnhout, 2004. That the actual words of the psalms were frequently taken up by illuminators has been demonstrated by F. P. Pickering, Art and Literature in the Middle Ages, London, 1970, pp. 273–85. I do not think, however, that this was the method employed in the instance under discussion. 30 R. Scheller, A Survey of Medieval Model-Books, Haarlem, 1963, provides the best overview. See also D.J.A. Ross, “A late twelfth-century artist’s pattern sheet”, Journal of Warburg and Courtauld Institutes, 25, 1962, pp. 119–28. See now Alexander, Medieval Illuminators. No pattern-sheets or books relating to the material under discussion have survived. 31 However, it is interesting to note in this context that a scene of martyrdom by stoning is used for the historiated initial to the book of Jeremiah in a group of Franco/Flemish 13th 29

54

The use of motifs derived from biblical (David) or sacred (St Stephen) sources in secular border decoration parallels the use of comparable motifs in the main miniatures or historiated initials in both types of book. What is more interesting, however, is the fact that, while the Bible of this group includes other irrelevant material, and contains a highly decorated border on every illuminated page, the secular MSS rarely employ borders at all, and such borders as do appear are very sparsely decorated.32 It might seem that the use of motifs out of context might be more admissible in a secular book, and yet, in the last decade of the thirteenth century, it is in liturgical books that the illuminators’ imagination is given freer rein. Ten years or so later there is more of a balance between the mingling of the sacred and the profane in secular and liturgical, or at least sacred, illumination. The well-known Montpellier Chansonnier MS Sect. Méd. H. 196 provides ample illustration of this.33 The MS includes poems that are century bibles: MSS BnF lat 16719–22; Basel, Öffentliche Kunstsammlung U. IX.30; Lille 835–8; New York, olim Kraus Collection, Marquette Bible, now Los Angeles, J. Paul Getty Museum, Ludwig I 8; Arras 1–3; Brussels, BR II.2523. See E. J. Beer, “Liller Bibelcodices, Tournai und die Scriptorien der Stadt Arras”, Aachener Kunstblätter, 1973, pp. 190–226. To Beer’s list one may add MSS Boulogne 4 and BL Yates Thompson 22. A convenient list of other motifs is found in L.M.C. Randall, Images in the Margins of Gothic Manuscripts, Berkeley/Los Angeles, 1966. Another example in S. Omer 5 that is probably based on a scene of martyrdom again used out of context is the marginal illustration to the Epistle of James (fol. 267) which shows a mitred figure being thrown from the roof of a church by a knight. This does not refer to the martyrdom of James, who was martyred by the sword, but is probably based on an illustration to the martyrdom of Saint Clement. A comparable example, though not from the same workshop, is in the Martyrology Valenciennes 838, fol. 124. For bible iconongraphy see now Stones, Gothic Manuscripts, Part II, vol. 2, pp. 35–128. 32 There is an added dimension in the border illustration to S. Omer 5; while much of it, as we have seen, is derived from non-biblical sources and seems irrelevant to the text on the page, there are also many cases where the border illustration of an Old Testament book contains an illustration from a New Testament scene, thus illustrating medieval ideas on the typological relationship between the Old and New Testaments. This arrangement is, however, most unusual. For yet another type of border illustration, but one that does not apply to the present material, see L.M.C. Randall, “Exempla as a source of gothic marginal illumination”, Art Bulletin, 57, 1957, pp. 97–107. On marginalia see now M. Camille, Image on the Edge: The Margins of Medieval Art, Cambridge MA, 1992, and J. Wirth, ed., Les Marges à drôleries des manuscrits gothiques, Geneva, 2008. 33 Facsimile ed. Y. Rockseth, Paris, 1935–9. Only one of the illuminators who worked on this MS uses the borders to present scenes. Few stylistic parallels have been suggested for his work. Closely related to this hand are the first two illuminated pages of the Bible Paris, Mazarine 34, made for Anthony Bek, bishop of Durham under Edward I (identification of Bek’s coat of arms by Dr. Adelaide Bennett). Although the illumination of this as yet

SACRED AND PROFANE ART

55

both sacred and profane in content; and while the historiated initials of each reflect the content of the poem they illustrate, the illuminators matched secular themes in the border with sacred themes in the historiated initial and the poem (figs. 18 and 19). At the same time a parallel development is witnessed in the outstanding liturgical books of the period like the Verdun Breviary34 (fig. 20) in which the borders contain scenes that are essentially secular in nature. The interpretation of this mingling of levels in late thirteenth-century French illumination is not easy. Even when we know who the patrons of illuminated MSS were, as in the case of the Verdun Breviary, we lack evidence to show the extent to which a patron might dictate in detail the designs and motifs used to decorate his book. Similarly, our knowledge of workshop structure and administration is insufficient to show whether an illuminated page is the result of careful planning by the head of the shop or whether the whim of an individual may not also have played a part in it. Liturgical and secular traditions were never very far apart as far as their artistic expression is concerned; the adaptation of the liturgical to suit the unpublished MS is English, I do not think it is necessary to assume that this hand of the Chansonnier was also English. Another MS related by less direct stylistic connection to the Chansonnier is a bible in the University Library at Santa Barbara; see L. Ayres, “The Miniatures of the Santa Barbara Bible: A Preliminary Report”, Soundings, 3, 1971, pp. 5–21. See now Stones, Gothic Manuscripts, Part I, vol. 2, Cat. no. 24. 34 Verdun 107 and BL Yates Thompson 8. The workshop of the Bar Breviary has not yet (in 1977) received the detailed study deserved by its exceptional quality. There are two main styles, both of which appear in Verdun 107, where the work of both is incomplete. Related to the second hand of Verdun 107 are YT 8; the Metz Pontifical, Cambridge, Fitzwilliam 298 (also incomplete), facsimile ed. E.S. Dewick, Roxburgh Club, 1902; a Book of Hours, Paris, Arsenal 288, to which my attention was kindly drawn by F. Avril. There is nothing that is of comparable quality to the first hand of Verdun 107 and in the same style (so I thought in 1977, see now the Prague part of Renaud’s Pontifical, Národní knihhovna MS XXIII C 120); hitherto unnoticed lesser MSS by an inferior hand of the same workshop are a Book of Hours, BnF lat 1361, and fragments of a Legenda Aurea, London, Victoria and Albert Museum, MS 706–16. I believe that the well-known Tournoi de Chauvency MS, Oxford, Bodleian Library, Douce 308, text ed. M. Delbouille, Liège/Paris, 1931, was also a product of this atelier. See now S.K. Davenport, Manuscripts Illuminated for Renaud de Bar Bishop of Metz (1303–1316), Ph.D. thesis, University of London, 1984 and ead., 2017; A. Stones, “Le contexte artistique du Tournoi de Chauvency,” in Lettres, musique et société en Lorraine médiévale: Autour du Tournoi de Chauvency (Ms. Oxford Bodleian Douce 308), ed. M. Chazan and N.F. Regalado (Publications romanes et françaises CCLV), Geneva, 2012, 151–204 and ead., “Les Manuscrits de Renaud de Bar,” in Écrire et peindre dans le diocèse de Verdun au Moyen Age, ed. A.-O. Poilpré with M. Besseyre, Turnhout, 2014, pp. 269–310.

56

secular is something that characterises twelfth and thirteenth-century literary as well as artistic works. While there are undoubtedly changes in the relative value placed upon secular as against liturgical books during the course of the thirteenth century, which result in the equalising of the aesthetic merit of both, what emerges most strikingly is the detachment of motif from context, whether it be sacred or profane, that characterises the last decade of the thirteenth century. Perhaps what we are really witnessing here is, after all, the emergence of the independent artistic personality which is free to choose motifs at will — art for art’s sake, rather than art subordinated either to the sacred or to the secular?

III Arthurian Art Since Loomis

T

oday, as in 1938, the starting-point for any inquiry into Arthurian art must still be Arthurian Legends in Medieval Art by Roger Sherman Loomis and Laura Hibberd Loomis.1 This book ranks, after fifty years, as an unsurpassed achievement that cuts across the disciplines of art and literature and contributes to each; its collection of Arthurian images is a corpus of visual data that remains unequalled in print. My purpose here is to present an overview — without claiming completeness, as I deliberately omit the fascinating dimensions of Arthurian archaeology and post-medieval art, to say nothing of over-simplifying what remains in which I assess what the Loomises did, see what fresh material has emerged and what new approaches have been brought to bear, and suggest some directions for further work on Arthurian art. Arthurian Legends begins with a brief outline of the particular problems of method that the study of the illustrations of literary or secular themes presents, and to which the image of the Tower of Memory with its twin doors, sight and hearing (veir e oir) depicted in the BnF fr 412 copy of Richard de Fournival’s Bestiaire d’amours provides an introduction (fig. 1).2 First published in Arturus Rex, volumen II (Acta Conventus Lovanensis 1987), ed. W. Van Hoecke, G. Tournoy, W. Verbeke, Leuven, 1991, pp. 21–78. In 2014 I gave an update entitled “Arthurian Art: the Past, the Present, and the Future,” at the International Arthurian Congress in Bucharest, reprinted as part of the present essays. Updates below are therefore kept to the minimum. 1 For full citation, see Bibliography; hereafter cited as Arthurian Legends. 2 P. 3, frontispiece, figure of BnF fr 412, fol. 228. The text accompanying the image in the manuscript reads, “Ceste memoire si a .ij. portes ueoir et oir et a chascune de ces .ij. portes si a .i. chemin par ou on i puet aler. painture et parole. painture si ert a oel et parole a oreille...” cf. Segre’s (1957) edition, p. 4 (based on MS ‘I’, Paris, Bibl. Ste-Geneviève 2200). On the image in BnF fr 412, see Nordenfalk (1976) 22–3 and fig. 9. Earlier mentions of the

58

The two roads leading to the doors, says Richard, are painture and parole.3 This image, in written and visual form, together with the aural sense to which its accompanying text refers, can be said to have taken on still more significance in the fifty years since the Loomises drew attention to it: questions of literacy and illiteracy, orality, memory, and of image as sign are among the issues of serious debate that have most strongly come to the fore in the past decade and are major concerns today for medievalists in both literature and art, and in Arthurian literature and Arthurian art.4 The theoretical issues for the Loomises took more traditional directions, governed primarily by organizational questions: should Arthurian images be looked at according to text, chronology, country? As they put it, “The nature of the subject has made it impossible to find a single principle of classification. The art student would naturally like to see the material discussed and classified according to the materials, forms, and schools of art. The student of letters would with equal justice prefer to find the material grouped according to the branches of the legend. A compromise has been adopted...”.5 The first part the book is iconographical rather than stylistic in approach: Part I deals with wall painting, ivories, sculpture and textiles, and arranges with material by subject, while Part II covers (much more selectively) the manuscripts and incunabula, and treats them geographically and chronologically, confronting (to a degree) the question of their stylistic relationships to other contemporary art. Both parts demonstrate a concern for the historical context of the works discussed and include a useful summary of the documentary evidence in relation to what has survived as well as in manuscript focussed on the name of the illuminator, Henri, whose mention is so unusual for the thirteenth century; see Millar (1959), pp. 9–10. Other MSS that can now be attributed to Henri are Paris, BnF naf 16251, Brussels, BR 582–9, BR 1787, BR 2512, The Hague, K.B. 74 G 31 and Vienna, Ö.N.B. s.n 12771, as I shall demonstrate more fully elsewhere (see now Stones 1997). For recent discussions of Richard’s words see Kolve (1984: 25–6) and Huot (1987: 138–144, 164–167). 3 Segre (1957: CLVI) thinks Richard’s archetype was certainly illustrated. 4 For the transition from oral to written transmission see the seminal studies by Clanchy (1979), Bäuml (1980), Scholz (1980), Curschmann (1984), and Camille (1985). The question of the rôle of images as mnemonic devices, to which Richard’s words point much more directly, has been little exploited since Yates (1966); see. however, Zinn’s (1974) analysis of the contribution of Hugh of St. Victor, and Evans (1980) where the issue is raised in relation to geometrical figures, and, more recently, Kolve (1984, ch. I), Friedman (1985), and now Carruthers (1990,1998). On the related issue of the uses of images for affective purposes see Berliner (1955), Suckale (1977) and Belting (1981). 5 Arthurian Legends, p. 7.

ARTHURIAN ART SINCE LOOMIS

59

relation to what is no longer extant. The Loomises shrewd recognition of the necessity of multiple approaches to a complex subject is I think one of the main reasons why there is still so much to learn from the book. Arthurian Legends united the two disciplines of art and literature at a time when the primary foci of each field led in somewhat separate directions, although in many ways the disciplines operated from a similar structural base, both formed in traditions of logical positivism.6 Depending on their location and training, art historians were focussing, by the 1930s, on the evolutionary processs of stylistic development, concomitant with which was an overwhelming concern with issues of quality, based on the ideal of classical art, and with a strong emphasis on the art of the Renaissance in Italy; another major focus was the search for meaning and the interpretation of symbolism, in which the painting of Northern Europe in the Renaissance provided an important corpus of material.7 Arthurian art fitted such patterns of art-historical scholarship only tangentially. Judged by qualitative criteria, much of Arthurian art simply did not measure up, and was, by omission, deemed unworthy of serious study on aesthetic grounds. The literary historian tended simply to ignore the illustrations, their presence rarely described, if mentioned at all, in text editions, and certainly not noticed with the kind of consistency that would have prepared the way for detailed investigation of the illustrations in individual manuscripts or decorative programmes, let alone for comparative study, or for dealing with subjects that did not fit a demonstrable textual situation. An instance of the qualitative difficulty facing the historian of Arthurian art, then as now, is that of the illustrated Chrétien. The illuminated manuscripts are substantially later than the composition of the texts, they are few in number and, at best, average in quality.8 The final image of Perceval 6 For a succinct outline of the origins of each, see Jauss (1970/1982), particularly ‘Literaturgeschichte als Provokation der Literaturwissenschaft’ and ‘Geschichte der Kunst und der Historie’. 7 Burckhardt’s Kultur der Renaissance in Italien, published in I860, had been translated into English in 1898; Wollflin’s Kunstgeschichtliche Grundbegriffe of 1898 was translated into English as Principles of Art History in 1932. See also Panofsky (1939), (1953). 8 I am grateful to Terry Nixon for circulating his list of 43 surviving manuscripts and fragments, which will be published and illustrated in the forthcoming volume The Manuscripts of Chrétien de Troyes, edited by K. Busby, T. Nixon, A. Stones, L. Walters (2 vols., Amsterdam and Athens, GA, 1993). A 44th fragment was published in Arthurus Rex (1987), n°4.23, p. 251; and a 45th fragment was published by L. Jefferson, “A New Fragment of the First

60

1. Paris, BnF fr 412, Richard de Fournival, Bestiaire d’amour, fol. 228, the Tower of Memory (© BnF)

2. Paris, BnF fr 12576, Chrétien de Troyes, Perceval and Continuations, fol. 261, Grail scene (© BnF)

ARTHURIAN ART SINCE LOOMIS

61

before the Grail in Paris, BnF fr 12576, f. 261 (fig. 2) exemplifies some of the problems.9 Judged qualitatively, against a background of classical antiquity, or even against the sophistication of the high quality Gothic painting with which it is contemporary, this image is open to criticism or even disparagement. The repetitious faces with their ineptly outlined features look almost like a parody of the angels, divine messenger and chosen hero they are supposed to depict; the Loomises described them as ‘grotesquely ugly’. Similarly the poor painterly technique and monotonous maroon, blue and grey palette, and the clumsily outlined border motifs all weigh against the ranking of this at an artistic level comparable to the greater literary merit of the text it illustrates (if, admittedly, it is placed at the end of the Continuations, and so does not actually illustrate Chrétien’s words) although the gold background shows that good money was paid for it. The general lack of artistic accomplishment extends also to the iconography in terms of the relation of picture to text. The miniature is placed at the end of the Third Continuation, composed by Manessier for countess Jeanne de Flandre (1205–1244).10 The text ends with Perceval’s death and a description of his tomb, complete with inscription, yet the miniature shows Perceval kneeling before the Grail, shown as a partially veiled chalice held by a woman who receives it (or hands it back) to an angel who leans out from a cloud above; another angel swings a thurible while a man on the other side of Perceval holds a taper-like object which ought to be the Holy Lance.11 The closest textual reference to the supporting figures is l. 42489, where the Grail is described as tot descouvert, not partially veiled as in the image, nor is the second woman, holding the tailloir d’argent (I. 42498–9), included in the picture. This appearance of Grail, tailloir and lance occurs in the presence of Arthur and his companions, who are likewise omitted, suggesting that, if a particular textual association is sought, it refers, rather, to Perceval receiving food and drink from the Grail during his final ten years at the hermitage (ll. 42584–ff.). Continuation of the Perceval (London, PRO, E122/100/13B),” Arthurian Literature, 15, ed. J.P. Carley and F. Riddy, 1997, pp. 55–76. 9 Arthurian Legends, p. 90, fig. 204. 10 Roach (1983), pp. 343–4, attributes the commissioning of the text to the period of the capitivity of Jeanne’s husband, Ferrand of Portugal, between 1214 and 1227. 11 Two parallel situations are the endings of Le roman de Judas Machabé and of Mort Artu, where the tombs of the central characters Judas Maccabeus and King Arthur, respectively, are described; while the Judas manuscripts illustrate the tomb, those of Mort Artu vary

62

Or else it is a generalized representation that is not aimed at the specific elucidation of, or commentary on, an ajacent text passage, but rather sums up what this text is fundamentally about, in the way that the single image of the tryst of Tristan and Isolde beneath the tree so popular on portable objects like the piece of shoe leather from the drains of Mechelen,12 discovered in 1972, is selected to stand for the whole story. The format of the Perceval miniature, its single scene larger then any other episode in the volume, supports the idea that the primary function of this image differs in emphasis from the others in the book.13 This example raises issues of fundamental importance that have still to be fully examined in relation to Chrétien texts and the Continuations, and more generally in Arthurian art. One is the purpose and function of the isolated or single image as sign or summary of the whole story. Which image is selected? The same one for each story? How is the meaning of the image determined when it is used without its supporting text? How was its meaning perceived? Closely related are questions about the links between single images in a narrative sequence and the corpus of narrative illustrations of a particular text as a whole. To what extent do the picture-cycles travel together? What is the overall pattern of inclusion and omission for each text’s illustrations? Do final illustrations, or initial ones, receive special treatment? What can be said about the comparative examination of each discrete image and each picture cycle as they are invented, repeated and changed in different regions and periods, and for different patrons? In what follows I attempt an overview of some of these issues as they have been addressed, or remain open for investigation, in Arthurian art. considerably in what is included as the final illustration, and never show the tomb itself. See Stones (1988). 12 Mechelen, Mechelse Vereniging voor Archeologie. See Arthurus Rex, (1987), vol. I, Cat. no. 2.2.6, pl. III.2. 13 Busby (1988), p. 44, considers that the only miniature in the Perceval section of BnF fr 12576 (T) that bears any relation to the text’s narrative is the opening one; his examination (published without photographs) excludes the Continuations, on which a separate essay is to appear in Busby, Nixon, Stones and Walters, referred to above in note 8. The highly subjective analysis of Mentre (1986) suggests there may be more textual relevance in the illustrations of Montpellier H 249 (M) and the prose Estoire, Rennes 255; but what is still needed is a comparative examination of these illustrations in relation to the rest of the illustrative tradition on the one hand and workshop practice on the other. For the latter, in relation to Rennes 255, see Stones (1977), noting its date of c. 1220, not 14th c. as in Mentré. For prose Lancelot manuscripts see now http://www.lancelot-project.pitt.edu.

ARTHURIAN ART SINCE LOOMIS

63

Most scholarly Arthurian activity since Arthurian Legends has been focussed on single questions, be they of one particular decorative program or medium, or on the illustrative tradition of one text or one episode. General and theoretical questions have usually been touched upon partially or indirectly. The only attempt at an all-embracing re-evaluation of Arthurian art is that of the exhibition Arthurus Rex held at the Museum in Leuven in 1987, in conjunction with the meeting of the International Arthurian Society.14 It presented a combination of photographs and artefacts of all kinds, from the archaeological sites to recent discoveries like the Mechelen shoe,15 and was a remarkable visual survey of the impact of Arthur in archaeology, art and literature from the Arthur of history to the present. It is hoped that the exhibition wili pave the way for a corpus of Arthurian images, the absence of which continues to be a serious drawback for comparative research. The exhibition served as a useful reminder that a considerable amount of new material has been discovered since Arthurian Legends. I review some of it, by country of origin, adding comments on some of the methodological approaches that have been brought to bear and crossing the boundaries of country and medium where appropriate, since any primary focus of inquiry, as the Loomises showed, will inevitably overlap with others. The enormous proliferation of publications dealing directly with Arthurian art, or important as providing context for it, present problems of organization for the writer of today that are as daunting as those faced by the Loomises; I take refuge in a summary bibliography included at the end. Germanic regions In monumental painting the extensive cycle of wall-paintings at Schloss Rodenegg16 discovered in 1972, made probably in the early years of the thirteenth-century in close association with the writing of Hartmann’s Iwein, is a fortunate witness to what must once have been one of the most Arturus Rex (1987). See also, now, La Légende du Roi Artur (exhibition catalogue), ed. T. Delcourt, Paris, 2009. 15 See note 12 above. 16 Rasmo (1973), Szklenar (1975), Ott and Walliczek (1979), Kühebacher (1982), Masser (1983), Bonnet (1986). See now J.A. Rushing, Images of Advenrue. Ywain in the Visual Arts, Philadelphia, 1995; id., “The Medieval German Pictorial Evidence,” in The Arthur of the Germans. The Arthurian Legend in Medieval German and Dutch Literature, ed. W.H. Jackson and S.A. Ranawake, Cardiff, 2000, pp. 257–79. 14

64

popular contexts for Arthurian art, that of the private home. The Rodenegg paintings also form a particularly interesting contrast with the slow development of traditions of illustration in manuscripts, a phenomenon still more pronounced in France than in Germany, and one which has still not been adequately accounted for. Despite the absence, and probably the loss, of an illustrated Iwein manuscript of comarable date, the Berlin copy of Heinrich von Veldecke’s Eneit bears witness to the survival of fully illustrated luxury books in Germany, whose full-page narrative miniatures remain almost without parallel elsewhere.17 If the Eneit facsimile was published in the 1930s, the past two decades have witnessed an enormous surge of interest in high-quality colour facsimiles of the important German literary manuscripts including the Parzival and Tristan in Munich, BSB Cgm 19 and 51 respectively. For manuscript material in German-language areas an indispensable resource is constituted by the files and photographs of the Kommission für deutsche Literatur des Mittelalters of the Bayerische Akademie der Wissenschaften, Munich, begun by Hella Frühmorgen-Voss. Her pioneering work using this data made a major impact before her untimely death in the early 1970s; under her successor, Norbert Ott, this material has now begun to appear in print.19 For the Low Countries J. Deschamps’ (1972) catalogue of manuscripts in Middle Dutch, which includes the Arthurian texts and Arthurian illustrations, and Gerritsen’s (1981) work on Jacob van Maerlant elucidate the context of Arthurian interest in those regions. In other media, the Tristan tapestries have received notable attention, both in monograph form (Fouquet 1971) and in the context of cross-media treatment of the subject-matter (Ott 1975, 1982). Britain The Arthurian interests of English monarchs are well known but, despite the Chertsey tiles, no painting directly associated with them seems to have survived, nor are there documentary references to Arthurian topics among

Boeckler (1939). Wolfram (1970); Becker (1977). 19 Frühmorgen-Voss (1969), ead. (1973), ead. (1975), ead. and N. Ott (1986). See now S.C. Van D’Elden, Tristan and Isolde: Medieval Illustrations of the Verse Romances, Turnhout, 2017. 17 18

ARTHURIAN ART SINCE LOOMIS

65

what is known of their artistic commissions.20 The Painted Chamber at Westminster, on which work is documented under both Henry III and Edward I, lacks Arthurian subjects, as do Henry’s other documented decorative schemes21. The only hint of a lost Arthurian decorative program is at Dover Castle, whose great hall, built c. 1240 by Henry III, came to be known as Arthur’s Hall, and the name Guinevere became associated with a smaller hall there — whether on account of painted portraits or programmes, tiles, or tapestries, is uncertain.22 In manuscripts, however, three recent discoveries, unknown to the Loomises, help fill a perplexing gap in the survival of Arthurian art in thirteenth-century England. The earliest of the newly discovered manuscripts is the series of Tristan drawings included (without an accompanying text) in the didactic 20 The Painted Chamber in the Palace of Westminster, on which work is documented under Henry III and Edward I, included scenes from the Life of St. Edward, a Maccabees cycle, and Virtues and Vices; see Binski (1985), reviewed Stones (1987). Henry also had Solomon and Marculph images in his ‘bassa camera’ (Tristram 1950: 575, Moralejo 1981, Binski 1985: 43). For the Chertsey tiles see Arthurian Legends, pp. 44–8 and Eames (1976: 7–9, 1980), with recent bibliography. For more on English Arthurian illustrated manuscripts see my “Fortunes of Arthur” below. 21 Tristram (1950), Borenius (1943). 22 Incomplete references to Arthur’s Hall and Guinevere’s Chamber are in Lethaby 1917: 139; Borenius 1943; 46; Morgan 1982: 145. W Dugdale (1823, IVn p. 533) cites, under the heading Historia Fundationis Ecclesie S. Martini Dover. Num.I: Du chatelle De Doure. coment il feust commence et establie, et de chanoignes que illoeques estoient jadis. [Brevia regis de anno 14 Edw. II]. ‘... L’an de Grace quarte centz seissante neuisne, regn in Bretaigne Arthour le Glorious; cesti amenda le dit chastiel e plusieurs choses, et fist la sale, que ore est appelle Artoursale; et la chambre sa femme est appellee Guaonebour, id est Thalamus Guanguare... This ascribes the rooms to the historical Arthur himself; T. Dugdale [1854–60], IV, 654. mentions the two rooms in the context of his description of work on the keep done by Godwin Earl of Kent, Arthur’s hall on the east side of the pre-Conquest keep and Arthur’s smaller hall or Guaonobour’s bedchamber in the last tower (presumably of the curtain wall rather than the keep itself ) next to the palace gate, the latter made into a magazine under Henry VIII. Allen Brown (1966; 4–6) attributes the present keep and inner curtain wall with its mural towers to the reign of Henry II; he identifies (p. 17) Arthur’s Hall with the great hall built in the southeast corner of the inner bailey by Henry III c. 1240. It was apparently known as Arthur’s Hall from the fourteenth century, and is labelled ‘Aula Arthuri Maior’ on John Bereblock’s sixteenth-century view (Allen Brown pl. VI), which also shows a smaller hall, ’Aula Arthuri Minor’, on the other side of the bailey. This is presumably the ‘Guinevere’s chamber’ of W. and T. Dugdale, unless the latter was located in the now destroyed Arthur’s Gate that adjoined the south curtain wall at the end of the walled passage running east from the south curtain wall to Penchester Tower on the outer curtain (see Allen Brown’s plan). Was Arthur’s Gate contemporary with the corresponding passage on the west side, which is attributed (ibid., 7) to the reign of King John, or was this also part of Henry III’s works?

66

3. London, BL Royal 20 C. VI, Mort Artu, fol. 150, Guinevere led to the fire (© British Library)

4. Paris, BnF fr. 123, Mort Artu, fol. 229, King and Queen with scribe (© BnF)

ARTHURIAN ART SINCE LOOMIS

67

compendium BL Add 11619.23 It was probably made c. 1250 and, at its best (more than one illustrator seems to have been at work), is very close stylistically, I suggest, to Geoffrey of Monmouth’s Prophecia Merlini, BL Cott. Claud. B. VII. The latter is illustrated with a single miniature of Merlin prophecying to King Vortigern and has been attributed to the Court School of Westminster c. 1250.24 What is exciting about the newly discovered drawings is that they constitute a sequence of full-page narratives, just the kind of illustration that the Berlin Eneit represents (although the BL manuscript is not as fully painted, and lacks the inscriptions, heraldry, and accompanying text that make the Berlin manuscript so unique). To judge by its stylistic relations with the Prophecia Merlini attributed to London and most likely commissioned in royal circles, perhaps the Tristan may also have been made for a member of the royal court.25 It is possible that the Agravain and Mort Artu, BL Roy 20 C.VI (fig. 3) might also have court associations, as the illustrations are quite close stylistically to the first hand (fol. 1) in the Douce Apocalypse, begun for Edward as Prince of Wales.26 Another Vulgate Cycle volume unknown to the Loomises, with illustrations by an English painter, is BnF fr 123 (fig. 4). It contains Queste and Mort Artu and was written by an English scribe for the Artesian nobility, most probably in celebration of the marriage in 1275 of Blanche d’Artois and Edmund Crouchback, earl of Lancaster.27 Roy 20 C.VI is sparsely illustrated, with only two miniatures, one for the opening of each branch of the cycle, while BnF fr 123 has a series of historiated initials for each text. Since these two manuscripts each include a copy of the Mort Artu, I use the comparison of their images as an example of the kind of investigation of comparative iconography that still remains in large part an unexploited Hunt (1987). Hunt omits mention of the quire structure and it is therefore not clear whether the drawings, on ff. 6r–9v, are an integral part of the compendium. As the binding is not contemporary, but is attributed to the fourteenth-century, it is quite likely the miscellany was put together then from disparate parts and that the textual component tells little about the origin of the Tristan drawings. 24 Arthurian Legends, p. 138, fig. 384. See also Morgan (1982), no. 94, fig. 298. 25 Nothing justifies Hunt’s claim (1987: 51) that the volume is a monastic book: secular canons, mendicants, or clercs in minor orders would be just as appropriate as candidates for the ownership of its didactic texts, while there is also the possibility that the Tristan pictures were not an original component. 26 Klein (1983); Bennett (1985). 27 Avril and Stirnemann (1987), cat.no.152, pl.I, J, LV1II, LIX. 23

68

aspect of Arthurian art.28 The opening initial of Mort Artu in BnF fr 123 follows the opening words of the text in depicting Henry II and Eleanor of Aquitaine commissioning Walter Map to write the text. It is a favourite subject for the beginning of Mort Artu manuscripts and there are many parallels, both earlier and later. The unusual feature is that Eleanor is included with Henry, and the depiction of the royal couple may be an allusion to the noble patron and patroness in honour of whose marriage the book was probably made. The illustration selected for Roy 20 C.VI, on the other hand, shows Guinevere’s trial by fire. This a subject that occurs in most of the fully illustrated manuscripts of Mort Artu (those that include more than a single miniature) but at its place in the narrative sequence, about halfway through the text. As the opening miniature, and as the only one in the manuscript, the choice is exceptional. The patron — Edward or a member of his court? — must surely have had reasons for selecting this over the other more commonly selected opening subjects: Map writing for Henry, Boort’s return to court, Arthur and his men, Lancelot receiving a sleeve, or Arthur’s death. He or she must have meant to underline — for whatever reason — the rôle of Guinevere’s adulterous relationship in bringing about the downfall of Arthur’s kingdom.29 These manuscripts show that English interest in Arthurian material was not limited to the chronicles illustrated in Loomis — fascinating though they are on account of their unusual iconography. In addition to the Vortigern image mentioned above, Langtoft’s Chronicle, to 1307, has the only image of King Arthur bearing the arms of the Virgin Mary, while the Egerton Brut, copied in 1328, includes some 188 miniatures and still awaits a detailed study.30 Yet the misericords, though they are still later in date, indicate that pictorial imagery derived from other Arthurian texts like Ywain For comparative work on the illustrations in Queste and Mort Artu see Stones (1970/1971, 1988). 29 Interesting in this context is Bloch’s (1974) interpretation of the political significance of the trial and the events that lead up to and follow it. 30 Both discussed and illustrated in Arthurian Legends, p. 138, figs. 385–6, 387–8, but, like the Auchinleck MS (Pearsall and Cunningham 1977; Shonk 1984) without mention (for reasons of quality?) in Sandler (1986). For ideas on the relationship between language (Latin vs. Anglo-Norman) and mode (relative quality) of illustration in early fourteenthcentury England see Camille (1988). Whether Camille’s correlation between language and mode holds true generally in this period is debatable; in France the existence of many more liturgical books (in Latin) painted in ‘vernacular mode’ suggests that the stylistic situation is governed by a more complex web of additional factors. See my “The Egerton Brut” below. 28

ARTHURIAN ART SINCE LOOMIS

69

also circulated in England and that Ywain’s horse, trapped by the portcullis, seems to have become by the late fourteenth century a motif that travelled independently of the rest of the story, like the Tryst beneath the Tree from the Tristan legend.31 Italy, Spain and the image as sign Whether these isolated images were understood as signs or symbols of the whole, or as amusing incidents without regard to their significance in their respective stories, is a question whose answer still remains elusive. The great difficulty of interpretation that such isolated subjects represent, whether Arthurian or otherwise — Alexander topoi are equally common — have been underlined in the most recent discussion centering on the famous mosaic of the cathedral of Otranto, dated 1163–5, in which figure both the historicolegendary heroes Arthur and Alexander the Great.32 Were these two heroes included as antitheses of each other, one as a positive force, the other negative? The meaning and rationale of the mosaic have long baffled iconographers. Haug’s (1975, 1977) interpretation focusses on an understanding of Arthur in relation to a circumspect analysis of the significance of the selection and formal grouping of the perplexing elements of the pictorial composition of the mosaic as a whole. Wierschin’s (1983) approach sees the mosaic as a direct comment on the contemporary political situation of southern Italy, with Arthur ‘as’ the emperor Frederick I and Alexander ‘as’ pope Alexander III. The arguments are not necessarily mutually exclusive and both depend in part on the structuralist approach elaborated by Bialostocki (1965/1981), Pickering (1966/1970) and Schapiro (1963/1980, 1973); central to the issue is whether, and to what degree, meaning as well as form is transferred when an image is copied from one context to another, and by whom the transferred associations were understood. Elsewhere in Italy, the Modena archivolt, noted by the Loomises, represents another familiar and puzzling example of the incursion of Arthurian Is the absence of illustrations in Rushing’s (1986) discussion meant to suggest that these motifs circulated as ideas rather than visual images? Such a proposition might satisfy the literary historian, but for the artist a more fitting analogy would be the reproduction of the photographs without the words. For Tristan, the Tryst was not the only image that stood for the rest, see the Santiago example below. 32 Haug (1975, 1977); Wierschin (1983). See also Settis-Frugoni (1968, 1970) and V.M Schmidt (1988: 12–15, 60. 185). 31

70

iconography into the decorative program of a major cathedral: the reasons for its inclusion and the way in which it was interpreted are still far from clear.33 Recent research has, however, provided a more general context for the interpretation of this kind of ‘secular’ topic in an ecclesiastical context where a moralistic interpretation is likely. The Ganymede capital at Vézelay is one such example, where the rape of Ganymede has been seen as a visual warning against the perils of homosexuality,34 or the Marculph image on the portal of Orense, part of a Solomon and Sheba group, and consequently dependent on literary texts of the story of Solomon and Marcuph, but also a topos of fool and rusticus with overtones of luxuria and consequently moralizing in intent.35 Yet it is only in the case of the woman holding the skull who accompanies the Temptations of Christ on the left tympanum of the Puerta de las Platerias at the Cathedral of Santiago de Compostela (fig. 5) that we have any direct evidence of how the contemporary viewer interpreted such sculpture. The Pilgrim’s Guide, based on first-hand observations made in the 1130s, describes the sculpture in its present location and offers moralizing comments.36 Was the little-known relief at Cambrai (fig. 6), showing the Arthurian Legends, pp. 32–36, figs. 5–8. For thought-provoking (if highly questionable) interpretations of visual aspects of epic themes see Nichols (1983, reviewed by Sears 1988). See also H. Nickel, “About Palug’s Cat and the Mosaic of Otranto,” Arthurian Interpretations, 3, 1989, pp. 96–105; M. Meuwese, “Het raadsel van koning Arthur; het mozaiek van Otranto,” Kelten 10, 2001, pp. 11–14. 34 Forsyth (1976). 35 Moralejo (1981). 36 ‘Nor should be forgotten the woman who stands next to the Lord’s Temptation, holding between her own hands the stinking head of her lover, cut off by her rightful husband, which she is forced by her husband to kiss twice a day. Oh, what ingenious and admirable justice for an adulterous wife, it should be recounted to everyone!’ (Pilgrims Guide, ch. 9, transl. P. Gerson, A. Shaver-Crandell. A. Stones (1998); Vielliard 1969: 102–3). There has been discussion as to whether or not the Guide’s interpretation should be taken at face value. It contains a few inaccuracies: the woman is not standing (Latin stat) but seated on a faldstool, and she holds a clean skull, not a rotting head (Latin caput fetidum). The woman has been variously interpreted as Eve (Azcarate 1963: 10; Moralejo 1977: 98) or as Mary Magdalene (Gaillard 1929 with references to previous literature), but neither explanation is convincing in terms of the visual parallels available at a comparable dale, nor in terms of possible literary sources. The twelfth-century Lai de l’Ignaure and the late thirteenth-century Châtelain de Coucy focus on a somewhat comparable motif, but that of an eaten heart rather than a kissed skull. Naesgaard (1962: 73) points to a parallel in the story of Lisabetta in Boccaccio, where a lover’s head is preserved in a jar, and for which Sicilian sources have been suggested. John Williams in an unpublished lecture of 1976 (see now J. Williams, “La Mujer del cráneo y la simbología románica,” Quintana, 2, 2003, pp. 13–27) cited a number of other literary 33

ARTHURIAN ART SINCE LOOMIS

5. Santiago de Compostela Cathedral, Puerta de las Platerías (south transept), left tympanum, scenes of temptation (photo: author)

6. Cambrai, Musée municipal, tympanum from Saint-Géry-auMont-des-boeufs, Cambrai, death of Pyramus and Thisbe (photo: author)

71

72

Ovidian theme of the death of Pyramus and Thisbe, similarly perceived as a topos of misfortune?37 A further recently identified Arthurian theme in the context of cathedral sculpture, is again at Santiago de Compostela — though not described in the Pilgrim’s Guide — on a column of the north portal (fig. 7).38 A series of combat scenes are shown, which Moralejo (1985b) identifies, on the basis of their accompanying details — a serpent-monster, a female figure attending a wounded male figure, a warrior fending off birds who peck at a dead horse, finally a warrior lying in a boat — as scenes from Tristan, but not the Tryst beneath the Tree, which seems to have been more popular in later periods and other regions. The selection of an Arthurian subject at Santiago is particularly interesting in relation to the relative paucity of Arthurian illuminated manuscripts from the Iberian peninsula. The only one in Loomis, the Tristan of 1278, BnF fr 750, has recently been reattributed to parallels: Paul the Deacon’s eight-century account of the sixth-century Lombard King Alboin, killed by his wife’s lover because he had vengefully made his wife use the skull of her father as a drinking-cup (MGH ed. Waitz); the fourteenth-century Italian writer Giovanni Fiorentino’s account of an adulterous wife forced by her husband to share a room with the corpse of her lover; and Marguerite de Navarre’s Heptameron, where the tale for the fourth day relates how an adulterous wife is forced by her husband to drink from her lover’s skull and keep his bones in a chest in her room. At Santiago the juxtaposition of the adultery punishment with the Temptations of Christ would certainly suggest that the motif was supposed to show what the Guide’s author says it does (although the link is not made explicit there), and that the aim was a didactic and moralizing one. Moralejo (in press: see now Lanfranco e Wiligelmo. Il Duomo di Modena (Quando le cattedrali erano bianche. Mostre sul Duomi di Modena dopo il restauro, Modena, 1984) accepts this view and notes (n. 43) a similar subject on a capital at Santa Marta de Tera, observed by Gomez-Moreno (1927: p. 185, no. 422). I thank Serafín Moralejo Álvarez and John Williams for helpful discussion of this sculpture. 37 Oursel (1978). no. 28. The episode shown is the discovery of the bodies of the lovers beneath the mulberry bush whose fruit is stained red by their blood. Further examples are at the cathedral of Basel, and in the Roman de la Poire, BnF fr 2186, fol. 7v. which shows the lovers talking through the wall and, below, the dead lovers with the lion. For the purpose and shifting emphasis in the portrayal of exemplary lovers in this manuscript see Huot (1987), pp. 174–90 and Marchello-Nizia (1984), where the manuscript is attributed to c.1275 rather than the 1250s as in Branner (1977: 102–7). For more on the date see now Stones, Gothic Manuscripts, Part I, vol. 2, Cat. no. I–10. 38 Moralejo (1985a: 67–68, 1985b: 419–421). with discussion of the question of textual sources for a somewhat enigmatic treatment of Tristan’s combat. Moralejo further suggests (1988) a consonance of themes between Chretien’s Perceval, its patron, and the sculptural programme of the Portico de la Gloria at Santiago. I am endebted to Serafin Moralejo Álvarez for helpful discussion of the iconography at Santiago and also for kindly permitting me to reproduce one of his drawings of the Tristan subjects from the north portal.

ARTHURIAN ART SINCE LOOMIS

73

Italy, where Arthurian manuscripts in French enjoyed enormous popularity in the thirteenth century and BnF fr 343 and Florence, BN Pal 556 attest to the particular inventiveness of text and picture in Italy in the fourteenth and fifteenth centuries; both manuscripts are strong candidates for facsimile treatment.39 Like the sculpture of Modena, the Tristan of Santiago antedates surviving written versions of the text and is another example of the precedence of imagery over words.40 The Santiago image, like the Arthur and Alexander of the Otranto mosaic, is also an example of the precedence of an iconic image over a narrative sequence.41 Both phenomena now need to be considered in relation to the important shift, under way in the twelfth century, from the culture of orality to that of reading and writing. This shift provides a framework that helps to explain why it is that the textual traditions of Arthurian literature came to be written down so long after the stories were in general circulation, and their manuscripts illustrated still later.42 39 For BnF fr 750 see Arthurian Legends, pp. 92–3, figs. 210–212; Avril and Gousset (1984), no. 194, pl. CXX, CXXI. There is the curious painting on the ceiling of the Sala de Justicia at the Alhambra, which seems to have some Arthurian subjects that were not fully understood (Dodds 1979). For Italy there is no single convenient source of information covering literary manuscripts illustrated and unillustrated. See, however, Avril and Gousset (1984) for the thirteenth-century Italian manuscripts in Paris. Their study pays careful attention to minor decoration in the form of pen-flourished initials as well as to full-colour initials and historiation; Genoa emerges from this study as a particularly important centre for the production of manuscripts with line-drawn illustrations. For BnF fr 343 see Arthurian Legends, pp. 118–120, figs. 328–334 and Avril (1984: no 84); a study by K. Sutton is anticipated (“Milanese Luxury Books: the patronage of Bernabò Visconti,” Apollo, 134, 1991). See now M. Pastoureau and M.-T. Gousset, Lancelot du Lac et la quête du Graal, Arcueil, 2002. For Florence, BN Pal 556 see Gardner (1930: passim): Arthurian Legends, p 121, figs. 337–9; Stones (1988: 71, 100). 40 Other possible early Arthurian sculptures, in addition to the Perros (Brittany) relief in Arthurian Legends, pp. 3, 10–11, fig. 3, are the tomb at Lantien (Cornwall) and the relief of Banagher (Ireland) (de Mandach 1972, 1975, 1978, 1980). Dating evidence for these is best treated with caution. 41 There are, of course, parallels in the art of the ancient and Early Christian worlds for the use of Iconic or ‘abbreviated’ images, adapted or compressed from narrative. See, for instance, the treatment of iconography in the catalogue, Age of Spirituality, ed. Weitzmann (1979). For the rôle of iconic images in the High Middle Ages in relation to affective uses and purposes see Berliner (1955), Suckale (1977), Belting (1981). At Santiago, as at Otranto, what is striking is the apparent precedence of the iconic, whereas at Modena Arthurian subjects are treated in a narrative mode. 42 See Clanchy (1979), Bäuml (1980), Scholz (1980), Curschmann (1984), Camille (1985).

74

Several studies have explored the impact of the change from orality to literacy on the visual imagery of twelfth-century manuscripts in relation to their texts.43 What we now need is a more broadly-based reassessment of the rôle of Arthurian imagery in relation to other artistic manifestations of ‘vernacular’ or ‘secular’ imagery in general in this period of transition. The rôle of the itinerant secular artist is important here, not only to a consideration of sculpture, but also to that of painting and illumination.44 Another underestimated consideration is the impact on books made for private ownership of the requirement, established for all the faithful at the Fourth Lateran Council of 1215, of communication at Easter, preceded by spiritual preparation that included confession.45 One result of this requirement is the general increase that occurs in the course of the thirteenth century in the production and dissemination among the laity of small devotional books, primarily psalters, psalter-hours and hours, with their additional prayers, saints lives, and images. The imagery in these devotional books comes to include, by the end of the thirteenth century, the range of obscenity, erotica, exempla, and literary motifs, that in the twelfth century found its primary expression in monumental sculpture.46 In the thirteenth century non-devotional books with pictures are also greatly on the increase — by-products of what becomes a secular book-and-image oriented culture, towards which the books made for private devotion lead the way.

Camille (ibid.,); Heslop (1986). For the Loomises, a primary explanation for the lime-lag between the composition of the texts and the production of illuminated copies was the predominance of monastic craftsmen in the field of illumination in the twelfth century and the emergence of the lay craftsman in the thirteenth (Arthurian Legends, p. 89). But the careers of important highquality painters like Master Hugo of Bury and the so-called Simon Master, among others, have shown that itinerant craftsmen already played an important rôle in the twelfth century (see particularly the excellent survey in Cahn [1982| ch. VII), and more convincing explanations for the time-lag must be sought elsewhere. 45 For aspects of thirteenth-century spirituality see Leclercq et al. (1961) and Bossy (1981). 46 The marginal illustrations in Randall (1966) are predominantly from books made for private devotion; but buildings continue to provide a context for ‘literary’ or vernacular themes, see Varty (1967) on Reynard the Fox, and Husband (1980) on the Wild Man, Rushing (1985) and others on misericords. 43 44

ARTHURIAN ART SINCE LOOMIS

75

France, and the creative process In France the largest body of Arthurian art is the extensive picture-cycles, developed during the thirteenth century, some including as many as 200 scenes or more, that illustrate the manuscripts, especially the prose romances. Many of these manuscripts were known to the Loomises and included, with some reproductions, in Arthurian Legends. The emergence of this important form of Arthurian art can now be seen as the product of a rather more complex set of circumstances than was thought in 1938. Since then, too, the resources available for reassessing French Arthurian art have expanded in two important ways: through the development of the research materials available as microfilms, fichiers, photographs and slides of the Institut de Recherche et d’Histoire des Textes, its Section romane and Section héraldique in Paris and its Section iconographique at Orléans; and through the publication of lists of manuscripts and text editions. Most notable among the latter are Woledge’s indispensible Bibliographie (1965/1975, 1975), Micha’s lists of manuscripts (1958, 1960, 1963), and the text editions by Frappier (1964), Pauphilet (1965), Curtis (1963, 1976), Kennedy (1980, 1986), Micha (1978–1983) and Ménard (1987–1997). Regrettably the presence or absence of illumination is not always mentioned in these publications, the manuscripts are not reproduced, and their date attributions are rather apt to be cavalier. Models are available, however, and an excellent example is the (1976) edition of the Chanson du Pseudo-Turpin by the late Ronald Walpole, which includes an outline of all the other texts in each manuscript as well as a photograph of each manuscript; Willard (1985) has taken a closer look at the manuscripts of Tristan, and Baumgartner (1987) has addressed the question of opening miniatures and their distinctive treatment, again in relation to the manuscripts of Tristan; Demaules and Marchello-Nizia (1989) have examined the iconography of dreams in Lancelot, Queste and Mort Artu. But the range of visual data that is conveniently available is still severely limited. What the iconographer would like is at least a list, better a description, and ideally a photograph, of every illustration in every manuscript.47 These are essential if one is to assess the question of whether 47 For Christian art there is the Index of Christian Art, housed at Princeton University, with copies of its files cards and photographs at the University of California, Los Angeles, Dumbarton Oaks, Washington D C., the University of Utrecht, and the Vatican Library, now online as the Index of Medieval Art, as well as many iconographical dictionaries, among

76

the illustrations of a particular text occur with regularity at the same textual breaks, whether the same subjects occur and to what extent their treatment is consistent, or varies, across time and space. In the absence of published lists, descriptions, or photographs the time and effort required to collect the basic data has inhibited detailed study in what remains an extremely fruitful area for research. Apart from the Chansonniers, facsimiles of French literary manuscripts, especially those in prose, are almost non-existent — König’s Roman de la Rose (1987) is an exception. At the same time, work on comparative iconography of “literary” texts has tended to be confined, for other countries as well as for France, to manuscripts whose textual tradition is small or whose picture-cycles are limited, so that the total corpus of illustrations becomes manageable. Models of such comparative study have tended to focus on ‘historical’ texts: examples are the studies by Buchthal on the Historia Troiana (1971), which includes manuscripts from Italy as well as France; Folda on the William of Tyre manuscripts from Acre (1976) and the forthcoming studies of the French Alexander in verse and in prose by D.J.A. Ross, which will compliment his (1971) study of Alexander in Germany and The Netherlands. Tables of comparative iconography are still rare, and hard to follow unless substantially illustrated; Segre’s (1957) edition of the Richard de Fournival Bestiaire d’amours is one of very few text editions that attempts to tackle the problem of comparative illustration in tabular form, and Meiss’s (1967, 1968, 1974) studies of French manuscripts in the time of Jean de Berry include a number of comparative lists among the appendices. Tables of the two earliest manuscripts and possible sources for their illustrations are included in Hindman’s (1986) study of Christine de Pizan’s Epistre Othéa, and my work on the illustrations of Queste and Mort Artu (1970/71, 1988) includes a comparative treatment of the iconography. Further models for comparative tabulation are available among studies of biblical iconography.48 Even so, the logistics of tabulating the heavily illustrated prose romances, which exist in large numbers of copies (50 or so of Mort Artu, for instance, and 34 of them illustrated) are daunting. Long rolls of architects’ tracing paper, a convenient portable edition of the text, and the time to see each manuscript are minimum requirements. which the Lexikon der christlichen Ikonographie, ed. E. Kirschbaum, 8 vols., Rome, 1968–76 is still the best. Nothing comparably comprehensive exists for profane art; van Marle (1931) is highly selective. 48 An example treating thirteenth-century Latin bibles from Flanders and Artois is Beer (1972). See now Stones, Gothic Manuscripts, Part II, vol. 2, pp. 61–75.

ARTHURIAN ART SINCE LOOMIS

77

What these comparative studies make clear is that, as with the illustrations of biblical, liturgical and devotional texts, sets or groups of images invented to illustrate a particular text are to a significant degree transmitted together, like their text, from one manuscript to another.49 In other words, the picturecycles tend not to be freshly invented each time a text is copied, although, as I indicated above, and have shown elsewhere, there can be omissions, substitutions, and changes of emphasis in the selection of illustrations that suggest that the process of copying and invention was also a flexible one.50 The transmission of these picture-cycles functions through a system of model and copy that is little different from the transmission of medieval images in general and is conditioned by the essentially conservative nature of the approach taken by the Early Christians and their Byzantine and Western medieval successors to the copying of illustrated bibles, as the fundamental studies of Weitzmann (1959, 1970), Kitzinger (1976), Kessler (1988) and others have shown. The artist is free to invent only to a limited degree. The image can be copied directly from an already existing one, or by way of one or more intermediaries in the form of a model in a copy book; a marginal sketch; a list of written subjects; a written marginal note; notes or symbols for colours, or many combination of the above.51 For Queste and Mort Artu see Stones (1970/71), ch. X, and for Mort Artu, ead., (1988). Such patterns of consistent copying do not always occur: my study of Wace’s St Margaret illustrations (1990), for instance, shows an absence of influence from other illustrated versions of the same story, and a similar independence is characteristic of St Margaret cycles in general. More important to the iconography of St Margaret’s life is which version of the text, with its particular list of tortures, is being illustrated. 50 In the illustrative tradition of Queste and Mort Artu, for instance, there is no correlation between the pictorial recensions and the text recensions (Stones 1970/71, ch. X). Similarly the pattern of how many illustrations were selected for inclusion can vary even between two products of the same workshop. Overall, however, there is a good deal of consistency in the selection and placing of subjects for illustration, suggesting a high degree of organization and coordination between planner, scribe, artist, and perhaps patron, quite apart from the question of continuity between models and copies. For related scribal questions see Kennedy (1970). 51 The most comprehensive survey of model-books is still Scheller (1963, 1995); see also Ross (1952) and for notes and sketches in Arthurian manuscripts Stones (1970/71, 1990). See now Alexander (1992). Particularly interesting marginal notes in ink referring to types of characters or settings shown in the illustration and to heraldic devices and their tinctures occur in M.805 and are transcribed by Cockerell (1907: 94–116) (but many were overpainted). The group of early fourteenth-century Vulgate Cycle MSS Add 10292–4, Roy 14 E.III and olim Amsterdam/Rylands French I/ Douce 215 include many traces of two sets of marginal notes, one for the rubricator, another describing the illustrative contents of the 49

78

7. Santiago de Compostela Cathedral, Puerta Francigena (north transept), column detail (drawing by Serafín Moralejo Álvarez) 8. New York, Pierpont Morgan Library M.805, Lancelot, fol. 67, First Kiss of Lancelot and Guinevere (© Morgan Library)

ARTHURIAN ART SINCE LOOMIS

79

Arthurian manuscripts provide many traces of such intermediaries. There are also instances among their illustrations which show that the choice of subject depended on the copying of a visual model and was unrelated to the words of the accompanying text. Again an interesting question is the extent to which the meaning of the visual model, particularly one lifted from another context, is also intended, a point which leads back to the question of the interpretation of the isolated images raised above. I select two examples from the French prose Lancelot cycle with which to explore some of these points. First the puzzling First Kiss of Lancelot and Guinevere from the Morgan Lancelot, New York, Pierpont Morgan Library, M.805 (fig. 8), which has been addressed in Haussherr’s important study (1975).52 What is immediately striking is the extraordinary triangular structure of the kiss group, caused by the presence of the third figure, Galehot, across whom the couple strain to reach each other; and this group of three figures is matched by another group of three in the other half of the miniature. No comparative study exists for the first kiss image in this context, although the Kiss motif as such has a long history in biblical illustration. If the Kiss of Righteousness and Peace in the mid-ninth-century Stuttgart Psalter, made in Northern France, is generally interpreted as showing the Visitation (fig.9 Psalm 84/85),53 the type reappears in a number of guises, and with a clearer male/female emphasis. Important in the development of the theme are the illustrations to the Commentary on the Song of Songs from Clairvaux, Troyes BM1869, c. 1200 (figs. 10, 11); 54 a further striking example that is almost exactly contemporary with the Morgan Lancelot miniatures and aimed at the illuminator. Both sets occur in the margins and are written in leadpoint; many are completely or partially erased (Stones 1970/71; 1987). 52 The most complete description of the manuscript is Cockerell (1907). For an attribution to Amiens c.1315, see Stones (1970/71: ch. VIII). For the text, see Sommer, III (1910), Kennedy (1980), p. 348. Kennedy (1980:7) incorrectly dates the MS to the fifteenth century, perhaps by confusion with M.807 which is a fifteenth-century manuscript. See now Stones, Gothic Manuscripts, Part I, vol. 2, Cat. no. III–93. 53 Facsimiles, ed. Dewald (1930) and Stuttgarter Bilderpsalter (1968). Both commentaries interpret the figures as Mary and Elizabeth (Dewald p. 76; Stuttgarter Bilderpsalter, p. 118), in conformity with Byzantine traditions, with Mary as the right-hand figure wearing a crown: it is interesting to speculate that the lack of clarity in the treatment of her hair might possibly indicate that this could be otherwise interpreted, although the halo lacks a cross which ought to preclude an interpretation of this figure as Christ. See my Lancelot and Guinevere contributions in these essays. 54 Morel-Payen (1935: figs. 39,41). See also Camille (1989), p. 150.

80

(though from Bohemia rather than Picardie) is the embrance of Christ and the Virgin Mary in the Passionary of Abbess Cunigunda of c. 1316–20 (fig. 12).55 A seated embracing couple that is also contemporary with the Morgan Lancelot is the embrace of Hosea and Gomer from the Bible of Jean Papeleu, made in Paris in 1317 (fig. 13),56 and vernacular illustration could also provide a similar, and earlier, model in the compendium of topoi of lovers mentioned above, the Roman de la Poire (fig. 14).57 What is missing in these examples is a parallel for the intermediary figure so dominant in the Morgan embrace, and for the exaggerated leaning pose of the lovers. Haussherr has suggested some ingenious sources that include borrowing from the Christ/St. John group common in Last Supper imagery, popular in South Germany in the early fourteenth century as a devotional image in sculpture.58 Further visual models might include images of marriage such as those that illustrate Causa xxviii of Gratian’s Decretum which have a priest as intermediary between the couple,59 or possibly even a compression and reserval of upward diagonals to downward ones in a image like the god of love and seated couple in another of the Roman de la Poire images (fig. 15), depending on how far one wishes to press the search for visual sources and related meanings. The importance of the threefold configuration in both halves of the Morgan miniature is attested by its reappearance and transformation in other manuscripts of the same text: in the contemporary BL Add 10293 of c. 1316, Guinevere is the central figure in the kiss group, and it is Lancelot and Galehot who join hands (fig. 16);60 in the fifteenth-century copy purchased by Jean de Berry and repainted half a century later for his great-grandson, Jacques d’Armagnac, BnF fr 117– 120 (fig. 17), Lancelot is the central figure, linked to the more distant Galehot by his diagonally projecting leg which seems to echo the strong diagonals of the Morgan composition.61 My second example is one where the composition seems at first sight not only curiously contrived but downright misleading. The subject is Lancelot’s Facsimile eds. Urbánková and Stejskal (1975). Paris, Ars 5059, see Diamond (1979). 57 See Marchello-Nizia (1984) and Oursel (1986) as in note 37. 58 Haussherr (1975). 59 Melnikas (1975), III, figs. 12, 15, pl. II. 60 For the manuscript and its group see Arthurian Legends, pp.97–8 and Stones (1970/ 71 ch. VI; 1987). 61 Arthurian Legends, pp. 105–7; Meiss (1967), 242, 252, passim. Stones (1988) 95. 55 56

ARTHURIAN ART SINCE LOOMIS

9. Stuttgart, Württembergische Landesbibliothek, fol. 123, Stuttgart Psalter, fol. 100v, Psalm 84/85, Embrace of Righteousness and Peace (© Württembergische Landesbibliothek) 10. Troyes, BM 1869, Gregorius ?, Expositio in canticum canticorum, fol. 176v, Embrace of Christ and Ecclesia (photo: author)

81

82

11. Troyes, BM 1869, Gregorius ?, Expositio in canticum canticorum, fol. 177v, Embrace of Christ and Ecclesia (photo: author) 12. Prague, Národní Knihovna XIV A 17, Passionary of Kunigunde, fol. 16v, Embrace of Christ and Ecclesia (©Národní Knihovna)

13. Paris, Bibl. de l’Arsenal 5059, Bible of Jean Papeleu, fol. 365, Embrace of Hosea and Gomer (© BnF)

ARTHURIAN ART SINCE LOOMIS

14. Paris, BnF fr 2186, Roman de la Poire, fol. 3v, Cligès and Fénice (© BnF)

83

84

15. Paris, BnF fr 2186, Roman de la Poire, fol. 1v, God of Love shoots arrows at lovers (© BnF)

ARTHURIAN ART SINCE LOOMIS

16. London, BL Add 10293, Lancelot, fol. 78, First Kiss of Lancelot and Guinevere (© British Library Board) 17. Paris, BnF fr 118, Lancelot, fol. 219v, First Kiss of Lancelot and Guinevere (© BnF)

85

86

18. Paris, BnF fr 110, Lancelot, fol. 206v, Lancelot visits a hermit (© BnF) 19. Oxford, Bodleian Library, MS Douce 215, Lancelot, fol. 14 Lancelot visits a hermit (© Bodleian Library)

ARTHURIAN ART SINCE LOOMIS

87

visit to a hermit, for which two illustrations are striking. One, in BnF fr 110, f. 206v (fig. 18), illustrates a hermit episode in the Lancelot proper,62 the other, in Oxford, Bodl. Douce 215, fol. 14 (fig. 19), a similar encounter in Queste.63 Both manuscripts show the hermit located in a tree-cell, with Lancelot on horseback or standing on the ground below. Nothing in either text justifies placing the cell in the tree, nor does such a tree-cell otherwise figure in the extensive range of images showing the numerous visits to hermits elsewhere in the illustrative of the Vulgate Cycle. Did one artist simply make a mistake, that was then copied by another? It is notable that both books come from roughly the same region, northern Artois or just possibly western Flanders, and are not too distant in date, BnF fr 110 probably from the 1290s and Douce 215 c. 1315–25.64 Formal parallels with subjects drawn from other contexts would suggest to the semioticians that these pictures are probably not simply errors but, to some extent at least, borrow meaning from those other contexts. Were it not for the hermitage structure, clearly made of wicker in Douce 215, an image of Moses before the Burning Bush image would provide an obvious formal parallel, a theme generally common in biblical narrative and typology, whose interpretations also included the Virgin Birth.65 Closer analogies of form and meaning are found in two images of saints in a compendium probably illustrated in the same region as the two Vulgate Cycle manuscripts, the so-called Rothschild Canticles, MS 404 in the Beineke Library of Yale University (figs. 20, 21).66 Neither hermitage is identical to those in the Vulgate Cycle images, yet the formal analogies are striking. Might these be the Not in Arthurian Legends. See Vitzthum (1907), 146 and Stones (1970/71, ch V) A less expansive version of this discussion was given at the 1983 Rennes conference on Artistes, artisans et production artistique (1990). For the text Sommer III, p. 185, line 37. 63 This is from the same workshop as Add10292–4, Roy l4 E.lll and belongs to the Amsterdam/Rylands fr.l set of volumes. The artist of this folio reappears nowhere in the other Add and Roy sets, see Stones (1987). For the text see Pauphilet (1965) 115. 64 Stones (1970/71, chs.V, VI; 1987). 65 See the ‘Moses-Typologie’ entry in Kirschbaum (1968–76). Might there be other significance to the use of wicker? The Glastonbury Chronicle says Joseph of Arimathea’s first church at Glastonbury, the first Christian shrine in England, was made of wattles (Latin virgis torquatis, Carley and Townsend [1985: 2–31]). I am grateful to Faith Lyons for helpful discussion of this point. See now Gilchrist and Green, Glastonbury Abbey (2015), under Joseph of Arimathea and St Joseph’s Chapel. 66 Described by James (1904). See also Hamburger’s description in Shailor (1987), in anticipation of his forthcoming monograph on the manuscript (Hamburger 1990). 62

88

kinds of compositions that the Vulgate Cycle miniaturists were drawing upon? A further example of a similar subject, in the margin of Cambridge, Trinity B. 11.22 (fig. 22), a Franciscan Book of Hours, from the same workshop as Douce 215, suggests that the motif was one of a number of stock scenes that could be made use of independently of a particular textual context. Or might the image be adapted from another literary context? Here an episode in the Continuations of Chretien’s Perceval provides another possible source: it is Perceval’s encounter with the child in the tree, a scene which is illustrated in three manuscripts, BnF fr 12576 (fig. 23), BnF fr 1453 and Montpellier H 24967. Of these, two are almost certainly earlier than BnF fr 110 and Douce 215 and show that a model was available prior to the creation of the Lancelot/hermit episode they show. Of course one cannot prove that those painters used it, and if did, they adapted it to the text at hand by substituting the hermit, together with his hermitage, for the child with his apple. Did the successive overlay of possible meanings also transfer with the image, and were those meanings understood? A hint that there might be a deliberate attempt to overlay one meaning with another is suggested by the fact that the texts on the book of hours page include, twice, the word ‘ignis’, suggesting that the inclusion of the knight with hermit in tree on this page might have been intended as a parody of the Burning Bush. This is a book whose margins are filled with all kinds of hybrids, animals and visual jokes of which only a few can be interpreted, but where an intended joke would not be unexpected.68 Might it also have been seen as a Lancelot joke? To what extent such an interpretation might also work without the presence of a particularly suggestive word in the text, is open to question. Was the Burning Bush joke — or the hermit-saint joke, or both, not to mention Perceval and the child in the tree — also transferred, with the image, to the Vulgate Cycle? By the same token, did the image of the embrace of Lancelot and Guinevere carry with it the visual connotations of sponsus and sponsa, Roach (1971) p. 568. In BnF fr 1453, the child has a distinctly female-looking hair style. The episode is also illustrated in BnF fr 12577, but with the child standing at the foot of the tree, not in its branches. For a wide-ranging survey, inspired by Loomis himself, of Wauchier’s textual sources and some of their major iconographical manifestations, see Greenhill (1954). For related themes in Christian iconography see the article ‘Baum’ by J. Flemming in Kirschbaum et al (1968–76), vol. I, cols. 258–68. 68 For its workshop see Stones (1987). Randall (1966) reproduces many of the marginal images. A recent survey of visual jokes in the Middle Ages is Schmidt (1984); for the literary equivalents see Ménard (1969); neither treat this image directly, but both provide the context in which some understanding can be arrived at. 67

ARTHURIAN ART SINCE LOOMIS

89

20. New Haven, Yale University, Beinecke Rare Books and Manuscript Library 404, fol. 27 Hermits (© Yale University)

21. New Haven, Yale University, Beinecke Rare Books and Manuscript Library 404, fol. 28 A dendrite hermit (© Yale University)

90

22. Cambridge, Trinity College, Book of Hours, fol. 207 Knight visits a dendrite hermit (© Trinity College)

23. Paris, BnF fr 12576 Second Continuation of Perceval, fol. 148, Perceval sees a child in a tree (© BnF)

ARTHURIAN ART SINCE LOOMIS

91

Christ and Ecclesia, Hosea and Gomer, Cligès and Fenice? The semioticians would argue that it did, but these examples also suggest that the boundaries between deliberate and incidental overlay of meaning, as those between the parody or joke and the ‘serious’ were often elusive. In the absence of a confirming text, one cannot be certain which level or levels were intended or understood, or by whom: artist or audience, neither or both, but the visual similarities between the images demonstrate how important it is to couple text-specific study with that of comparative iconography and style. Towards the next 50 years of Arthurian art If the study of comparative iconography and style contributes anything to our understanding of Arthurian art, it is that it provides an artistic context in time and space which enables us to see more clearly where particular artefacts or images came from and what they meant to the culture that commissioned and produced them. Since 1938 we have learned much more about who it was that commissioned Arthurian literature and art, from Keller (1985) on literary patronage of the age of Philip Augustus, to Stanger (1957) on the thirteenth century court of Flanders, de Winter (1985) on Philippe le Hardi’s inventories and Meiss (1967, 1968, 1974) on Jean de Berry and his circle, and it is hoped that careful attention to the inventories like those listed by Bautier and Sornay (1985) will extend our knowledge of who the patrons were. Of especial interest are the still scanty, but nontheless important, mentions of Arthurian interests among the bourgeois class, like Jean Cole of Tournai, whose Merlin is listed among the six or seven literary manuscripts in his will of 1303.69 Another important aspect of the socio-historical importance of Arthurian art is demonstrated in the recent surge of interest in heraldry (Brault 1959, 1960, 1966, 1972; Nickel 1983; Pastoureau 1975, 1976, 1979, 1980, 1983). These studies emphasize the links between the historical owners or patrons and the Knights of the Round Table, their heroes (or anti-heroes), and between those patrons of the past and their artefacts still present to us today. The few images discussed here serve to underline how much work is still to be done in the area of comparative iconography, text by text, motif by motif, within and between media and periods, and between Arthurian art and art of other subjects, both sacred and secular. The Loomises amassed an 69

Stones (1987).

92

impressive body of Arthurian art and artefacts. Much more has been added since, and new ways of looking at it have emerged. But the most significant change, to which the numerous publications of the past half-century bear witness, is the realization that Arthurian art, and ‘secular’ art in general, are more than a footnote to the history of medieval art, or a handmaid to medieval literature. They have come to be seen rather as pointers towards important aspects of the cultural and intellectual history of the Middle Ages, and this promises well for the next fifty years of Arthurian study.

IV Indications écrites et modèles picturaux, guides aux peintres de manuscrits enluminés aux environs de 1300

A

u XIIIe siècle la nouvelle littérature en langue vulgaire, composée principalement au siècle précédent, fit sentir l’impact de son influence dans le domaine de l’enluminure. Les textes nouveaux réclamaient l’élaboration de nouveaux cycles d’illustrations et les enlumineurs avaient besoin de directives pour illustrer des scènes qui leur étaient peu familières ou dont ils n’étaient pas sûrs de l’emplacement dans une longue série de miniatures telle que caractérise, par exemple, l’enluminure des romans en prose. Au XIIIe et au début du XIVe siècle, quand les programmes d’illustrations de textes en langue vulgaire étaient encore en cours de développement, quelques manuscrits présentent des exemples révélateurs du genre de directives dont avaient besoin les artistes. Celles-ci étaient en partie visuelles, vu que beaucoup de scènes existantes, de type bibliques, hagiographiques ou autres, déjà familiaires aux artistes, par exemple les scènes de batailles, les scènes de banquet, les personnages au lit, pouvaient facilement s’adapter aux sujets profanes.1 D’autres scènes Certaines parties de cet article ont été présentées sous une forme différente aux colloques de la College Art Association et de l’Institut médiéval de Western Michigan University à Kalamazoo en 1972, 1973 et 1979. Nous remercions MM. E. Kitzinger, A. Cutler, D. Coq, D. Muzerelle et V. Pace, ainsi que Mme P. Stirnemann, de leurs interventions à Rennes. Publié dans Artistes, artisans et production artistique (Actes du colloque, Rennes, 1983), ed. X. Barral I Altet, 3 vols., Paris, 1990, vol. III, pp. 321–349. 1 Pour l’adaptation en général, voir notre article « Sacred and profane », repris dans ces essais. Pour quelques cas spécifiques dans un atelier parisien du début du XIIIe siècle: ead., « The earliest » repris dans ces essais.

94

sont semblables par la disposition de personnages ou d’objets, d’éléments de paysage ou d’architecture ou d’autre motif particulier. On ne s’en étonne pas puisque ce procédé d’emprunt de modèle pour en faire la copie est fondamental à la transmission de scènes entières et de motifs particuliers dans l’art médiéval.2 C’est ainsi qu’à l’occasion on utilisait des modèles impropres au sujet ou des modèles qui, de toute évidence, avaient été inventés pour un autre contexte. Souvent il existait en même temps que ces modèles picturaux, et tout aussi importants qu’eux, des directives sous forme de notes ou de symboles apparaissant en marge ou dans l’espace même de l’illustration. Ce genre de directives, destinées à l’enlumineur, n’étaient pas inconnues avant le XIIIe siècle, mais leur nombre plus abondant à cette époque suggère le développement d’une approche nouvelle de la tâche pour les cycles d’illustrations et pour l’adaptation de scènes appropriées à un certain passage dans un texte particulier. Beaucoup de manuscrits séculiers,3 particulièrement ceux des longs manuscrits en prose ou des recueils de vies de saints tout aussi longs, avaient aussi un système de rubriques explicatives qui à la fois résumait le contenu du chapitre et expliquait celui des miniatures. Dans la plupart des textes en langue vulgaire, la place des enluminures n’était pas standardisée. Le contenu des illustrations ne l’était pas non plus. Il y avait une telle prolifération de versions différentes de textes qu’il était rare qu’un atélier puisse posséder une copie illustrée complète de chacune d’elles pour servir directement de modèle à l’enlumineur.4 Ces textes étaient souvent très longs et un lecteur pressé, ne désirant pas passer de longues heures à tout lire, était sans doute tenté de parcourir le livre en regardant les images; les miniatures et les rubriques ont dû lui être utiles car elles lui Parmi de nombreux exemples citons le cas du psautier d’Utrecht: Dufrenne, 1976. Nous employons le mot « séculier » tout en reconnaissant son caractère ambigu. II nous semble que cette ambiguïté attire l’attention sur la possibilité que le clergé séculier des ordres mineurs ait joué un rôle dans la production livresque à l’époque. 4 II est particulièrement frappant, à cet égard, de comparer les enluminures dans les multiples exemplaires d’un même texte copiés et enluminés dans le même atelier, par exemple les trois Romans d’Alexandre Berlin, Kupf. 78.C.1, BR 11040 et BL Harley 4979, discutés dans notre article « Alexander ». Un quatrième exemplaire illustré de ce même texte a été récemment découvert dans une collection privée. Il contient plusieurs scènes qui manquent dans les trois autres manuscrits. Nous remercions J. Marrow d’avoir eu la gentillesse de nous signaler sa découverte et de nous avoir fourni sa description du manuscrit. Voir Stones et Ross, « Alexandre ». Voir aussi les situations analogues des Lancelot BN ms. fr 110 et Bonn 526, ainsi que celle de BL Add 10292–10294, Roy. 14 E.III, Ryl/Douce/Kraus, cf. infra. 2 3

INDICATIONS ÉCRITES ET MODÈLES PICTURAUX

95

permettaient de reconnaître les personnages principaux et de comprendre la logique liant chaque épisode au reste de l’histoire. Un examen des relations mutuelles entre la miniature, l’esquisse, le modèle visuel, la note et la rubrique permet de mieux comprendre les pratiques d’ateliers en cette période qui devait être passionnante dans le domaine de l’enluminure séculière. Les modèles visuels feront d’abord l’objet de notre attention. Les indications écrites seront analysées plus loin. Dans le domaine de l’enluminure des textes en langue vulgaire, on se retrouve, parfois, devant une situation complexe: certaines scènes se trouvent utilisées dans des situations différentes de celles pour lesquelles elles avaient été inventées à l’origine, avec des conséquences inattendues, même confuses, pour l’interprétation de la miniature, mais qui sont d’autant plus révélatrices du point de vue du procédé de création. Examinons quelques exemples tirés du Lancelot en prose français. Une illustration du ms. Bonn 526 représente une saignée (fig. I). Le bras d’une femme assise est pris dans un tourniquet et le sang est recueilli dans un bol; la main de la femme est soutenue par un bâton. Le texte décrit l’usage qu’on va faire du sang, et qui se trouve représenté également dans la même miniature: la sœur de Perceval, pour sauver une femme atteinte de la lèpre, fait don de son sang avec lequel on lave la poitrine de la femme mourante.5 Le texte décrit l’ouverture de la veine à l’aide d’une lame, ce qui n’est pas représenté dans l’image, mais ne fournit pas les détails du bâton ni du tourniquet. On dirait que l’artiste a dû adapter une scène médicale, mais il est difficile d’en retrouver dans la tradition illustrée médicale qui corresponde en tout détail à la version du Lancelot. Un manuscrit médical à peu près contemporain, BnF lat 7134, nous présente une scène qui montre l’ouverture de la veine (fig. 2) mais qui n’a pas servi de modèle pour la scène de Lancelot de Bonn.6 C’est dans le contexte du calendrier qu’on retrouve un exemple antérieur au Lancelot qui lui ressemble un peu plus, puisqu’il fournit le bâton et le bol (fig. 3); 7 mais aucune de ces scènes ne presente exactement les mêmes Cet épisode fait partie du texte de la Queste del saint Graal, quatrième branche du Lancelot en prose, éd. Pauphilet, p. 237. Pour l’iconographie des manuscrits de la Queste voir Stones, Illustrated Lancelot, p. 319 pour cet épisode. 6 Gourdon et Sournia, éd., Médecine, n° 44. Le manuscrit y est attribué à la fin du XIIIe siècle ou au début du XIV e. 7 C’est P. Stirnemann qui nous a montré cet exemple et nous a fourni la référence à Gordon, « Fécamp Psalter ». Nous remercions Mme P. Stirnemann ici. C. Hohler a remarqué que ce psautier, soi-disant de Fécamp, contient un calendrier de Ham. Ses notes se trouvent à la Conway Library, Courtauld Institute. 5

96

détails. L’artiste a-t-il donc utilisé un modèle tiré d’un manuscrit médical maintenant perdu, ou a-t-il inventé ces détails d’après la pratique médicale qu’il connaissait dans la vie quotidienne ? Beaucoup plus tard dans la tradition illustrée des manuscrits de médecine, on retrouve la combinaison de motifs bâton, bol, sang, mais les mains du médecin tiennent lieu de tourniquet (fig. 4). 8 En tout cas, le choix de modèle convient fort bien au texte du Lancelot et l’emploi de détails supplémentaires dans la miniature en enrichissent l’intérêt. Un second exemple d’emprunt iconographique, dont les sources semblent dériver d’un texte différent, aboutit à un autre résultat quant à la compréhension de la miniature. Dans le Lancelot en prose, BnF fr 110, on trouve au fol. 206 v (fig. 5) la visite de Lancelot à l’ermite, sujet assez courant dans la tradition des Lancelot illustrés. De façon étrange, l’ermitage, dans ce cas, est une structure avec un toit arrondi, de nature assez indéterminée, placée dans un arbre; Lancelot s’en approche à cheval. Le texte ne donne aucune justification pour ce traitement de l’ermitage.9 Malgré la fréquence de visites aux ermites dans le texte, comme dans l’illustration du cycle de Lancelot, le motif de l’ermitage dans l’arbre, tel qu’il apparaît dans le ms. fr 110, est relativement isolé. Il se retrouve, néanmoins, transposé dans un autre manuscrit un peu plus tardif, Douce 215, pour illustrer un épisode différent (fig. 6). 10 Cette fois l’ermitage, également une structure arrondie, Nous remercions L. Voigts de nous avoir signalé ce manuscrit anglais, qu’on pense être la copie d’un exemplaire fait pour le duc de Bedford, et de nous en avoir fourni la photographie, ainsi que la référence à Ogden, ‘Cyrurgie’. 9 Sommer, Vulgate Version, 3, pp. 185–237. A noter que l’édition de Micha commence au vol. 4 de Sommer. 10 Pauphilet, Queste, p. 115; Stones, Illustration of Lancelot, p. 316. Le fragment Douce 215 de la Bibliothèque bodléienne d’Oxford fait partie du Lancelot de la Bibliothèque John Rylands de l’Université de Manchester, ms. fr 1: Stones, «Short note ». Ce recueil de Lancelot se complète par trois volumes qui se trouvaient à New York, en possession de H.P. Kraus, Catalog, n° 165, New York, 1983, n°3 et qui ont été acquis par la Bibliotheca Hermetica, Amsterdam (depuis le 7 décembre 2010 vendu à la vente chez Sotheby’s, lot no. 33, à un collectionneur privé). Le premier tome contient l’Estoire; il est numéroté avec le chiffre imprimé 1045 au dos et 1045/3630 sur la feuille de garde, dans le dernier cas le 5 de 1045 a été corrigé à l’encre (au XIX* siècle?) en 7. Nous pensons que la correction est juste, et qu’il s’agit ici de Phillipps 1047. Le ms. Phillipps 1045 est actuellement Yale 227: Cahn et Marrow, Manuscripts at Yale, n° 34. La meilleure liste des manuscrits du Lancelot se trouve dans Woledge, Bibliographie. Les deux autres tomes ayant appartenu à H.P. Kraus sont les anciens Phillipps 3630, Merlin et Lancelot. Le premier tome de Phillipps 3630 contient Merlin et, commençant au fol. 37, la première partie du Lancelot (Sommer, Vulgate Version, 3, 3/1) le 8

INDICATIONS ÉCRITES ET MODÈLES PICTURAUX

97

1. Bonn, Landes-undUniversitätsbibl. 526, Queste, fol. 436/447, Saignée et guérison de la lèpre (©Landes-undUniversitätsbibl. Bonn)

2. Paris, BnF lat. 7134, Magister Rollandus, Chirurgia, fol. 18v, Part III, Saignée (©BnF)

98

3. La Haye, KB 76 F 13, Psautier de Ham, fol. 2, Saignée (© La Haye, KB)

4. Bristol, Avon County Library, Reference Library 10, Gui de Chauliac, Cyrurgie, fol. 230, Saignée (© L. Voigts)

INDICATIONS ÉCRITES ET MODÈLES PICTURAUX

5. Paris, BnF fr 110, Lancelot, fol. 206v, Lancelot rend visite à un hermite (© BnF)

6. Oxford, Bodleian Library, MS Douce 215, Queste, fol. 14 Lancelot rend visite à un hermite (© Bodleian Library)

99

100

7. Cambridge, Trinity College, Livre d’heures, fol. 207 Un chevalier rend visite à un hermite dendrite (© Trinity College)

INDICATIONS ÉCRITES ET MODÈLES PICTURAUX

101

est fabriquée en osier tressé, et Lancelot est descendu de son cheval pour s’entretenir plus directement avec l’ermite. Il faut admettre un lien de parenté étroit entre les deux manuscrits pour la représentation de cette scène à cause de sa rareté. 11 Les menus changements dans le détail témoigneraient de l’esprit créateur d’un artiste plus tardif, plus réaliste, mais qui dépend, néanmoins, du modèle du ms. fr 110. Ce rapprochement entre le ms. fr 110 et Douce 215 est d’autant plus intéressant que, en ce qui concerne la tradition du cycle en général, les deux manuscrits appartiennent à des groupes différents, aussi bien du point de vue stylistique que du point de vue du cycle de miniatures. Le ms. fr 110 contient le cycle court; tandis que Douce 215 fait partie du groupe des manuscrits qui contiennent le cycle long. 12 Le ms. fr 110 s’apparente stylistiquement au Lancelot de Bonn, écrit en 1286: on peut les attribuer au même atelier. Le ms. Bonn 526 contient également le cycle court de miniatures, mais l’ermite dendrite n’y paraît pas. La chronologie relative entre les deux manuscrits s’établit difficilement.13 On pourrait proposer une date légèrement plus tardive que 1286 pour le ms. fr 110 et supposer que l’artiste ait rencontré deuxième tome continue le Lancelot jusqu’à la fin de la partie centrale (ibid., 4, p. 362); la dernière ligne est la rubrique pour le commencement de la dernière partie du Lancelot propre, l’Agravain (ibid., 5). Le début de l’Agravain se trouve dans le ms. Douce 215 puis dans Ryl. C’est le regretté N. Ker qui nous a signalé ces manuscrits et qui a proposé le rapport avec la partie Ryl/Douce. Nous exprimons nos remerciements ici, ainsi qu’à M. H.P. Kraus qui nous a gentilment permis d’étudier et de photographier ces volumes. 11 Nous ne connaissons aucun autre exemple de la visite au dendrite dans les manuscrits du cycle de Lancelot. 12 Ce groupement de l’enluminure en cycles longs et courts est valable pour le Lancelot propre comme pour la Queste et la Mort Artu: Stones, Illustration of Lancelot, pp. 310, 340. Micha, Lancelot. 13 Nous reconnaissons qu’à cette époque l’atelier artistique est encore mal compris et difficile à définir; voir surtout les études d’Avril, « A quand remontent » et de Branner, Manuscript Painting. Soulignons également l’importance des rôles de la taille de la fin du XIIIe siècle pour la définition de l’atelier d’enluminure, mise en valeur par Baron, « Enlumineurs », ainsi que les autres sources documentaires du nord de la France publiées par Dehaisnes, Documents et extraits. Voir également notre discussion des ateliers du nord de la France entre 1250 et 1340 dans notre thèse lllustration of Lancelot, pp. 27–127. Voir maintenant Alexander, Medieval Illuminators, ainsi que Rouse et Rouse, Manuscripts and their Makers. Nous n’essayons pas ici de présenter les arguments en toutes dimensions pour la structure de l’atelier artistique à cette époque, mais simplement d’en définir quelques caractéristiques sur le plan technique de la création. Il faudra attendre une analyse stylistique beaucoup plus détaillée des manuscrits discutés ici pour prouver la validité de notre groupement ainsi que pour définir de façon plus raffinée notre conception de l’atelier même et de l’individualité

102

un motif qu’il jugât convenable pour cet épisode, bien que le motif en manquât dans son modèle, ms. Bonn 526. Reprenons plus tard la question de la source du motif. Soulignons le fait que la saignée déjà décrite ne paraît pas dans le cycle du ms. fr 110; 14 il ne s’agit donc pas de voir ce peintre plus inventeur dans l’ensemble de son programme iconographique. Ce ne sera pas la dernière fois que deux Lancelot issus du même atelier présentent des variantes dans le nombre de miniatures dans leur cycle comme dans le traitement de scènes particulières. Les ateliers n’ont apparemment pas gardé un exemplaire complètement illustré de chaque texte qu’illustraient leurs

artistique. Nous proposons d’y revenir. Nous avons présenté une esquisse des caractéristiques stylistiques de l’enluminure de ces manuscrits, dans notre thèse (pp. 208–224) et une liste de manuscrits est citée dans notre article « Sacred and profane » p. 108. En attendant de traiter cet atelier en détail ailleurs, nous proposons d’y inclure les suivants; a) les débuts: Ars 3516 (recueil avec un calendrier de Thérouanne et des tables qui commencent en 1268), BL YT 43 (psautier avec calendrier de Saint-Omer, écrit en 1277), BnF, ms. fr 19162 Lancelot et ms. fr 24394 Lancelot, Sainte-Geneviève ms. 2200 Image du Monde etc. écrit en 1276 et 1277; b) psautiers: Douce 24 (calendrier de Saint-Omer avant 1298), Douce 49 (calendrier de Saint-Omer avant 1298), Morgan M.79 (calendrier de Saint-Omer avant 1298), Arras Mus. 47 (calendrier de Thérouanne ou de Cambrai avant 1298), Douai 193 (calendrier d’Arras avant 1298), Abbeville 3 (calendrier de Villers-Saint-Josse); bibles: Saint-Omer 5, BR 10753; bréviaire de Marquette avant 1298: Cambrai 99; missels de la cathédrale de Cambrai: Cambrai 153, 154 (avant 1298); livres d’heures: Cambrai ms. 87 (avant 1298; frontispice seulement); Getty/Ludwig IX,3 (« Ruskin Hours », peintre de la litanie seulement); Yale 404 (« Rothschild Canticles », un des peintres des miniatures en couleurs); Walters 90 (Heures de Thérouanne après 1298); c) textes littéraires ou historiques: Morgan M.108 Livre d’échecs; Bonn 526 Lancelot (écrit en 1286 par Arnulphus de Kayo à Amiens); Getty/Ludwig XV, 5 Bestiaire (écrit après 1277); Bibliothèque nationale, ms. fr 110 Lancelot; British Library, Add. 5474 Tristan; Boulogne ms. 192 Guillaume d’Orange (écrit en 1295, premier peintre); Douai ms. 797 Vincent de Beauvais, Speculum historiale (provient de Marchiennes); Boulogne ms. 131 Vincent de Beauvais, Speculum historiale (écrit en 1297 pour Eustache Gomer de Lille, abbé de Saint-Bertin); Boulogne ms: 130 Vincent de Beauvais Speculum historiale (copie de Boulogne ms. 131; premier peintre); Morgan M.751 Petrus Pictaviensis, Abrégée de l’histoire. Sur le plan stylistique, nous rapprochons le ms. BnF fr 110 des missels Cambrai ms. 153 et ms. 154, et du frontispice de Cambrai ms. 87, de la litanie de Getty/Ludwig IX, 3 et d’une partie de Yale 404; en général ce sous-groupe nous semble plus proche de Boulogne ms. 192 et de Boulogne ms. 131, écrits en 1295 et 1297, que de Bonn 526, écrit en 1286. Mais nous reconnaissons qu’il est très difficile d’établir la chronologie de ce groupe ainsi que la location de l’atelier sans une analyse beaucoup plus détaillée. Pour les manuscrits de Vincent de Beauvais et la question de production livresque dans le milieu cistercien voir notre article The Minnesota Vincent. Voir maintenant Stones, « Entre Cambrai et Saint-Omer ». 14 Le copiste n’a pas prévu une miniature à cet endroit dans le texte du ms. fr 110 de la BnF. Voir notre thèse, Illustration of Lancelot, p. 319.

INDICATIONS ÉCRITES ET MODÈLES PICTURAUX

103

employés. Il faudrait admettre d’autres sortes d’intermédiaires entre modèle et copie. Parmi les Lancelot illustrés, les pendants stylistiques du ms. Douce sont BL Add 10292–10294 (enluminé en ou après 1317 n.s.), et BL Roy. 14 E.III. 15 Comme le ms. Douce, ces derniers contiennent le cycle long de miniatures, mais à cet endroit le ms. Add montre Lancelot devant un ermitage, situé par terre, représenté comme une chapelle, et le ms. Roy. montre Lancelot avec l’ermite dans l’ermitage, représenté également sous forme de chapelle située par terre.16 Les artistes de Roy. et d’Add (ou peut-être l’artiste, car les enluminures sont extrêmement proches) ont-ils supprimé un motif qu’ils jugèrent inapproprié ? Ou bien cette miniature dans le ms. Douce a-t-elle été ajoutée après-coup à la suite de la parution, dans l’atelier, du ms. fr 110, ou d’une autre source contenant le même motif ? Cette dernière possibilité est suggérée par le fait que cette représentation dans le ms. Douce appartient, il nous semble, à l’œuvre d’un deuxième artiste, peut-être légèrement plus tardif que le premier, mais dont l’œuvre est très proche stylistiquement de celui du premier peintre. Ce deuxième artiste est l’enlumineur d’un groupe de manuscrits, publié récemment par K. Carlvant et qu’elle a nommé le «Copenhagen Master».17 Loomis, Arthurian et cf. supra, 10. Stones, Illustration of Lancelot, p. 316. 17 Carlvant, « Collaboration ». Nous avons traité cet atelier dans notre thèse, Illustration of Lancelot, pp. 225–238. La participation, peut-être la collaboration des deux artistes dans ms. Douce 215 nous semble indiquer l’existence de liens très proches entre le «Copenhagen Master» et le reste du groupe auquel appartient le peintre principal de Douce/Ryl/Kraus (olim Amsterdam), BL Add 10292–10294 et Roy. 14 E.III: BL, Stowe 17 (psautier avec un calendrier de Liège), Saint-Omer ms. 270 (psautier présenté à la Chartreuse de Longuenesse, près de Saint-Omer, par Gilbert de Sainte-Aldegonde en 1325, frontispice seulement); nous trouvons également de liens stylistiques très proches avec l’Add 28784 A et B, de la BL, publiés par Oliver, « Reconstruction ». Signalons en même temps un courant stylistique différent, qui parait dans l’Add 10292–10294. Celui-ci a également été un projet collaboratif entre deux peintres, dont l’œuvre du premier ressemble tellement au style du peintre de Roy. 14 E.III et du premier peintre de Douce/Ryl/Kraus (olim Amsterdam) qu’on pourrait y reconnaître la main d’un même peintre. Le deuxième peintre de l’Add 10292–4, qui se distingue nettement du premier, semble avoir travaillé aussi sur les mss. Morgan M.754/Add 36684 (Heures de Thérouanne, après 1318) et Boulogne 130 (après 1297), dont le premier peintre fait partie de l’atelier du ms. fr 110 de la BnF et de Bonn 526 (cf. supra, n. 11). Nous proposons de traiter ailleurs ce réseau compliqué de liens stylistiques; nous attirons l’attention ici sur ce rapport entre l’atelier du ms. fr 110 et Bonn 526 et celui de Douce 215/Ryl/Kraus (olim Amsterdam), qui présente un pendant sur le plan stylistique au rapport iconographique suggéré par l’emprunt du motif dendrite d’un atelier à l’autre. Nous reviendrons ailleurs sur la 15 16

104

Parmi les manuscrits que K. Carlvant rapproche de l’œuvre du « Copenhagen Master» se trouve un curieux livre d’heures, Cambridge, Trinity B. 11.22. Sur la bordure du fol. 207 paraît un chevalier qui salue un ermite dendrite, dont l’ermitage, dans l’arbre, prend maintenant la forme d’une petite chapelle (fig. 7). Ce motif fait partie des illustrations dans le texte des heures du Saint-Esprit, sans avoir apparemment un rapport direct avec ce texte. Il est intéressant de remarquer ailleurs dans l’œuvre des artistes de cet atelier, dans le BL Stowe 17, une image de saint Siméon Stylite dans la marge du fol. 123 v.18 Peut-on conclure à un intérêt particulier aux saints du désert dans ce groupe de livres religieux et séculiers, et qui dépendrait du Lancelot ms. fr 110? Cherchons maintenant la source du motif dans ce dernier manuscrit. En fait, les cycles de vies des saints du désert sont rares dans l’art franco-flamand, comme dans l’art d’Europe de l’Ouest en général, à cette époque. Ainsi la présence d’un tel cycle dans le ms. Yale 404 (Rothschild Canticles) est d’autant plus frappante qu’il existe un certain rapport stylistique entre l’autre partie du cycle de Yale 404 et l’atelier du ms. fr 110.19 Nous signalons ici la présence de deux scènes dans ce manuscrit qui, ensemble, fournissent les éléments du rencontre avec l’ermite dendrite dans les manuscrits Bibliothèque nationale, ms. fr 110, Douce 215 et Trinity B. 11.22: la rencontre avec l’ermite dans son ermitage par terre, mais de forme arrondie et couvert de feuilles (fig. 8); l’ermite avec son ermitage, structure de pierre, dans un arbre (fig. 9). Le peintre du ms. fr 110 aurait adapté ces compositions pour en faire sou image; l’artiste de Douce 215, qui dépend du ms. fr 110, en a élaboré les détails et a peut-être été lui-même responsable pour une deuxième version dans Trinity B. 11.22. Il est probable que l’emploi du motif dans ce dernier manuscrit dépend de la tradition des Lancelot plutôt que celle de Yale 404 à cause de l’emploi du chevalier armé qui ne se trouve pas parmi les dendrites du question de la localisation et de la chronologie de ces ateliers, ainsi que sur l’analyse détaillée de leur structure. Voir Stones, « Entre Cambrai et Saint-Omer ». 18 Cité dans Randall, Images, p. 82: « cleric on column with club and shield». 19 Voir Cahn et Marrow, Manuscripts at Yale, n° 29. Nous avons présenté une discussion de style de Yale 404 à Kalamazoo en 1979 et nous proposons d’y revenir ailleurs. Pour le style d’un des peintres du cycle de miniatures en couleurs cf. supra, n. 13. Le cycle des saints du désert est fait en encre avec une couche légère de peinture, par un autre artiste; nous l’avons traité brièvement dans notre article «Arthurian Art Since Loomis», repris dans ces essais. Nous pensons retrouver le peintre des saints du désert dans le ms. BR 456–457, que Gilles le Muisis a fait écrire pour l’abbaye de Saint-Martin à Tournai quand il était encore moine (avant 1331).

INDICATIONS ÉCRITES ET MODÈLES PICTURAUX

105

manuscrit de Yale. C’est la rareté des représentations de dendrites qui mène à la supposition qu’une parenté proche existe, pour cette scène, entre ces quatre manuscrits, parenté confirmée en partie par les rapports stylistiques. L’interprétation de la scène dans les Lancelot n’est pas encore tout à fait évidente: l’artiste du ms. fr 110 a-t-il voulu souligner le caractère lointain, mystique peut-être, de l’ermite, en utilisant le modèle d’un dendrite, ou s’agit-il tout simplement de l’emprunt d’un modèle quelconque qui convenait à peu près au contexte ? L’emploi du motif a été réussi à l’époque, puisqu’il a été recopié plusieurs fois; pour l’observateur moderne, cette intervention, quoi qu’en ait été le but original, fait la transformation d’une épisode courante en une scène pleine de complexité et d’intérêt inattendus. Ces exemples, dont l’explication se complique par le manque de certitude quant à la façon exacte dont s’est produit le rapport entre modèle et copie et quant au but de l’artiste, suggèrent néanmoins une certaine habitude de copie selon laquelle l’artiste ne dépend pas d’un exemplaire illustré du texte particulier mais utilise des modèles ou motifs qui viennent d’un stock de scènes ou de parties de scènes qui existent indépendemment du texte. La présence d’esquisses et parfois de notes dans les marges de quelques manuscrits confirme l’idée qu’il y a eu une ou plusieurs étapes intermédiaires entre texte et miniature et ajoute ainsi un autre niveau de complexité entre l’image modèle et l’image copie. Parfois les esquisses ont été préservées dans les marges à côté des miniatures. Elles sont généralement exécutées à la pointe sèche ou à la mine de plomb, plus rarement à l’encre. Elles sont surtout évidentes dans les ateliers où la production semble avoir été organisée à grande échelle, avec de nombreux participants, comme était le cas dans un atelier parisien dans le deuxième quart du XIVe siècle, où la qualité du produit final était médiocre et où l’on n’avait même pas pris soin d’effacer ces esquisses. Quelques exemples en seront mentionnés ci-après. Cette sorte de pratique est attestée bien plus tôt, mais elle réapparaît de nouveau là où de nouveaux textes étaient illustrés, là où un nouveau cycle d’illustration était créé. Il en est de même pour les indications écrites pour l’enlumineur qui apparaissent souvent en même temps que les esquisses, ou qui les remplacent. L’emploi d’instructions écrites pour l’enlumineur nous ramène aux débuts mêmes de la miniature dans l’Itala de Quedlinbourg,20 cas pourtant isolé en ce que les notes consistent en de longues phrases écrites sous la peinture, dans l’espace de la miniature, 20

Weitzmann, Age of Spirituality, n° 424 avec bibliographie.

106

procédé qui semble se modifier considérablement par la suite. C’est ainsi que les notes dans le recueil d’auteurs classiques de la BnF, fait vers la fin du XIIe siècle, ont été placées dans les marges sous forme abrégée.21 Les listes écrites pour l’illustration du psautier qu’a publiées F. Wormald et qui existent également aux environs de 1300 dans l’œuvre du Pictor in carmine, suggèrent que ces notes dépendent elles-mêmes d’une étape intermédiaire.22 Dans la deuxième moitié du XIIIe siècle les romans ne sont pas les seuls textes à recevoir un nouveau cycle de miniatures. On retrouve les notes et les esquisses dans les manuscrits liturgiques et hagiographiques. Quelques exemples indiquent de légères différences de traitement dans l’emploi des esquisses et des notes. Le Pontifical de Cambrai à Tolède présente un cas intéressant où l’on trouve de nombreuses esquisses à la pointe sèche, accompagnées souvent par des notes écrites en français (fig. 10 et 11).23 Ces pratiques se voient encore avec difficulté puisqu’on s’est donné la peine de les effacer en partie, comme on a dû faire très généralement. On peut constater toutefois qu’il y a plus de notes et d’esquisses dans les parties du texte qui concernent la consécration de l’autel et des éléments utilisés pour le construire, sujets assez rares dans l’illustration du pontifical pour que l’artiste ait eu besoin d’aides-mémoire. Ces guides ont dû en même temps éviter la nécessité de lire le texte, latin ou autre. Avec une série de directives, données peut-être par un chef d’atelier ou par un autre personnage intermédiaire, l’artiste aurait pu procéder beaucoup plus rapidement que s’il devait s’arrêter à chaque espace vide pour se rassurer d’avoir choisi le sujet qu’il fallait y mettre. C’est ainsi que s’explique la présence d’une note écrite à côté d’une initiale dans le Missel de la cathédrale d’Arras, qui représente un prêtre célébrant la messe, sujet très courant pour ce contexte, mais où l’on retrouve quand même une directive détaillée, ayant repère aux topoi généralisés: « un prêtre les bras étendus devant un autel» (fig. 12). 24 On s’étonne, pourtant, de Avril, « Un manuscrit d’auteurs classiques ». Wormald, «Two illuminated psalters »; James, « Pictor in carmine »; Brown, «Pictor in carmine ». 23 Janini et Gonzalvez, Manuscritos liturgicos, n° 216. Pour les autres manuscrits de cet atelier cambrésien voir Stones et Steyaert, Minnesota Collections, pp. 12–21. Au fol. 114 du manuscrit on lit, dans la marge inférieure, « li ..oint lautel de le s.oile..po.. fait une s... »; dessus la note se trouve une esquisse qui montre un personnage mitré, le bras étendu sur un autel (très difficile à déchiffrer). A côté de la note se trouve une marque, qui renvoit à la marge à droite de l’initiale historiée. On y remarque également le mot « dor », directive pour le fond d’or de l’initiale. Nous reprenons ci-après la question de directives pour les couleurs. 24 Leroquais, Sacramentaires et missels, p. 175. 21 22

INDICATIONS ÉCRITES ET MODÈLES PICTURAUX

107

8. New Haven, Yale University, Beinecke Rare books and Manuscript Library 404, Rothschild Canticles, fol. 27 Deux dendrites (© Yale University)

9. New Haven, Yale University, Beinecke Rare books and Manuscript Library 404, Rothschild Canticles, fol. 28 Un hermite dendrite (© Yale University)

108

10. Tolède, Bibliothèque capitulaire, 56. 19, Pontifical de Cambrai, fol. 114, Consécration de l’autel (photo: auteure) 11. Tolède, Bibliothèque capitulaire, 56. 19, Pontifical de Cambrai, fol. 114, note et esquisse à la mine de plomb (photo: auteure)

INDICATIONS ÉCRITES ET MODÈLES PICTURAUX

12. Arras, BM 303 (960), Missel, fol. 61v Consécration (photo: auteure)

13. San Marino, Henry E. Huntington Library, HM 3027, Jacobus de Voragine, Legenda aurea, fol. 118v, Saint Gilles et le cerf, miniature, note et esquisse (© Henry E. Huntington Library)

109

110

14. New York, J. Pierpont Morgan Library and Museum, M.805, Lancelot, fol. 50, Lancelot sur le bord d’une rivière devant un château (© Morgan Library and Museum)

15. Los Angeles, J. Paul Getty Museum, Getty Ludwig XV 3, Bestiaire, fol. 90v, Balaine(© J. Paul Getty Museum)

INDICATIONS ÉCRITES ET MODÈLES PICTURAUX

16. Paris, BnF fr. 241, Jacobus de Voragine, Legenda aurea, tr. Jean de Vignay, fol. 209v, Dormition de la Vierge (© BnF)

111

112

17. Bonn, Landes-undUniversitätsbibliothek 526, Lancelot, fol. 409/420, Queste del saint Graal: Galaad prend un écu suspendu à une croix rouge (© Landes-und-Universitätsbibl. Bonn)

18. Bonn, Landes-und-Universitätsbibliothek 526, Queste del saint Graal, fol. 409/420, Rubrique (© Landes-und-Universitätsbibl. Bonn)

INDICATIONS ÉCRITES ET MODÈLES PICTURAUX

113

19. Olim Amsterdam, Bibliotheca Philosophica Hermetica 1, Estoire del saint Graal, fol. 113, Couronnement de Galaad, frère de Josephé (photo: Lancelot-Grail Project)

20. Londres, BL Add 10294, Queste del saint Graal, fol. 68, Combat entre Lancelot et un chevalier (© British Library Board)

114

voir conserver cette note si lisible; on attribue sa préservation à un manque de soin dans un atelier où la production était faite à la hâte. Dans la Légende dorée HM 3027 les notes écrites sont en latin, la langue du texte. Ce manuscrit, dont les enluminures n’ont pas encore fait l’objet d’une étude détaillée, semble se placer au début de la tradition illustrée de la Légende dorée.25 Les notes sont écrites à l’encre tandis que les esquisses sont faites à la pointe sèche et sont relativement bien conservées dans l’ensemble du manuscrit. Les notes font mention des noms de saints, saint Gilles par exemple, dont l’incident de la protection du cerf fait l’objet d’une miniature (fig. 13), ce qui fait penser que les notes dépendraient d’une liste intermédiaire faite au moment de l’établissement du programme d’illustration de ce texte; la présence des esquisses indiquerait que l’artiste avait encore besoin d’aidesmémoire pour la disposition des éléments de la composition dans l’espace de la miniature. On remarque, dans la miniature de saint Gilles, qu’il a changé la position du roi avec son cor et de l’archet, qui sont placés à l’envers dans l’esquisse. Peut-on conclure à la participation de plusieurs personnes dans la création de cette miniature et du cycle dont elle fait partie ? L’emploi de l’encre pour les notes et de la pointe sèche pour les esquisses semble indiquer, sinon de différents participants, au moins deux campagnes de travail, sans compter celle de l’écriture de la miniature, du texte et de la rubrique qui l’accompagne. Revenons plus tard à cette dernière. Dans l’ensemble, l’emploi et la conservation d’esquisses et de notes dans les Lancelot paraît dépendre de la pratique différente de chaque atelier. Les esquisses y sont relativement rares, les notes sont plus souvent conservées.26 Le ms. Morgan 805 contient quelques notes, écrites, comme celles de la Légende dorée, à l’encre, et qui ont été préservées dans les marges tout près du bord de la page; plusieurs en ont été en partie découpées, révélant la pratique qui a dû être très générale d’éliminer les notes une fois qu’elles avaient été

25 Nous ne connaissons pas d’étude sur les légendes dorées illustrées (voir maintenant Fleith, Legenda aurea; Easton, Huntington Library Legenda Aurea). D’après les recherches que nous avons pu faire jusqu’à présent, ce manuscrit nous semble se placer tôt, sinon au début, de la tradition illustrée. Il a été attribué à Paris vers 1300 par Preston, Aspects of Medieval England, n° 23. 26 Le ms. Bonn 526 contient une esquisse à la mine de plomb au fol. 409 v; au fol. 433 paraît une tête de femme dessinée dans l’encre rouge des initiales fleuronnées de la même page, dans le style de l’artiste des miniatures. Cet artiste a-t-il pu être l’auteur des initiales fleuronnées ? Pour ce rapport voir Stones, Minnesota Vincent, p. 14.

INDICATIONS ÉCRITES ET MODÈLES PICTURAUX

115

utilisées.27 Au fol. 50 (fig. 14), la note se lit: « .i. chevalier a unes armes blenches qui esgardera .i. chastel sor une rivière. » On remarque l’emploi de l’impératif, qui rappelle l’usage de l’Itala de Quedlinbourg, et l’emploi de topoi, qui s’en distingue, mais qui se retrouve dans d’autres exemples tel le Missel d’Arras cité ci-dessus. On remarque également la mention des armes blanches, indication de l’importance croissante à cette époque de l’héraldique des chevaliers de la table ronde,28 mais rappelant en même temps la pratique, documentée dès le XIIe siècle, d’inclure, dans l’espace de la miniature ou à côté d’elle, une notation pour les couleurs.29 Déjà au XIIe siècle les notations pour les couleurs sont écrites en latin et en français; on voit un curieux mélange de langues dans l’aviaire et bestiaire au Jean-Paul Getty Museum, ms. Ludwig XV 3, c. 1270, où les couleurs sont indiquées parfois par des mots: « wit », « vermilien » (vermillon), et parfois par les lettres: « m » pour « minium » (rouge), « r » pour « rose » ou « rosa », « sa » pour « sable » (noir) (fig. 15).30 Dans ce manuscrit, les notations ou symboles pour les couleurs s’entrevoient sous les couches légères de la peinture; ils sont faits à l’encre. 27 Celles-ci se trouvent aux folios 48v, 49v, 50, 121, 123v, 155; elles ont été signalées par Cockerell dans Twenty Manuscripts, n° 88, p. 94–116. Nous remarquons que les notes sont écrites en littera cursiva dans une encre moins foncée que celle qu’a employée le copiste du texte; les lettres guides pour les initiales fleuronnées, par contre, sont en littera textualis et dans la même encre que le texte. On ne voudrait pas exclure la possibilité que les notes aient été quand même écrites par le copiste du texte, puisqu’un copiste était capable d’écrire dans plusieurs styles différents, comme l’a démontré van Dijk, Pattern Sheet. Il est pourtant rare de trouver une note écrite dans le même style que le texte. Voir le Lancelot de Kraus (olim Amsterdam), cf. infra. 28 Pour une table de l’héraldique des chevaliers de la table ronde dans les manuscrits de la Queste et de la Mort Artu avant 1340 voir Stones, Illustration of Lancelot, p. 359–384. Pour les XIVe et XVe siècles voir Pastoureau, Armorial. 29 Stirnemann, Nouvelles pratiques. Nous trouvons dans les marges à côté de plusieurs miniatures des mss. fr 110 et Bonn 526 des marques sous formes de traits diagonaux du genre que présente Madame Stirnemann, mais nous n’avons pas pu reconstruire le système selon lequel elles ont été utilisées dans ces derniers manuscrits. 30 Von Euw and Plotzek, Sammlung Ludwig, 4. Nous remercions les auteurs de ce catalogue d’avoir eu la gentillesse de nous montrer leur texte inédit et de discuter ce manuscrit avec nous. Le système de symboles qu’on a utilisé dans ce manuscrit est très difficile à déchiffrer puisqu’il ne semble pas toujours avoir de rapport entre le symbole et la couleur mise au-dessus. « Wi » ou «wit» paraît signifier « blanc » au fol. 3 (coulombe blanche, «wi» à côté), et au fol. 102 (deux moutons blancs, «wit» sur les dos des moutons); mais au fol. 88 on voit « wit » avec « sa » (?) sur un chien marron et au fol. 90 v « wit » paraît sous la mer verte, au-dessus d’un poisson bleu-gris. « Vermlien » paraît sur un loup marron au fol. 83 v; « m » =

116

Finalement au XIVe siècle, les notations sont surtout écrites en langue vulgaire, par exemple le ms. fr 241, de la Bibliothèque nationale, écrit en 1348, et enluminé dans un très grand atelier parisien dont au moins soixante manuscrits survivent, contient une série de notes, à côté de presque toutes les miniatures, qui indiquent les couleurs de l’arrière-plan: « o » pour « or », « r » pour un diapré rose, « a » pour un diapré azur (fig. 16). 31 Reprenons la question des rapports entre les miniatures, les directives, et les rubriques. Le Lancelot M.805 montre que les rubriques n’ont pas toujours été utilisées dans la tradition du Lancelot. Le lecteur doit avoir lu ou écouté la lecture du texte en entier, car c’est du texte total que dépend sa compréhension des miniatures. D’autres Lancelot montrent qu’il s’est créé un réseau compliqué entre le texte, la rubrique, la note de la rubrique, la note de l’enluminure et l’enluminure. Parfois, le fait que l’artiste se soit basé sur des indications écrites plutôt que sur des modèles visuels aboutit à des erreurs dans la miniature. Un folio du Lancelot de Bonn montre une miniature dans laquelle on voit un écu blanc accroché à une croix rouge et un chevalier occupé, soit à décrocher l’écu de la croix, soit à l’y accrocher (fig. 17). Cette scène et la section suivante du texte de la Queste sont décrites et résumées dans la rubrique écrite en rouge au-dessus de la miniature: « coment Galaad le fils Lanselot pent .i. escu blanc a une crois vermeille que onques hom ne pot pendre a son col et il le pendi el sien. » (fig. 18) La première partie de la rubrique décrit la scène dans la miniature (« coment Galaad ... pent i escu blanc ... ») tandis que la seconde partie explique le sens de l’épisode («que onques hom ne pot pendre a son col ») et fournit un sommaire des événements suivants (« et il le pendi el sien »). Cette rubrique est écrite plutôt pour le lecteur que pour l’artiste, « minium » sur un renard rouge et sous les vêtements d’un homme au fol. 83; « sab » = « sable » paraît également à côté d’une coulombe noire au fol. 3 et sur un hydre au fol. 85 v. On croit entrevoir « o » = « ocre » au fol. 102 sous des pierres ocre foncé. Autrement « r » = « rose » est courant, aux fol. 18. 28 v, 32 v, 36 v, 38v, 43 v, 88, 91. On comprend moins facilement «g» sous une couche rose de peinture aux fol. 27 v, 30 v, 38, 40 v, « a » également sous rose aux fol. 32 et 34 v; « vsb » au fol. 3; « go » (?) et « p » sous la mer verte au fol. 90 v. Signalons également dans la marge du fol. 88 un petit dessin à l’encre qui paraît être une esquisse pour le motif décoratif du fond colorié de la miniature. Parmi ces mots français ou latin, c’est le «wit» qui frappe. On pense bien qu’il s’agit de « wit », qui indiquerait la participation d’un anglais, plutôt que «wis» qui rapporterait à un flamand. 31 Pour une liste préliminaire de manuscrits de cet atelier, voir Illustration of Lancelot, pp. 283–286. L’emploi de directives dans cet atelier est très courant; ibid., p. 303. Voir Ross, Methods of book-production. in a fourteenth century French miscellany (London, British Museum, Royal 19.D.i), dans Scriptorium, 6, 1952, p. 63–75.

INDICATIONS ÉCRITES ET MODÈLES PICTURAUX

117

mais dans ce cas particulier la rubrique comme la miniature donnent une mauvaise représentation de l’événement. Ce que cette scène devrait montrer, c’est Galaad prenant un écu dont l’emblème est une croix rouge.32 Cet écu joue un rôle central dans la Queste: c’est Josephé fils de Josèphe d’Arimathie qui a mis la croix rouge sur lecu et seul un parfait chevalier peut le porter; tous les autres sont impuissants à l’enlever de l’endroit où il est gardé. Dans la Queste, l’écu est destiné à Galaad et il devrait être montré occupé à le prendre et à le porter. L’artiste ne connaissait pas l’histoire; celui qui a écrit la rubrique ne paraît pas l’avoir connu non plus, ou bien il n’a pas fait attention en copiant la rubrique. Il nous semble que l’erreur faite dans la représentation de la scène est liée à deux mots de la rubrique, l’un mal copié ou mal transcrit et l’autre simplement ambigu. Il est encore plus probable que la faute vienne d’une note en marge, petite, écrite en cursive avec des abréviations, et difficile à lire. Elle a dû être effacée ou coupée par la suite. Dans la rubrique, le mot « pent » («coment Galaad pent...») est une erreur. Nous suggérons qu’on s’est trompé en transcrivant une abréviation « pnt », qui devait s’écrire » prent ». Il y a aussi la phrase « a une crois vermeille » qui peut prêter à confusion. L’artiste l’a interprétée comme modifiant le verbe, «pendre a», plutôt que qualifiant le nom, «un escu a une crois vermeille». Si quelqu’un se base sur les mots de la rubrique, sans lire le texte ou sans regarder une autre illustration de cette scène, les deux interprétations deviennent également valables. Cette erreur montre clairement que l‘artiste s’est basé sur les mots plutôt que sur la lecture ou le souvenir qu’il avait du texte, et qu’il n’a pas utilisé un autre exemplaire illustré du même texte.33 A part une notation pour la couleur, les notes pour l’enlumineur et pour le rubricateur n’ont pas été conservées dans le ms. Bonn 526.34 Par contre, les exemples préservés dans le Lancelot de Kraus (olim Amsterdam BPH 1) et des autres Lancelot venant du même atelier flamand, révèlent des liens très proches entre la rubrique et la note pour la rubrique, qui se distinguent des notes destinées à l’artiste. Au fol. 113 du ms. Kraus se présente un cas particulièrement lisible: la miniature montre le couronnement de Galaad Pauphilet, Queste, p. 26; Stones, Illustration of Lancelot, p. 314. Le ms. fr 110 de la BnF contient encore des notes pour le rubricateur aux fol. 57, 57, 73v, 81v, 87v, 93v, 97, 101v, raison de plus de les reconstruire dans Bonn 526, fait dans le même atelier. 34 Au fol. 205 on lit « blanc » à côté d’une miniature qui montre le combat du chevalier blanc. 32 33

118

par son frère Josephe (épisode de l’Estoire); Josephe s’accompagne par deux clercs à gauche et par deux anges à droite (fig. 19). 35 La rubrique reprend exactement les mots de la note, écrite à la mine de plomb, au fond de la marge: « Chi couronne iosephus galaad son frere & le fait roy ». L’autre note, écrite également à la mine de plomb, semble avoir été écrite soit par un copiste différent, soit à un autre moment de la production de ce livre, car l’écriture est plus grande et moins serrée que celle de l’autre note. Elle se lit: « Ensi quns evesques sacre la couronne... un roy en un faudesteef. » On reconnaît ici l’emploi de topoi comme dans le Lancelot de la Bibliothèque Morgan, mais ce qui frappe en même temps est le manque de mention des autres personnages qui se trouvent dans cette miniature, les deux clercs et les deux anges, ainsi que les détails: un ange est thurifère, l’autre tient une aiguière, tandis qu’un des clercs tient un livre. D’autre part, le roi est assis sur un trône de type généralisé, non sur le « faudesteef » spécifié dans la note, et la couronne est déjà sur la tête de Galaad. Ces changements, ainsi que l’inclusion des autres personnages, sont-ils dus à l’esprit individuel de cet artiste, qui se donnait la liberté de ne pas suivre les directives d’un de ses collègues ? Les liens entre le ms. Kraus et les deux autres Lancelot enluminés dans le même atelier sont très compliqués. Il y a de nombreuses variantes entre les trois manuscrits à cet endroit dans le texte et dans le cycle de miniatures; on comprend facilement pourquoi le rubricateur et l’artiste ont eu besoin de directives. Le ms. Add ne contient pas de miniature à cet endroit, tandis que le ms. Roy la place à un autre endroit dans le texte,36 et présente quelques différences dans le traitement: un clerc tient la couronne qui n’a pas encore été mise sur la tête de Galaad, les anges manquent et un acolyte asperge. Le ms. Roy. contient dans l’ensemble très peu de rubriques; quelques lignes ont été laissées vides au-dessus de la miniature ici. Le manuscrit contient Sommer, Vulgate Version, I, pp. 282/33. Fol. 83v: ibid., p. 281/9. L’étude définitive de l’Estoire enluminée est encore à faire; nous remarquons que l’emplacement de cette scène dans Roy se retrouve dans Bonn 526, fol. 56 v (Josephé couronne Galaad assis), mais le plus souvent on rencontre un épisode différent à cet emplacement: les mss. fr 749, fol. 116v, ms. fr 24394, fol. 102, ms. fr 770, fol. 115v de la BnF montrent tous Josephé avec Pierron et Pharain; cette dernière scène se trouve également dans le ms. fr 95, fol. 106 v dans le registre supérieur de la miniature, avec le couronnement de Galaad agenouillé dans la partie inférieure. Dans le ms. 255 de Rennes, fol. 96, la scène du couronnement se place à un endroit différent dans le texte: Sommer, Vulgate Version, p. 284/21. Le traitement de la scène y est très simple: Josephé bénit Galaad, couronné, assis sur un trône; reproduction dans Stones, The earliest, pl. 7a, repris dans ces essais. 35 36

INDICATIONS ÉCRITES ET MODÈLES PICTURAUX

119

également de nombreuses traces de notes dans la marge inférieure de chaque folio enluminé, mais qui ont été soigneusement effacées, de sorte qu’elles ne soient guère lisibles. Il est impossible de déterminer si ces notes ont été adressées à l’artiste ou au rubricateur ou à tous les deux, comme dans le ms. Kraus. Dans le ms. Add les notes sont un peu plus lisibles mais il est difficile de trouver un exemple où l’on peut faire la comparaison entre la rubrique, les notes pour la rubrique et les notes pour la miniature. Une comparaison entre la même scène avec notes et rubriques dans les trois Lancelot qui proviennent de cet atelier serait intéressante à faire, mais un exemple nous échappe. Un examen plus attentif en révélera peut-être un cas. Pour le moment il suffira de signaler que la même pratique d’écrire les notes à la mine de plomb dans les marges inférieures se voit régulièrement dans les trois manuscrits et suggère qu’il s’agit de la pratique normale de l’atelier, au moins pour les Lancelot (fig. 20). 37 Ce n’est pas facile de tirer des conclusions générales à partir de ces exemples.Trop souvent les manuscrits ne donnent que quelques indications sur les méthodes de production, parmi un grand nombre d’illustrations. Il est évident que le grattage ou la rognure ont fait disparaître beaucoup – sinon la plus grande partie – des indications, dont seulement quelques exemples ont survécu. De ce fait, on ne peut pas toujours savoir dans quelle mesure les différentes pratiques étaient répandues dans les différents ateliers; quelquefois on croit, toutefois, pouvoir remarquer certaines préférences dans les multiples exemplaires d’un même texte avec un cycle de miniatures similaires. Les notes et les esquisses semblent avoir été utilisées surtout dans certaines circonstances: au moment de la création d’un nouveau cycle de miniatures et quand il s’agissait d’illustrer une longue série de scènes où il y avait des questions d’emplacement, de contenu et souvent d’explication dans la rubrique. La présence des notes et des esquisses témoignent de la production rapide et une attitude d’insouciance envers le produit final, qui Nous n’avons pas retrouvé d’exemples comparables dans les manuscrits liturgiques de cet atelier, ni dans ceux des autres ateliers cités ci-dessus, mais nous ne voulons pas trop insister sur le manque de notes dans les manuscrits liturgiques ou quasi-liturgiques. Nous avons déjà signalé celles qui paraissent dans le pontifical de Cambrai et le missel d’Arras cités ci-dessus. Nous signalons en outre la présence de notes dans la bible de Winchester et la bible Puiset, que nous connaissons grâce à P. Stirnemann, et dans la bible moralisée, British Library Harl. 1527. 37

120

se manifestent surtout dans les textes en langue vulgaire et dans les œuvres de qualité inférieure. Elles témoignent d’autre part de la complexité du procédé de création artistique en révélant quelques-unes des étapes intermédiaires entre le texte et la miniature; elles suggèrent même la possibilité que ces étapes aient été dirigées ou peut-être complétées par une ou plusieurs personnes intermédiaires entre le copiste et l’artiste. Elles nous aident ainsi à voir un peu plus clairement comment a fonctionné l’atelier dans une époque où les documents nous fournissent peu de renseignements sur cette question.

V Fabrication et illustration des manuscrits arthuriens*

I

l n’existe pas à proprement parler de production spécifique de manuscrits arthuriens. Les romans de la Table ronde sont copiés et enluminés dans les mêmes circuits que les autres manuscrits contemporains: on peut d’abord supposer la présence de copistes et de peintres auprès des comtes de Champagne ou des Plantagenêt, à la fin du XIIe siècle. Puis des centres de production se développent, principalement dans le Nord du royaume de France et à Paris. Des artistes, sans doute itinérants, y réalisent des manuscrits où cohabitent chansons de geste, romans antiques, textes arthuriens, psautiers ou livres d’heures, ouvrages de piété, à destination d’un public laïque ou religieux. À la fin du Moyen Âge cohabitent deux types de production: des manuscrits commandés par de grands mécènes, comme Jean duc de Berry ou Jacques d’Armagnac, et une production plus courante, mais souvent de qualité, que des libraires plus ou moins spécialisés vendent à une clientèle variée de nobles et de bourgeois, mais également à de grands personnages. Pour autant, le nombre très élevé des manuscrits arthuriens qui subsistent aujourd’hui – signe parmi d’autres d’une diffusion considérable – et la richesse de leur tradition textuelle et iconographique permettent de dégager certaines particularités par rapport au reste de la production livresque du temps. Les débuts de l’illustration arthurienne sont difficiles à établir avec certitude. Du point de vue iconographique, on constate que le décor

*Publié dans le catalogue de l’exposition La Légende du roi Arthur, éd. T. Delcourt, Paris, 2009, pp. 19–29.

122

architectural l’emporte d’abord nettement sur l’enluminure et que les premiers témoins de la matière de Bretagne se trouvent sculptés loin des lieux d’origine des légendes. Citons à titre d’exemples les scènes du combat de Tristan représentées dès le début du XIIe siècle sur l’une des colonnes du portail nord de la cathédrale de Saint-Jacques-de-Compostelle; encore en Galice, Yvain et son lion sculptés sur le tympan de l’église de Santa María de Peñamaior; ou bien les scènes arthuriennes déclinées sur le tympan du portail nord de la cathédrale de Modène. Il faut également évoquer le cas du roi Arthur représenté, chevauchant un bouc, face à Alexandre le Grand, dans la mosaïque de la cathédrale d’Otrante. Le domaine anglo-normand C’est dans l’aire anglo-normande que l’on trouve les premiers témoins de l’enluminure arthurienne, avec le portrait d’Arthur réalisé vers 1150 au Mont-Saint-Michel (cat. 10) ou celui d’un personnage (Tristan ou Iseut ?) jouant de la harpe dans l’unique initiale historiée du fragment du Tristan de Thomas (Oxford, Bodl., ms. French d. 16, f. 10, vers 1200). Pourtant, les manuscrits arthuriens illustrés par la suite dans les îles Britanniques sont peu nombreux. On ne peut même pas parler d’une tradition illustrative, car chaque manuscrit constitue un cas spécial et isolé. Ainsi, les bifeuillets présentant des scènes du roman de Tristan insérés dans le manuscrit Londres, BL, Add. 11619 (cat.76) contiennent des dessins à l’encre rehaussés d’un lavis de couleur disposés en pleine page (format peu courant dans l’illustration arthurienne), sans rapport avec le reste du manuscrit. Apparentés stylistiquement à l’école de Matthew Paris, ces bifeuillets restent néanmoins sans parallèles parmi les chroniques et les vies de saints produites dans le cercle de Matthew. Le Lancelot-Graal a également laissé quelques témoins illustrés en Angleterre aux environs de 1275, notamment les manuscrits BnF, fr 123 (cat. 84) et Londres, BL, Royal 20 C.VI, où l’on remarque des choix iconographiques insolites. En tête de la Mort Artu du BnF, fr 123 (f. 229), on voit une reine (Guenièvre ou Aliénor d’Aquitaine) à côté d’un roi (Henri II ou Arthur) donnant des ordres à un copiste assis à ses pieds. On peut supposer que la présence de la reine à côté du roi fait allusion au couple commanditaire, car cet exemplaire fut confectionné pour le mariage de Blanche d’Artois et d’Edmond de Lancaster. Dans Londres, BL, Royal 20 C.VI, l’unique illustration de la Mort Artu (f. 150) s’explique moins

FABRICATION ET ILLUSTRATION

123

facilement: Guenièvre va être soumise à l’épreuve du feu – choix d’en-tête sans équivalent qui souligne la part qui revient à la reine dans l’effondrement à venir du royaume d’Arthur. Peut-on supposer que cet exemplaire était destiné au futur roi Édouard Ier, puisque le peintre a également été responsable de l’initiale historiée en tête d’un manuscrit de l’Apocalypse confectionné pour le même Édouard alors qu’il était encore prince de Galles (Oxford, Bodl., Douce 180)? Deux manuscrits anglais du début du XIVe siècle se font remarquer par leurs illustrations arthuriennes: la Chronique de Langtoft jusqu’en 1307 (Londres, BL, Royal 20 A.II), où figurent Vortiger brûlé dans son château et le roi Arthur avec un écu orné d’une image de la Vierge à l’Enfant, et le Brut dans la version unique transmise par le manuscrit Londres, BL, Egerton 3028 (cat. 15). Copié peu après le décès du roi Édouard II, en 1327, ce dernier est très amplement illustré, là encore par des dessins rehaussés de lavis, et le Brut y fait pendant au roman de Fierabras, mettant ainsi en parallèle les rois d’Angleterre et Charlemagne avec ses descendants. L’illustration des romans en vers Les débuts de l’illustration arthurienne en France remontent au début du XIIIe siècle mais ils ne correspondent pas du tout à ce que l’on sait sur l’évolution des traditions littéraires. Hormis les fragments d’un recueil champenois aujourd’hui conservés en Ardèche et qui peuvent être datés de 1200 environ («Fragments d’Annonay»), les plus anciens manuscrits de Chrétien de Troyes ne datent que d’une génération après la disparition du poète. Ils restent sans illustration jusqu’au dernier quart du XIIIe siècle, à deux exceptions près: on attribue au deuxième quart du siècle le célèbre portrait de Marie de Champagne qui figure en tête du Lancelot dans le recueil champenois copié par Guiot (BNF, fr 794, f. 27, cat. 23), ainsi que l’initiale fort abîmée représentant Perceval à cheval dans le recueil conservé à la Burgerbibliothek de Berne, ms. 354. Ce n’est qu’après 1275 que l’on trouve des manuscrits de Chrétien de Troyes contenant des illustrations en série – encore s’agit-il surtout de Perceval et des Continuations (BnF, fr 12576, cat. 25; Montpellier, BIU, Section médecine H 249; Mons, BUMH, 331/206; BnF, fr 1453 et fr 12577, cat. 24). Les provenances sont souvent difficiles à établir: les illustrations de BnF, fr 1453 et fr 12577 sont dues à des peintres parisiens, alors que le manuscrit de Mons est vraisemblablement tournaisien et que BnF, fr 12576

124

et le manuscrit de Montpellier émanent probablement du Nord du royaume de France, malgré quelques éléments parisiens. L’illustration des autres textes de Chrétien est beaucoup plus limitée. Pour Yvain, on ne connaît que deux manuscrits contenant une série d’images, Princeton, PUL, Garrett 125 (attribué à Amiens) et BnF, fr 1433 (Nord-Est, cat. 27), plus un manuscrit présentant une seule initiale historiée, BnF, fr 12603 (copié à Arras). Pour Lancelot, un manuscrit illustré (là aussi PUL, Garrett 125). Pour Érec et Énide, un manuscrit contenant une initiale historiée (BnF, fr 1376, attribué à Dijon) et un exemplaire très amplement illustré par deux peintres, l’un d’Arras, l’autre peut-être de Thérouanne (BnF, fr 24403). Enfin, pour Cligès, aucune illustration, à part une miniature du Roman de la Poire (BnF, fr 2186, f. 3 v°, cat. 79). On remarque par ailleurs que les romans de Chrétien de Troyes sont accompagnés le plus souvent d’autres textes, ce qui indique une prédilection des commanditaires pour des recueils à contenu varié. Le manuscrit récemment acquis par l’université de Nottingham, ms. WLC Lm 6 (olim Mi LM 6) constitue une exception précoce. Il s’agit d’un vaste ensemble incluant le Roman deTroie, llle et Galeron, le Roman d’Alexandre en vers, la Chanson d’Aspremont, des fabliaux et deux textes arthuriens: la Vengeance Radiguel et le Roman de Silence. Son importance a été reconnue par Terry Nixon, qui y voit à juste titre la plus ancienne compilation française illustrée, alors que la plupart des critiques l’avaient placée beaucoup plus tard dans le XIIIe siècle. L’illustration consiste en une série de petites miniatures de la largeur d’une moitié de colonne de texte, peintes sur fond d’or mat et encadrées de bordures colorées orange et blanc. Malgré leur caractère simplifié et stéréotypé, le traitement des personnages est vivant, avec les jambes des chevaux et les lances sortant de l’encadrement, les gestes expressifs et le tombé des draperies bien modelé. La gamme des couleurs s’appuie surtout sur l’orange, le bleu clair, le vert clair et le blanc. Tout ceci rappelle les peintures religieuses beaucoup plus monumentales des artistes flamands qui travaillèrent, à la fin du XIIe ou au début du XIIIe siècle, pour le scriptorium du monastère de Saint-Bertin à Saint-Omer en Flandre. La production des manuscrits en prose Si certains noms de copistes et de commanditaires nous ont été transmis dans les manuscrits mêmes, la plupart des textes ne contiennent pas de tels

FABRICATION ET ILLUSTRATION

125

renseignements. C’est donc sur des critères stylistiques de l’illustration et par comparaison avec d’autres productions que nous arrivons à proposer quelques attributions géographiques et chronologiques. Les caractéristiques de l’écriture et du décor fournissent aussi des données utiles, mais restent des guides d’attribution moins sûrs que les illustrations. Toujours est-il que, à l’exception de ce qui se passe à Paris, nous sommes mal renseignés sur le mouvement des artistes avant le XVe siècle et sur leurs conditions de travail, de sorte que les attributions géographiques sont incertaines. Le Robert de Boron de Modène (Bibl. Estense, alpha E 39) remonte peut-être très haut dans le XIIIe siècle, mais ses rapports avec la production livresque de l’époque sont difficiles à cerner: si l’écriture est beaucoup plus élégante que celle du recueil de Nottingham, les initiales historiées sur fond d’or, avec des bordures à antennes sur fond géométrique, ne se laissent pas classifier avec certitude et on a du mal à relier texte et scènes représentées, à quelques exceptions près, comme le Christ aux enfers pour le début du Merlin, sujet très courant par la suite à cet emplacement. Les initiales historiées aux personnages imposants et expressifs qu’on trouve dans le Lancelot-Graal de Rennes (BM, 255) rapprochent ce manuscrit des psautiers royaux et des Bibles moralisées des environs de 1220. Là encore, l’iconographie est remplie de détails insolites, tels Mordrain transporté dans son lit vers l’île déserte ou l’initiale en tête du Lancelot qui montre Aramont, seigneur de Bretagne, devenant vassal d’Uterpendragon et embarquant avec lui pour aller combattre le roi Claudas en Gaule. Dès le milieu du XIIIe siècle, la production de manuscrits arthuriens est nettement dominée par les deux grands cycles du Lancelot-Graal puis du Tristan en prose, même si d’autres romans antérieurs, en vers ou en prose, continuent d’être copiés jusqu’au XVe siècle. Les grands textes en prose plus récents, Guiron le Courtois ou Perceforest, ne connaîtront jamais la même diffusion. Si l’on omet les rédactions particulières comme la Queste dite PostVulgate lombarde (BnF, fr 343, cat. 87) ou les adaptations faites pour Jacques d’Armagnac (BnF, fr 112, cat. 6 et 95) ou pour Jean-Louis de Savoie (Bruxelles, KBR, 9246 et BNF, fr 91, cat. 96 et 97), les manuscrits du Lancelot-Graal se divisent en deux versions principales, l’une longue et l’autre courte, qui se transmettent côte à côte pendant trois siècles. Les images jouent un rôle important dans cette transmission et dans l’accueil que leur réserve le public, quoiqu’il existe aussi de nombreux exemplaires dans lesquels la décoration se limite aux initiales coloriées ou filigranées, ou à une

126

illustration unique en début de manuscrit, par exemple BnF, fr 768. Flandres, Hainaut et Artois Le Nord de la France a été un centre actif de production de manuscrits arthuriens, sans doute parce que l’aristocratie locale s’intéressait particulièrement à ce genre littéraire. Mais, comme toujours, les artistes ne se sont pas cantonnés aux seuls romans courtois. C’est probablement à Saint-Omer que furent produits les deux manuscrits jumeaux BnF, fr 19162 et 24394, Estoire, Merlin, que l’on peut rapprocher stylistiquement d’un psautier à l’usage de Saint-Omer copié en 1277, Londres, BL, Yates Thompson 43. Les manuscrits Le Mans, MM, 354 (Estoire, cat. 3), BnF, fr 770 (Estoire, Merlin et d’autres textes, cat. 34) et Oxford, Bodl., Digby 223 (Agravain, Queste, Mort Artu) ont été tous illustrés par un peintre actif dans la région de Douai, qui a enluminé aussi un martyrologe et un obituaire pour les cisterciennes de Notre-Dame-des-Prés (Valenciennes, BM, 838) et un psautier-livre d’heures à l’usage de l’abbaye de Saint-Amé (Bruxelles, KBR, 9391); il a également enluminé les vies de saints aujourd’hui conservées à Saint-Pétersbourg (Bibliothèque de l’Académie des sciences, F 403) et plusieurs exemplaires du Trésor de Brunet Latin et du Roman des sept sages. D’autres Lancelot-Graal se rattachent de moins près à son œuvre: Oxford, Bodl., Douce 303 et BnF, fr 342, copié en 1274. Le copiste du Mans 354 a signé «Walterus de Kayo». On suppose que sa ville natale était Cayeux-sur-Mer. Un deuxième fils de ce même pays, Arnulfus de Kayo, a enluminé en 1286, à Amiens, un manuscrit complet du Lancelot-Graal (Bonn, LUB, 526). Étaient-ils apparentés? On ne saurait le dire car l’écriture des manuscrits n’est semblable qu’en termes généraux et leurs collaborateurs ne sont pas les mêmes. On voudrait savoir ce que signifie la présence à Amiens d’Arnoul de Cayeux, car la miniature amiénoise n’offre pas de parallèles stylistiques pour le Lancelot-Graal de Bonn. C’est plutôt à Cambrai ou à Saint-Omer et Thérouanne que se situent les manuscrits apparentés: des missels à l’usage de Cambrai (Cambrai, MM, 153 et 154), des Bibles (Saint-Omer, BM, 5 et New York, Morgan, M.79 et M.969), et encore un recueil d’Hugues de Fouilloy (Los Angeles, Getty, Ludwig XV 4). Un autre Lancelot-Graal complet, qui date probablement d’une dizaine d’années plus tard, lui est aussi apparenté (BnF, fr 110), tout comme les Tristan en prose Londres, BL, Add. 5474 et BnF, naf 6579 (cat. 74) et une

FABRICATION ET ILLUSTRATION

127

partie du recueil contenant Êrec et Énide dont il a déjà été question, BnF, fr 24403. On invoque le mécénat probable d’un membre de la famille des comtes de Flandre pour une autre paire de manuscrits provenant du diocèse de Thérouanne, BnF, fr 95 (Estoire, Merlin, cat. 3) et New Haven, Yale University, Beinecke, 229 (Agravain, Queste, Mort Artu) dont l’artiste principal réapparaît dans le psautier BnF, latin 1076 et son proche parent, le livre d’heures à l’usage de Thérouanne, Marseille, BM, 111. Un très grand nombre d’autres romans, de livres juridiques, scientifiques et de dévotion leur sont apparentés. Un sous-groupe arthurien peint par un artiste inférieur reflète ce même courant stylistique: Oxford, Bodl., Ash. 828 (Lancelot); BnF, fr 749 (Estoire, Merlin, cat. 63); Bologne, AS b.i.bis (Estoire, fragments); Turin, BNU, L.III.12 (Merlin). C’est encore vers Saint-Omer, ou peut-être Tournai ou Gand, qu’on doit se tourner pour placer Royal 14 E.III (Estoire, Queste, Mort Artu, cat. 35) et les deux triplets constitués d’une part par Londres, BL, Add 10292–4 (Lancelot-Graal), qui contient un tombeau gravé portant la date du 17 février 1317 ns, et d’autre part par olim Amsterdam, BPH 1, Manchester, JRUL, French 1 et Oxford, Bodl., Douce 215 (Lancelot-Graal). Le peintre principal était encore actif en 1323. On sait par les testaments que, dès 1300, les bourgeois de Tournai possédaient et léguaient leurs romans arthuriens, et bien des documents nous renseignent sur des possesseurs dont les manuscrits arthuriens n’ont pas survécu. La Queste du manuscrit 5218 de l’Arsenal, dans laquelle Pierart dou Thielt nous dit qu’il copia, enlumina et relia le manuscrit en 1351, est un cas exceptionnel (cat. 86). On peut se demander si ce manuscrit faisait partie des commandes faites par Gilles li Muisit, abbé de Saint-Martin de Tournai, pour qui Pierart enlumina les Chroniques composées par l’abbé lui-même. Le Lancelot BnF, fr 122, copié en 1345 ns (cat. 66), anticipe ce même courant stylistique alors que BnF, fr 1422–1424 (Agravain, Queste, Mort Artu, cat. 61) se rattache à un autre peintre de l’école de Tournai, dont on retrouve la main dans un Roman d’Alexandre, Oxford, Bodl., ms. 264. Aux manuscrits de Chrétien produits à Arras on peut ajouter deux Lancelot-Graal: Londres, BL, Royal 20 D.IV, illustré par le Maître au menton fuyant, et Darmstadt, ULB 2543, dont le peintre a également enluminé des romans de Tristan en prose, BnF, fr 758, fr 776 et Vienne, ÖNB, 2542. Au Maître de la Vie de sainte Benoîte d’Origny (Berlin, Kupferstichkab., 78 B16), commandée en 1312, sont dues les grandes miniatures du Lancelot

128

en prose New York, Morgan, M.805–806; toutefois la plupart des initiales historiées sont l’œuvre du peintre du missel de Saint-Remi de Reims (Reims, BM, 217). Un collaborateur du Maître de sainte Benoîte, qui travailla à ses côtés en 1311 dans le recueil Turin, BNU, L.II.14, fut luimême l’artiste du manuscrit dit Huth, Londres, BL, Add 38117 (Joseph, Merlin), et de BnF, fr 12573 (Agravain, Queste, Mort Artu). Ces peintres ont dû être actifs dans la région de Laon/Noyon/Saint-Quentin ou bien à Reims. Ces filiations complexes, ces interactions multiples, nous posent d’innombrables questions: on ne sait pas jusqu’à quel point les artistes ont été itinérants, ni si les clients cherchaient leurs livres dans des endroits éloignés de leurs résidences. Les autres régions du royaume et l’Italie D’autres régions sont moins riches en manuscrits arthuriens. On ne connaît aucun manuscrit arthurien émanant de Bretagne et peu de manuscrits normands: le manuscrit Didot, BnF, naf 4166, copié en 1301 (cat.32), et le manuscrit Berkeley, UCB, 106 (Estoire reliée avec les Vies des pères), dans lequel François Avril a discerné la présence d’un artiste connu par ses livres illustrés pour l’abbaye de Jumièges aux environs de 1275. Enfin le magnifique frontispice du Tristan abrégé, BnF, fr 103, est attribué à Rouen au troisième quart du XVe siècle (cat. 73). Dans l’Est, l’exemplaire du Lancelot-Graal complet entré dans la bibliothèque de Marguerite de Luxembourg (BnF, fr 344) et son jumeau partiel (Estoire, Merlin), aujourd’hui en mains privées, sont les témoins isolés d’une vaste entreprise messine qui a produit nombre de chroniques, livres de médecine, chartes, dans un style relativement rustique, à la fin du XIIIe siècle. Les productions méridionales sont aussi rares: en 1319 a été confectionnée à Avignon, dans l’entourage du pape Jean XXII, une copie de la Queste del saint Graal (Florence, BML, Ash. 121). D’autre part le Roman de Jaufré, en occitan, subsiste en deux exemplaires dont seul BnF, fr 2164 est illustré (cat. 58). Le Tristan en prose BnF, fr 104, qui ne contient qu’un décor d’initiales champies et filigranées, pourrait aussi être un manuscrit méridional. Un cas intéressant est le Tristan BnF, fr 750, copié en 1278 par un scribe normand, Pierre de Tiergeville, mais illustré en Italie du Sud ou en Terre sainte. En effet, dans les vingt dernières années du XIIIe et au début du XIVe siècle, le goût du Lancelot-Graal et du Tristan se répand en Italie.

FABRICATION ET ILLUSTRATION

129

Les exemples de manuscrits produits pour l’aristocratie lombarde, ligure ou vénitienne sont innombrables. On sait aussi que l’empereur et roi de Sicile Frédéric II possédait un manuscrit de Palamède (appelé aujourd’hui Guiron le Courtois): le royaume angevin de Naples sera ensuite un centre de production de manuscrits arthuriens, comme en témoigne le BnF, fr 756–757, qui fut confectionné probablement pour la famille Carafa de Naples. La production parisienne La production parisienne est tout aussi robuste que celle de Flandre et d’Artois. Le manuscrit BnF, fr 2455 (Estoire), mentionne ainsi le mécénat de Philippe Auguste (au f. 293), ce qui pourrait faire penser à une production précoce, du début du XIIIe siècle, mais il s’agit en fait d’une copie de la fin du siècle. Produits au milieu du XIIIe siècle, Bruxelles, KBR, 9627-9628 (Queste, Mort Artu), et l’ensemble formé par Chicago, Newberry Libr., f21Ry.3412261 (Lancelot), et Arsenal 3347 (Agravain, Queste, Mort Artu), sont certainement parisiens, comme Berkeley, UCB, 107 (Lancelot), et BnF, fr 339 (Lancelot, Queste, Mort Artu, cat. 69). Vers la fin du XIIIe siècle, le Maître hospitalier, actif en Terre sainte et à Paris, collabore avec un peintre méditerranéen dans Tours, BM, 951 (Estoire, Joseph, Merlin), alors qu’il travaille seul dans BnF, fr 12580 (Agravain, Queste, Mort Artu) et dans Londres, BL, Royal 20 D.II (Tristan). D’autres peintres sur lesquels nous sommes bien renseignés ont produit des manuscrits arthuriens: le Maître de Méliacin – BnF, fr 772 (Tristan); l’entourage des suiveurs de Maître Honoré – Saint-Pétersbourg, BNR, Fr. F.v.XV.5 (Estoire), et Oxford, Bodl., Rawlinson Q.b.6 (Lancelot, Queste, Mort Artu), copié par Ernoul d’Amiens; le Maître de Maubeuge – Vatican, BAV, Pal. lat. 1964 (Tristan); le Maître de Jean Papeleu – BnF, fr 333 (Agravain, cat. 64), BnF, fr 334 (Tristan), BnF, fr 1453 (Perceval et Continuations), déjà mentionné; le Maître de Fauvel et ses collaborateurs – BnF, fr 105 (Estoire, Merlin), et Arsenal, ms. 3481, BnF, fr 9123 (Estoire, Merlin, cat. 2); le Maître du libraire Richard de Montbaston – Los Angeles, Getty, Ludwig XV 5 (Tristan) – et ses successeurs – BnF, fr 16999 (Lancelot), et Arsenal, 3482, (Merlin, Agravain, Queste, Mort Artu). C’est à Paris que Jean de Berry a acheté un Lancelot-Graal complet (BnF, fr 117–120, cat. 1, 70 et 90) et qu’a été fait son jumeau Arsenal, ms. 3479– 3480 (cat. 4). Jean de Berry a possédé le Tristan aujourd’hui conservé à

130

Vienne, ÖNB, 2537 (vers 1410): un exemplaire similaire, illustré en dessins au lavis de couleur, BnF, fr 100–101 (cat. 72), a appartenu à Marguerite d’Écosse. On peut attribuer à Paris le Tristan, BnF, fr 335–336, copié en 1399, la Compilation de Rusticien de Pise (BnF, fr 340, cat. 60) et un autre Tristan de Vienne, ÖNB, 2539–2540. Au milieu du XVe siècle, le Maître de Dunois a enluminé un Lancelot-Graal complet, aujourd’hui dispersé, pour Prigent de Coëtivy, amiral de France. Enfin le milieu et la deuxième moitié du XVe siècle ont vu l’expansion de la production de manuscrits arthuriens vers le Centre et l’Ouest de la France, en relation avec le déplacement des centres du pouvoir dans le Val de Loire. On attribue au Maître d’Adélaïde de Savoie, actif à Angers ou Poitiers, le Lancelot-Graal BnF, fr 96 et au Maître de Charles du Maine, actif à Tours et Angers, l’ensemble formé par Chantilly Musée Condé 648 et Dijon, BM, 527 (cat. 88). Le Maître du Missel de Yale semble avoir été actif à Bourges ou à Tours; on lui doit l’enluminure de trois exemplaires du Tristan dont on ne connaît pas les commanditaires (New York, Morgan, M.41, daté de 1468, BnF, fr 102 et Genève, Bibliothèque de Genève, 189). Deux autres manuscrits du Lancelot-Graal conservés à la Pierpont Morgan Library sont attribuables au Centre de la France, au troisième quart du XVe siècle (M.207–208 et M.807). La production de manuscrits arthuriens culmine avec le plus célèbre bibliophile de l’époque, Jacques d’Armagnac, dont la bibliothèque est présentée en détail plus loin, et qui fit appel au peintre Evrard d’Espinques. Le grand artiste berruyer Jean Colombe illustra également quelques romans de la Table ronde (cat. 96 et 97). L’iconographie des manuscrits arthuriens Sur le plan iconographique, les manuscrits illustrés présentent énormément de variations, aussi bien dans le placement des images que dans leur sélection et leur traitement. La comparaison entre les Lancelot-Graal complets est parlante: avec 747 miniatures, le manuscrit Londres, BL, Add 10292–4, l’emporte nettement sur BnF, fr 344 (344 images), Bonn, ULB, 526 (346 miniatures) et son pendant stylistique BnF, fr 110 (99 miniatures), BnF, fr 117–120 (131 miniatures) et son jumeau Arsenal, ms. 3479–3480 (130 miniatures), ou encore BnF, fr 113–116 (209 miniatures). Par contre, les manuscrits incomplets sont souvent très copieusement illustrés. Certains thèmes du Lancelot-Graal sont très populaires: par exemple la Nef enchantée créée par Salomon, qui apparaît dans l’Estoire et de nouveau

FABRICATION ET ILLUSTRATION

131

dans la Queste, est présente dans la plupart des manuscrits illustrés, tout comme le premier baiser de Lancelot et de Guenièvre. D’autres sujets y figurent en revanche rarement, tels la saignée de la sœur de Perceval dans la Queste (représentée seulement dans Bonn, ULB, 526, Florence, BML, Ash. 121 et BnF, fr 343), ou bien la guérison de Lancelot lorsqu’on lui souffle une poudre dans le nez – sujet insolite dans Londres, BL, Add 10293 et là aussi lié à l’illustration médicale. Arthur perché sur la roue de Fortune à la fin de la Mort Artu n’est représenté que dans Add 10294, et on ne le voit découvrant l’adultère de Lancelot et Guenièvre dans la Chapelle aux images que dans deux manuscrits de Jacques d’Armagnac, BnF, fr 112 et 116. Parfois, on peut discerner une forte prédilection pour certains aspects du texte: par exemple, dans Londres, BL, Add 10292–10294, où de très nombreuses scènes portent sur les aspects juridiques du Lancelot-Graal, tout ce qui concerne l’usurpation territoriale de Lancelot et la légitimité du statut de Guenièvre doit manifestement rencontrer l’intérêt particulier d’un commanditaire grand feudataire ou juriste. Le traitement du Graal varie quant à lui considérablement selon les manuscrits: il apparaît sous la forme d’une écuelle, d’un ciboire couvert, d’un calice vide, tantôt visible, tantôt caché par une châsse, contenant une croix ou le personnage du Christ-enfant. Les études comparatives sur l’iconographie arthurienne n’en sont encore qu’à leurs débuts. La mise en ligne rapide des images devrait permettre sans doute d’expliquer les raisons des choix opérés par les artistes ou leurs commanditaires et leur signification culturelle.

Arthur triomphant de l’épreuve de l’épée 3. Lancelot-Graal (Estoire, Merlin, Suite Vulgate). Roman des Sept Sages. André, Pénitence Adam Thérouanne. vers 1290–1300. Parchemin. III + 394 + 3f. (2 col. de 40 lignes). 470 x 330 mm. Provenance: Bibliothèque des ducs de Milan, à Pavie; rapporté à Blois par Louis XII. BnF, Manuscrits. français 95 (f. 159 vo) Ce manuscrit a été depuis longtemps reconnu comme le plus somptueux roman arthurien du XIIIe siècle. Comme le manuscrit conservé à la Beinecke

132

Library de l’université de Yale (ms. 229, fin du Lancelot en prose, Quête du Saint Graal, Mort Arthur), qui pourrait être sa suite, il transmet la version longue du Lancelot-Graal et offre un nombre exceptionnel d’illustrations, réalisées par deux peintres. Le décor mineur s’étale sur chaque page, sous la forme d’initiales champies accompagnées de bordures dessinées en rouge et bleu à cinq feuilles et à motifs végétaux, sans parler des initiales historiées et des miniatures, dont la plupart, disposées sur deux registres, permettent de montrer une suite d’événements dans une seule image. Des antennes souples à terminaisons en forme d’hybrides ou d’animaux supportent de petits personnages, des animaux, des oiseaux et des hybrides qui complètent les scènes principales ou en présentent le contresens. L’emploi d’antennes contenant des motifs secondaires, qu’on retrouve également dans le manuscrit du Mans (MM, 354, cat. 33), est insolite dans les manuscrits du LancelotGraal, où ce type de décor est normalement limité aux frontispices. Cette miniature (fig. 1) est la dernière illustration de la branche de Merlin. Elle représente le moment du triomphe d’Arthur, seul capable de retirer l’épée de l’enclume. Il se montre ainsi digne de succéder à Uterpendragon comme roi de Bretagne. La partie supérieure de la miniature le montre retirant l’épée (pour la deuxième fois) devant l’évêque, le clergé et le peuple; dans la partie inférieure, il la place sur l’autel à côté d’un calice à moitié voilé, pendant qu’il est couronné et béni par l’évêque. L’illustration met l’accent à la fois sur la prouesse d’Arthur et sa légitimité en tant que roi chrétien accueilli par l’Église en présence des barons. Sur les terminaisons des antennes, deux grues tiennent des cruches et boivent dans des bols alors qu’en bas quatre dragons mordent les antennes. Peut-on y voir l’anticipation de la Quête du Saint Graal et de la chute du royaume d’Arthur dans la Mort Artu, ou s’agit-il simplement de motifs pris dans le répertoire décoratif des peintres? On peut attribuer cette peinture à la région de Thérouanne, car le même artiste a notamment enluminé un psautier-heures à l’usage de ce diocèse (BnF, latin 1076 et Marseille, BM, 111), vraisemblablement confectionnés avant la canonisation de saint Louis, en 1297. Les écus de la famille du comte de Flandre, discrètement distribués dans les marges de ce manuscrit et du 229 de Yale, laissent entrevoir le mécénat d’un membre de la famille comtale ou de son entourage. La présence d’autres textes dans un manuscrit du Lancelot-Graal n’est pas sans parallèle (voir par exemple Cologny, Bodmer, 147, cat. 8, et BnF, fr 770, cat. 34). Toutefois les Sept Sages commencent un nouveau cahier et la

FABRICATION ET ILLUSTRATION

1. Paris, BnF fr 95, Merlin, fol. 159v: Arthur retirant l’épée de l’enclume; Arthur plaçant l’épée sur l’autel et recevant la couronne (© BnF)

133

134

Pénitence Adam leur fait directement suite. Ils ont donc pu avoir été ajoutés après coup au Lancelot-Graal. On ne sait pas non plus si Yale 229 constitue la suite de ce manuscrit (après la lacune du début du Lancelot en prose proprement dit) ou une partie d’un autre manuscrit jumeau. Début de la MORT ARTU 8. Lancelot-Graal avec interpolations de passages de la Bible en français, de sermons de Maurice de Sully, du Roman de Troie et des Faits des Romains France (ouest de la Marne ?), vers 1300. Parchemin. 1 + 388 +1 f. (2 col de 49–51 lignes), 350 x 260 mm. Provenance: Antoney de Racygnano (XIVe siècle): bibliothèque de Sir Thomas Phillipps; acheté par Martin Bodmer en 1946. Cologny. Fondation Martin Bodmer 147 (f. 344) Ce manuscrit, copié par deux scribes, reproduit le cycle du Lancelot Graal, sauf le Lancelot en prose proprement dit, dont l’absence semble révéler un projet de concentration romanesque. Le premier scribe a pourtant étoffé sa matière en interpolant des extraits bibliques et des sermons de Maurice de Sully en français; il place aussi dans la bouche de Merlin une version isolée du Roman de Troie en prose et les Faits des Romains. Ces ajouts, que renforcent des leçons textuelles propres, infléchissent la lecture des romans et font de ce volume un témoin exceptionnel de la réception arthurienne. De même que le célèbre manuscrit BnF, fr 1450 (cat. 22) introduit les œuvres de Chrétien de Troyes dans la suite Troie, Énéas, Brut, inscrivant ainsi les aventures chevaleresques dans le sillage du monde antique, de même la composition du recueil Bodmer invite ostensiblement à lire l’histoire du Graal jusqu’à l’affrontement final entre Arthur et son fils Mordred à la lumière de la destruction de Troie et de la vie de César. Les Livres des Maccabées et de Judith ou les passages de la Genèse et des Évangiles sont quant à eux légitimés par le sujet de l’Estoire del saint Graal, qui raconte comment Joseph d’Arimathie recueillit le sang du Christ dans la coupe du Graal: leur présence affermit les parallèles entre les récits arthuriens et bibliques. Le manuscrit est très difficile à placer stylistiquement, malgré le nombre considérable d’illustrations (167), constituées de miniatures à plusieurs compartiments pour les ouvertures de branches, accompagnées de bordures à motifs variés et à armoiries et, pour le corps des textes, de plus petites

FABRICATION ET ILLUSTRATION

2. Cologny, Fondation Martin Bodmer, 147, Lancelot-Graal avec interpolations, fol. 344, Mort Artu: Retour de Boort à la cour du roi Arthur ; la bataille finale (© Fondation Martin Bodmer, ecodices)

135

136

3. Londres, BL Egerton 3028, Wace, Roman de Brut, fol. 49, Confrontation d’Arthur et du géant du Mont-Saint-Michel (© British Library Board)

FABRICATION ET ILLUSTRATION

137

miniatures, de la moitié d’une colonne et d’initiales historiées (surtout pour les passages de la Bible en français interpolés dans l’Estoire). La miniature qui ouvre la Mort Artu (fig. 2) est accompagnée d’une bordure ornée d’oiseaux, d’animaux terrestres et d’hybrides. L’un d’eux tient une épée et un bouclier, comme pour participer à la bataille qui se déroule dans le registre inférieur de la miniature. Plutôt que le tournoi de Winchester, qui apparaît dans la suite du texte, la présence d’une tête coupée, sur le sol à droite, et des ondes, à gauche, évoque la bataille finale de la Mort Artu. La partie supérieure montre Bohort, le seul à survivre parmi les trois héros de la Quête du Saint Graal, revenant à la cour, où le roi Arthur l’attend en compagnie de ses barons. Six écus non identifiés sont suspendus aux antennes des bordures: tous sauf un (d’or à un sautoir en forme de mailles de gueules) ont été repeints (Sylviane Messerli et Alison Stones). Arthur et le géant du Mont Saint-Michel 15. Wace, Roman de Brut. Destruction de Rome. Fierabras Angleterre (Gloucester ?) ou Sud du Pays de Galles, vers 1338–1340. Parchemin, iii + 118 + III f. (1 col. de 38 lignes), 198 x 120 mm. Provenance: Narcissus Luttrell, vers 1693; acheté en 1920 à W. C Pendarves. Londres, British Library, ms. Egerton 3028 (f. 49) Le manuscrit Egerton 3028 tient une place à part dans l’illustration anglaise en raison du nombre exceptionnel de ses miniatures (53 pour le Brut, 65 pour les deux autres textes). Couvrant une demi-page, elles prêtent une allure très particulière au défilé de l’histoire de l’Angleterre. Le manuscrit contient une version condensée du Brut de Wace suivie de continuations, jusqu’au règne d’Édouard III et au début de la guerre de Cent Ans. Celle-ci est liée aux légendes continentales grâce à la présence de la Destruction de Rome et de Fierabras. La réunion de ces différents textes semble exprimer le désir, quand commence la guerre de Cent Ans, en 1338, de maintenir une présence anglaise en France. Le récit arthurien proprement dit occupe les folios 24 à 53. Les dessins à l’encre, avec un léger lavis de couleur, se placent difficilement dans la peinture anglaise du début du XIVe siècle, en raison de leur aspect plutôt rustique et d’un manque de sophistication qui tranche avec la peinture de la cour de Londres ou celle des établissements ecclésiastiques de l’East Anglia. Pourtant, la présence de nombreuses scènes insolites présentées avec

138

vivacité fait de cet ensemble un recueil séduisant à sa façon. La confrontation d’Arthur et du géant du Mont Saint-Michel, qui a enlevé la fille du duc de Bretagne, Hélène (celle-ci donnera son nom à l’îlot de Tombelaine, où elle serait ensevelie), commence dans la miniature du folio 49 (fig. 3), qui forme une sorte de préambule. Arthur arrive devant l’énorme géant, vu de profil et à moitié masqué par la colline, en train de rôtir un sanglier. Le monstre est effrayant avec sa taille démesurée, ses sourcils broussailleux, sa barbe et ses cheveux rouges, ses oreilles pointues et poilues, sa langue semblable à celle d’un serpent et ses yeux exorbités: autant d’attributs diaboliques. Le combat du roi et du géant est un thème folklorique classique, qui renvoie aux plus anciens récits gallois consacrés à Arthur. Il est fréquent dans les manuscrits de la Suite du Merlin du Lancelot -Graal et apparaît aussi dans la Chronique de Geoffroy de Monmouth, où les images sont rares: dans le célèbre exemplaire conservé à Douai (cat. 11), c’est le combat même qui est représenté, alors que dans le manuscrit de Bonn ULB 526 le géant fuit devant Arthur. Quant au manuscrit Londres, BL, Add 10292, il montre le triomphe d’Arthur devant le géant qu’il vient de tuer. Deux manuscrits enluminés par un peintre de Douai 33. Lancelot-Graal (Estoire) Douai, vers 1285. Parchemin, 207 f, (2 col. de 36–37 lignes), signé par Walterus de Kayo, 298 x 213 mm. Provenance: Jean, seigneur d’Aigreville, fait chevalier pendant le siège de Rouen en octobre 1449, le même jour que Jean de la Rivière, qui fut en possession du ms BnF, fr 770 à cette époque; trois écus et les noms de Jehan du Roux. M. de Geresme, C. de Brichanteau (XVIe siècle?); offert à la bibliothèque municipale du Mans en 1824 par les héritiers de M. Lambert, avocat au Mans. Le Mans, médiathèque municipale Louis Aragon, ms. 354 (f. 59) 34. Lancelot-Graal (Estoire, Merlin, Suite du Merlin). Histoire d’Outremer et du roi Saladin avec La Fille du comte de Ponthieu et L’Ordre de chevalerie interpolés Douai, vers 1285. Parchemin, [3J–354–[1] f. (3 col., 48 lignes). 235 x 317 mm. Provenance: une note sur le dernier f. indique que le ms fut prêté par Pierre des Essarts au duc de Normandie (le futur roi Jean II le Bon, duc

FABRICATION ET ILLUSTRATION

4. Le Mans, Médiathèque Louis Aragon, 354, L’Estoire, fol. 59 L’écu à la croix rouge révélé à Évalac et Séraphe; leur baptême (photo: Lancelot-Grail Project)

139

140

5. Paris, BnF fr 770, Lancelot-Graal avec interpolations, fol. 236v–237, L’annonce de l’invasion des Saxons; Merlin raconte les événements à Blaise (© BnF)

FABRICATION ET ILLUSTRATION

6. Paris, BnF fr 770, Lancelot-Graal avec interpolations, fol. 237 ; l’armée s’avance (© BnF)

141

142

de Normandie entre 1332 et 1350): au XVe siècle en la possession de Jean de la Rivière, seigneur de Champlemy, mort en 1468, devenu chevalier en octobre 1449 en compagnie de Jean, seigneur d’Aigreville, possesseur du ms. Le Mans 354. BnF. Manuscrits, français 770 (f. 236 v°–237) Ces deux manuscrits, illustrés par le même peintre, transmettent la même version de l’Estoire contenant des interpolations tirées du Joseph de Robert de Boron. Deux autres manuscrits appartiennent à ce groupe textuel: Chantilly, Musée Condé, ms. 476 (644) et Saint-Pétersbourg, BNR, ms. F.v.XV.5. – dont le texte se poursuit dans le manuscrit Oxford, Bodl. Rawlinson Q.b.6. On doit à Roger Middleton la suggestion selon laquelle le manuscrit de Saint-Pétersbourg aurait été copié d’après le BnF, fr 770 lorsque ce dernier se trouvait à Paris, dans les mains de Pierre des Essarts. Cet exemple illustre bien la circulation complexe des textes arthuriens, et l’histoire encore plus complexe de leur dispersion dans les bibliothèques du monde entier à l’époque moderne. Les deux manuscrits ont été illustrés par un peintre de Douai, qu’on retrouve dans des productions destinées aux abbayes de la région, mais également dans deux copies du Lancelot-Graal (Oxford, Bodl., Digby 223 et Londres, BL Add 17443). On reconnaît son style à ses petits personnages trapus, d’expression simple. Le contenu du BnF, fr 770 laisse supposer qu’il a été commandé par un mécène laïque proche des maisons des comtes de Ponthieu ou de Flandre: en effet, les textes qui suivent le Lancelot mettent en avant l’histoire mythique des comtes de Ponthieu et de la maison de Béthune (les comtes de Flandre), qui prétendaient avoir Saladin parmi leurs ancêtres. Les formats choisis pour la disposition du texte et pour l’illustration ne sont pas les mêmes dans les deux manuscrits présentés ici: Le Mans, BM, 354 (copié sur deux colonnes par Gautier de Cayeux qui signa «Walters de Kayo» en latin) contient à la fois des initiales historiées et des miniatures qui s’inscrivent dans la colonne de texte, alors que le BnF, fr 770, copié sur trois colonnes par un écrivain anonyme, utilise surtout les miniatures. Le contenu et l’emplacement de celles-ci diffèrent aussi. Ainsi les deux miniatures du Mans (fig. 4) montrant, d’une part, l’écu blanc à la croix vermeille et l’image du Christ crucifié révélée miraculeusement à Évalac et Séraphe pendant la bataille contre l’armée du roi Tholomer et, d’autre part, leur baptême n’ont pas d’équivalent dans le BnF, fr 770. Si la scène

FABRICATION ET ILLUSTRATION

143

du baptême est relativement courante dans l’illustration de l’Estoire (elle est présente dans le manuscrit de Saint-Pétersbourg), celle de l’écu est unique. L’artiste a dû disposer de plusieurs modèles en créant ces programmes iconographiques différents. Le BnF, fr 770 présente une ample série d’illustrations pour le Merlin et sa Suite. Le choix des sujets est très comparable à celui de BnF, fr 95 (cat.3) et de Londres, BL, Add 10292. Dans la Suite, Merlin joue le rôle principal sous la direction de son tuteur Blaise, qui met par écrit tout ce que Merlin lui raconte (fig. 5). Ici, un messager venu de Grèce annonce l’invasion des Saxons: les dix rois chrétiens lèvent une armée contre eux et remportent la victoire à Clarence (fig. 6). Peu de temps après, Arthur épousera Guenièvre, la fille du roi Léodagant. Fin de l’ESTOIRE et début de la QUESTE 35. Lancelot-Graal (Estoire, Queste, Mort Artu) Tournai, Gand ou Saint-Omer, vers 1317. Parchemin, 162 + 11 f. (3 col. de 51 lignes). 480/484 x 338 mm. Provenance: signatures d’Elizabeth (1466– 1503) et de Cécile (1469–1507), filles du roi Edouard IV d’Angleterre, et de Jane Grey (f. 1, un folio n’appartenant peut-être pas au reste du manuscrit à l’origine, comme l’a remarqué Roger Middleton); note de possession de Richard Roos, poète partisan des Lancastre durant la guerre des Deux Roses, qui, dans son testament du 8 mars 1482 légua à sa nièce Alianor Hawte « my grete booke called Saint Graal ». Londres, British Library, ms. Royal 14 E.III (f. 88v–89) (Non exposé) Ce manuscrit fait partie d’un groupe de trois exemplaires du LancelotGraal exécutés en Flandre pour des mécènes anonymes dans les premières décennies du XIVe siècle. Le second est aujourd’hui divisé entre la Bibliotheca Philosophica Hermetica d’Amsterdam (ms. 1, vendu à Sotheby’s le 7.12.2010, lot 33 et maintenant en mains privées), la John Rylands University Library de Manchester (ms. French 1) et la bibliothèque Bodleienne d’Oxford (ms. Douce 215). Le seul exemplaire complet, Londres, BL Add 10292–4, contient au folio 55 v° du manuscrit 10292 une image montrant la duchesse Flegentine commandant des tombeaux qu’un sculpteur est en train de graver. L’une des dalles porte la date du 17 février 1316 (1317 selon la datation

144

d’aujourd’hui). C’est donc autour de cette date qu’il convient de placer leur exécution. L’exemplaire Royal est plus grand et plus luxueux que les deux autres; il contient sur toutes les pages des initiales à filigranes et bordures alternant le rouge, le bleu et l’or. Il est illustré par un artiste qui a exécuté le portrait ajouté par Guilbert de Sainte-Aldegonde à son psautier à l’usage de Tournai lorsqu’il l’offrit à la chartreuse de Longuenesse en 1323 (Saint-Omer, BM, 270), et fut aussi l’artiste principal des autres manuscrits de la série, même s’il recourut également à des collaborateurs. Royal et Amsterdam mettent tous les deux l’accent, à la fin de l’Estoire, sur la punition du duc responsable de la mort du roi Lancelot l’Ancien, l’ancêtre de Lancelot du Lac, qui sera le seul à pouvoir retirer la tête de son aïeul de la fontaine bouillante pour la réunir à son corps (fig. 7). En face (fig. 8) se trouve le début de la Queste del saint Graal. Arthur et sa compagnie, en train de jouir du banquet de la Pentecôte, reçoivent la visite de la demoiselle qui vient chercher Lancelot pour qu’il adoube son fils Galaad. La deuxième miniature représente l’adoubement de Galaad, devant

7. Londres, Royal 14 E.III, LancelotGraal, fol. 88v, l’Estoire: La tête du roi Lancelot jetée dans la fontaine bouillante (© British Library Board)

FABRICATION ET ILLUSTRATION

145

8. Londres, Royal 14 E.III, Lancelot-Graal, fol. 89 Queste del saint Graal: Le banquet de la Pentecôte à la cour du roi Arthur; l’adoubement de Galaad devant le couvent; un musicien, un nu jouant de la trompette d’où sort un lapin, un écu, un singe tirant à l’arc, un papillon, un hybride, deux chevaliers affrontés (© British Library Board)

146

le couvent où il a été élevé. Cette page, dont les sujets sont pratiquement de rigueur pour l’ouverture de la Queste, se distingue par son luxueux encadrement, qui contient une grande variété de personnages, d’animaux et d’oiseaux souvent fantaisistes. Lutte contre un lépreux 58. Le Roman de Jaufré Sud de la France, fin du XIIIe ou début du XIVe siècle. Parchemin, 110 f (2 col. de 37 lignes), 205 x 160 mm. Provenance: acquis par Louis XIV en 1662 avec la collection Béthune. BnF. Manuscrits, français 2164 (f 28 vo–29) Avec l’autre exemplaire du Roman de Jaufré (BnF, fr 12571), copié en Italie et ne comportant pas d’illustration, ce manuscrit suscita l’intérêt des premiers chercheurs en littérature du Moyen Âge. Ainsi, comme de nombreux manuscrits de Chrétien de Troyes, les deux exemplaires de Jaufré servirent à Henri-Pascal de Rochegude, qui en fit une copie au début du XIXe siècle. BnF fr. 2164 est en effet l’unique manuscrit arthurien en occitan qui comporte des illustrations. Pour Clovis Brunel, la graphie du manuscrit indique une production du Sud du Languedoc, avec de nombreux éléments catalans, surtout dans la première partie. Comme l’avait déjà noté Brunel, on trouve d’autres traces de Jaufré en Catalogne: on sait par des sources du XIVe siècle qu’une salle du palais des rois d’Aragon à Saragosse était ornée de peintures inspirées de ce roman. Les dessins du manuscrit sont réalisés très simplement, au lavis de couleur, sans modelé. Les sujets se répètent (Jaufré chevauchant, en particulier) mais les événements les plus importants du récit sont absents de l’illustration alors que le manuscrit ne compte pas moins de 250 images ! Sur le plan stylistique, on peut faire des rapprochements avec le milieu avignonnais, où l’on trouve de très grands personnages dessinés à l’encre dans les manuscrits juridiques conservés à Berlin (DSB, Ham 365) ou Avignon (BM, 749) par exemple. On peut se demander en fin de compte si ce curieux manuscrit était un livre d’enfants et si la pièce peinte du palais de Saragosse servait de chambre d’enfants et de salle de jeux. Toutefois, les miniatures, comme le texte, peuvent être d’une extrême violence, surtout dans les nombreuses scènes de combat. Aux folios 28 v–29,

FABRICATION ET ILLUSTRATION

9. Paris, BnF fr 2164, Roman de Jaufré, fol. 28v, Jaufré en combat avec un lépreux muni d’une arme rectangulaire (lo mezel) (© BnF)

147

148

10. Paris, BnF fr 2164, Roman de Jaufré, fol. 29, Une demoiselle en prière devant une tige fleurie; Jaufré tranche le bras du lépreux (© BnF)

FABRICATION ET ILLUSTRATION

11. Paris, BnF fr 2164, Roman de Jaufré, fol. 30v, la demoiselle guérit les blessures de Jaufré devant un arbre en fleurs (© BnF)

149

150

Jaufré lutte contre un lépreux muni d’une arme rectangulaire (lo mezel) pour défendre une demoiselle (fig. 9); il réussit à couper le bras puis la jambe du lépreux, qu’il finit par tuer quelques miniatures plus tard (fig. 10). Sur ces entrefaites, la demoiselle guérit les blessures de Jaufré en lui versant de l’eau sur la tête, alors qu’ils contemplent le lépreux mutilé qui gît mort à leurs pieds (fig. 11). On peut soupçonner là des éléments satiriques et humoristiques... qui anticipent à certains égards les aventures pseudoarthuriennes des Monty Python ! Prouesse suprême de Lancelot et marges facétieuses 66. Lancelot-Graal (Lancelot, Queste, Mort Artu) Tournai, achevé le 14 mars 1345. Parchemin. [4]–322–[4] f. (2 col, de 50 lignes). 350 x 430 mm. Provenance: Louis de Bruges; Louis XII; présenta la Bibliothèque du roi à Blois en 1500 BNF, Manuscrits, français 122 (f. 1) Ce manuscrit transmet une version spéciale de cette partie du Lancelot qui repose sur Le Chevalier de la Charrette de Chrétien de Troyes. On peut attribuer sa décoration au cercle artistique tournaisien qui a décoré le BnF, fr 1424 (cat. 61) et le Roman d’Alexandre, illustré par Jean de Grise en 1344 (Oxford, Bodl., ms. 264). Malgré l’absence d’illustrations en pleine page comme dans l’Alexandre, l’artiste utilise de grandes miniatures qui occupent toute la largeur des deux colonnes de texte. Comme son contemporain Jean de Grise, il prête beaucoup d’attention aux détails du costume, des armes et des armoiries, aux fonds de miniatures ornés de rinceaux délicats ainsi qu’aux motifs architecturaux. La magnifique miniature d’ouverture représente les épreuves les plus remarquables de la prouesse suprême de Lancelot (fig. 12). D’abord il traverse le Pont de l’épée, puis il vainc trois lions grâce à l’anneau magique que lui a donné la Dame du Lac et, enfin, il triomphe au combat à la lance contre Méléagant, le fils du roi Baudemagu, qui tient la reine Guenièvre prisonnière. Guenièvre et Baudemagu regardent les exploits de Lancelot depuis les tours du château. C’est l’une des plus remarquables miniatures représentant ces aventures merveilleuses, pour une fois groupées dans une seule image à l’ouverture du volume. Il n’existe pas à notre connaissance

FABRICATION ET ILLUSTRATION

151

12. Paris, BnF fr 122, Lancelot, fol. 1, Lancelot traverse le pont de l’épée, tue les lions, et combat Méléagant, fils du roi Baudemagu qui tient Guenièvre prisonnière; Guenièvre et Baudemagu regardent depuis les tours du château. Dans la marge et paon et une série de jeux entre hommes, femmes et singes (© BnF)

152

13. Paris, BnF fr 122, Lancelot, fol. 1, Une série de jeux entre femmes, hommes, et singes (© BnF) 14. Paris, BnF fr. 339, Lancelot, fol. 23, La dame de Malehaut tombe amoureuse de Lancelot (© BnF)

FABRICATION ET ILLUSTRATION

153

d’autre exemplaire commençant par cet épisode et l’on peut se demander si le commanditaire s’y intéressait particulièrement. La page est entourée d’une bordure à rinceaux en antennes souples et motifs de feuilles. En bas, un singe est prêt à fouetter son petit et, sur la bordure inférieure, plusieurs scènes présentent des jeux quasiment courtois (fig. 13): une femme tirant à l’arc sur une cornemuse (symbole sexuel) devant une autre femme qui lève les bras, alors qu’une troisième s’approche à genoux d’un homme; deux couples jouent à colin-maillard; enfin un singe tenant une cruche plante un dard dans les fesses d’un homme nu. Autant d’images qui prennent à contre-pied la bravoure de Lancelot et son amour courtois pour la reine. La Dame de Malehaut tombe amoureuse de Lancelot 69. Lancelot-Graal (Lancelot, Queste, Mort Artu) Paris?, vers 1250. Parchemin. 283 f. (2 col de 59 lignes), 344 x 247 mm. Provenance: Colbert; entré à la Bibliothèque du roi en 1732 BNF, Manuscrits, français 339 (f. 22vo–23) En dépit d’une mise en page très soignée et d’une division en paragraphes régulière, l’illustration de ce manuscrit est modeste, les initiales historiées et les rares miniatures n’occupant que la moitié de la largeur d’une colonne de texte (fig. 14). La gamme de couleurs, qui s’en tient au bleu et au rouge foncé, au gris et au blanc, est caractéristique de la peinture parisienne du milieu du XIIIe siècle, mais l’artiste anonyme qui a enluminé ce manuscrit est difficile à retrouver ailleurs. Cet exemplaire se distingue toutefois parmi les LancelotGraal des années 1200–1250 par le nombre élevé de ses illustrations – 120 au total. Trois initiales historiées évoquent les efforts faits par la Dame de Malehaut et sa cousine, chez lesquelles Lancelot est prisonnier, pour découvrir son identité pendant qu’il dort, épuisé après avoir remporté un premier tournoi. Entre-temps Arthur, entouré de ses barons, devient pensif, car il voudrait connaître l’identité de ce rouge chevalier, le meilleur du monde. Enfin, Lancelot refuse de dire qui il est à la Dame de Malehaut. Ces événements précèdent la deuxième assemblée, durant laquelle Lancelot, portant des armes noires, sera de nouveau victorieux. Il révélera finalement son nom à la reine Guenièvre et lui dira tout ce qu’il a fait depuis qu’elle lui a donné son épée.

154

La Dame de Malehaut, veuve et mère, manifeste, comme un peu plus tard la Demoiselle d’Escalot, la séduction presque involontaire qu’exerce Lancelot sur les femmes. Tombée immédiatement amoureuse de son prisonnier, elle pressent pourtant que c’est pour une autre qu’il réalise ses exploits chevaleresques. Elle peut donc être considérée comme l’une des incarnations, un peu négatives, de l’amour courtois dans le Lancelot en prose.

VI Text and Image in Arthurian Manuscripts

W

hat do images do ? Clearly they were important elements in the production of medieval Arthurian manuscripts, as most copies have pictures, and those without narrative images invariably articulate the written page with other kinds of textual markers: painted in colours and gold, penflourished in red and blue, or large initials in a single colour. And small text capitals may also be coloured. They all contribute, to a greater or lesser degree, to the decorative appearance of the manuscript page and break up the text, serving as markers enabling a reader/viewer to find her place again when reading/viewing was interrupted. And of course images, particularly those embellished with gold and colours, add substantially to the cost, and the value, of the book. But images, and other markers, offer a far more complex and interesting network of meanings and functions, which are the focus of this chapter.1 Many medieval texts comment specifically on the need for, and the purpose of, illustrations: to help understand aspects of the text and to remember what was written, and even to confirm the truth of what is being said in words. The Image du monde of Gossuin de Metz, the Breviari d’amor of Matfre Ermengau, and the Roman de Fauvel are examples where the purpose of the pictures is to better understand the text, while Richard de Fournival’s twin paths of ‘parole’ and ‘painture’ leading to the two doors of seeing and hearing in Memory’s Tower is a famous instance of pictures for remembering. And for Henri de Mondeville, author of a surgery treatise First written for Handbook of Arthurian Romance, ed. Leah Tether and Johnny McFayden, pp. 215–33 (Berlin: De Gruyter, 2017). 1

156

dedicated to King Philippe le Bel, his set of anatomical pictures represented the way to truth. There is nothing quite as explicit in the Lancelot-Grail romance, nor, so far as I know, in any other Arthurian text. The presence and usefulness of illustrations must have been taken for granted, as indeed it was in all kinds of books both secular and sacred throughout the thirteenth century and beyond. Lancelot, Morgan and Arthur and the rôle of pictures There is however one well-known sequence in the text of the Lancelot where pictures are the means of unveiling truth. When Lancelot is a captive in Morgan’s prison, he looks through the window of his cell and sees in an adjacent room an artist painting on the wall a picture of Aeneas escaping from Troy. It occurs to Lancelot that he could paint pictures of his amorous adventures with Queen Guinevere. The result is so successful that he embraces his picture of the Queen. Morgan plans to show the pictures to King Arthur, so that he will finally realize that the rumours about Lancelot and Guinevere were true. She does so much later in the text, in the Mort Artu, and the result is exactly what Morgan had hoped for. This revelation of the betrayal of Arthur’s favourite knight signals the beginning of the downfall of the king. One might imagine that this sequence would have been an intrinsic part of the iconography of the Lancelot, not only for its significance in the story, but at the same time to epitomize the importance of the painted image as a measure of what was considered important in the text in general. On the contrary, only a single illustration of Lancelot’s artistic activity has survived, and it does not correspond to what is in the text: Lancelot is shown embracing not his picture of the Guinevere on the wall, but a statue of her (London, BL Add. 10293, f. 325v), on the model of Pygmalion or the Roman de la Manekine where sculpted images take on life-like qualities. And Morgan’s revelation to Arthur was also a subject rarely treated pictorially: there are only four instances, two of which simply show Arthur and Morgan in conversation (Paris, BnF fr 342, f. 167; Yale University, Beinecke 229, f. 289), seated together on a bench, with no signs of any painting on the walls. The other two both occur in manuscripts made for Jacques d’Armagnac, one showing Morgan explaining the meaning of the paintings to Arthur (BnF fr 116, f. 688v); the other showing Arthur alone, seeing the pictures and reading the captions placed beneath them (BnF fr 112, vol. III, f. 193v). One can infer that for Jacques d’Armagnac the revelation through pictures was indeed a

TEXT AND IMAGE IN ARTHURIAN MANUSCRIPTS

157

significant episode even if his books do not include an image of the making of those pictures. But why did Add. not include images of Arthur’s discovery of the adultery through Lancelot’s paintings ? We cannot know. Many of the questions raised by the selection of images are simply unanswerable. Choices and emphases in the illustrations of LANCELOT Variations in the number, placing in the text, and treatment of subjects among Lancelot-Grail manuscripts indicate that choices were made. No two copies contain exactly the same selections of images and none are identical in pictorial treatment, even when the same artists made them, as the two versions of Arthur seeing Lancelot’s pictures show. Another striking example of differences in selection and treatment of images from two manuscripts made in the same workshop is the False Guinevere episode in London, British Library MS Additional 10293 (hereafter Add.) and olim Amsterdam, Biblioteca Philosophica Hermetica MS 1 (Amst.), both made in Flanders in the early years of the 14th century. Add. treats the episode in a lengthy sequence of 24 single-column miniatures while Amst. limits its selection to only five images — no doubt a much cheaper option but one that required a fuller treatment of the subject in just a few miniatures than a more developed sequence of pictures would need. The five miniatures in Amst. depict subjects that are also present among Add.’s 24, but they are sometimes placed at different points in the text and they treat the subjects in different ways. Both manuscripts begin the sequence at the same place in the text, with the arrival of the False Guinevere’s messenger at Arthur’s court. In Add.’s version (f. 131) everyone is standing, in two facing groups, one led by King Arthur with his barons behind him, the other the messenger and her entourage; Arthur and the messenger raise hands in speech gestures. The version in Amst. (ii f. 202) is much more elaborate: the scene takes place in a room framed with pointed arches, turrets, and quatrefoil tracery. Queen Guinevere and King Arthur are seated, with barons behind them; Bertholais, standing, presents the False Guinevere’s messenger whose veil has been thrown onto the checkered floor and she opens the box containing the False Guinevere’s letter, all details specified in the text but absent in Add.’s opening miniature. Add. continues the story with an image of the reading of the letter before the court, the arrival of the False Guinevere and her companions, and King Arthur seized by King Tholomer’s knights. Both manuscripts then have two miniatures each, at the same places in the text,

158

depicting King Arthur imprisoned by the False Guinevere, followed by King Galehot and Queen Guinevere offering the crown to Gauvain because of Arthur’s absence. After that, Add. has 9 miniatures depicting debates with King Arthur about his treatment of Queen Guinevere and Lancelot’s duel fought as her champion; to these there is no pictorial counterpart in Amst. Both manuscripts include an illustration at S IV 72. 9 and LM III 89 but this time the subjects are different. Add. illustrates the interdiction on Arthur delivered by the pope using a lighted taper, a motif borrowed from legal illustration such as is found in Gratian’s Decretum. One can surmise that the legal aspects of the case were of interest to the anonymous patron of Add. but not to the patron of Amst., where the chosen illustration is the death-bed confession of the False Guinevere. Add. also includes a death-bed scene but not until three miniatures later, following illustrations of Arthur’s own repentance and absolution. The end of the episode is marked in both manuscripts by reconciliation: Lancelot embraced by Arthur in Amst. and Lancelot begged by Guinevere to rejoin the Round Table: but in Add. this is the last of five concluding scenes and the corresponding miniature in Amst. occurs at the same place in the text as the fourth scene. There are also differences of format between the two manuscripts, in part because Amst. was copied in two columns and Add. in three; and Add. uses champie initials for the text opening below the miniatures whereas Amst. has pen-flourished initials. The wording of the rubrics is also different, another aspect of the creation and transmission process that requires further work. Variations in ESTOIRE manuscripts between a model and its copy In one case one can show a relationship between two manuscripts that suggests the pictures in one were copied from the other, but even then, additions were made and the treatment of the subjects is far from identical. The version of the Estoire del saint Graal adapted for Jean-Louis de Savoie, Bishop of Geneva († 1482), by Guillaume de la Pierre and attributed to Jean Colombe or a painter in his entourage: Brussels, Bibliothèque Royale 9246 (hereafter BR), was most likely based directly on a Parisian copy of the early 14th century illuminated by an artist I have called the Sub-Fauvel Master (Paris, BnF fr 105). Two copies were produced in the Sub-Fauvel workshop, fr 105 and fr 9123 (the lattter a collaboration between the SubFauvel Master and the so-called Maubeuge Master). Comparison with BR 9246 shows that it was mostly fr 105 and not fr 9123 that served as model

TEXT AND IMAGE IN ARTHURIAN MANUSCRIPTS

159

to BR 9246, with two possible exceptions. There are 11 miniatures in fr 9123 that have no corresponding miniatures at the same places in fr 105 and BR, and 22 places where fr 9123 lacks a miniature to correspond with what is in fr 105 and BR. Occasionally a miniature in fr 9123 corresponds to one of the same subject at a different place in the text in fr 105 and BR. And there are two cases where a subject included in fr 9123 but not in fr 105 takes up part of a miniature in BR (on ff. 95 and 111), suggesting that perhaps the BR artist was aware of fr 9123 and its illustrative program but had decided not to adopt it as a model for the whole, preferring fr 105; unless there was another intermediary that is now lost. However the placing of the miniatures, their subjects, and their rubrics are the same in fr 105 and BR, with two exceptions: BR 9246 introduces two miniatures that are not in the model, and not in fr 9123 either — and for these there are no rubrics. Given the discrepancy in date between fr 105 and BR 9246 it is not surprising that the illustrations are differently treated: Jean Colombe’s style is characterized by a developed interest in architecture, landscape and seascape, in interaction among the protagonists, and in a sophisticated painterly technique, to a far greater degree than fr 105, where the miniatures are sketchy and summarily washed in colour. There are differences in format too, BR’s pictures are mostly bigger, often taking up half a page of a twocolumn text, and sometimes presenting a sequential narrative, expanding what is depicted in fr 105 and emphasizing the importance of illustration in relation to text. The two additional miniatures depict Titus and Vespasian bidding farewell to Josephé and his followers (f. 13v), and Josephé and his followers before King Elzalach and his men in a room decorated wtih relief scuptures (or grisaille paintings) of the Life of Christ on the walls (f. 18v). Why these subjects ? Both add emphasis to the rôle of Josephé as hero of the Estoire and his evangelizing mission, as would be appropriate for a patron who was also a bishop. The final miniature of the Estoire in fr 9123 depicts the lions who guard the bleeding tomb of King Lancelot (grandfather of the eponymous hero of the Lancelot branch) and lick the blood. Neither fr 105 nor BR include a picture at this point in the text, but both give a minor initial and accompany it by a rubric. The rôle of initials as markers The rôle of large initials as substitutes for, or in anticipation of, illustrations, is an area that needs work. Amst. is a manuscript where large champie

160

initials, mostly accompanied by rubrics, often mark points in the text where in the related manuscripts Add. and Royal 14 E.III there is a miniature. In this case the use of champies must have been in part an economic choice since champies would have cost substantially less than full-scale miniatures. The issue is more complicated than that, however, as it raises the question of relative precedence: did the placing of large champies or large party-bar (puzzle) initials precede the establishment of pictorial cycles, or vice-versa ? An examination of the second part of the Estoire del saint Graal in some of the earliest unillustrated manuscripts in relation to the earliest illustrated copy shows a high correlation in the placing of large pen-flourished initials and narrative illustrations. The earliest illustrated Estoire is Rennes, BM 255 (hereafter Rennes), c. 1220, which has historiated initials at its major textual breaks. There is also another copy that may be about the same date, with only pen-flourished initials: WLC/Lm7 in Nottingham University Library (hereafter Lm7). Both are written above top line. From a generation or so later — it is written below top line (c. 1250–60 ?) is London, BL Royal 19 C.XII (hereafter Royal), which also has only pen-flourished initials. Since Lm7 is incomplete, preserving only the second part of the Estoire, my comparison also begins there. There are 19 large pen-flourished initials in Lm7 and these places are also marked by large pen-flourished initials in Royal. Rennes has 16 historiated initials at corresponding places but at the beginning of the sequence (no. 1 in Lm7 and Royal) it has only a small champie initial, and in two more instances (corresponding to nos. 3 and 4 in Lm7 and Royal) Rennes omits any decoration at all. When we include two copies of the Estoire from the same workshop, probably based in Douai, c. 1285, we find that all 19 of the large pen-flourished initials of Lm7 and Royal are places also marked in Le Mans BM 354 and Paris BnF fr 770, this time with miniatures, historiated initials or large champie or party-bar pen-flourished initials — with the sole exception of no. 1 which, as also in Rennes, has only a small initial (pen-flourished in Le Mans and fr 770, a small champie in Rennes). These correlations suggest that Lm7 played an important part in the establishment of a system of break-points that stands at the beginning of the illustrative tradition and was drawn upon to a greater or lesser degree in the copies produced subsequently. And the differences between Lm7 and Rennes show that there was more than one model in the early stages of the emergence of patterns of textual marking in Estoire manuscripts.

TEXT AND IMAGE IN ARTHURIAN MANUSCRIPTS

161

Conclusion The examples outlined above have all been examined comparatively. Only through a comparative approach can one detect patterns of similarity or of difference in the choice, placement, and treatment of images and thus determine the ways in which a particular copy follows established patterns or breaks new ground. Sometimes we can suggest reasons for the choices made, particularly when we know who the patron and/or the makers were, but in the absence of detailed payments or other documentary evidence, any hypothesis must remain conjectural.1

Several of the points made here are discussed in more detail in other essays in this collection, especially nos. IX, XII, XXVI, XXVII, XXX. 1

B The Lancelot-Grail Romance Methods and Subjects

VII The ‘Lancelot-Graal’ Project

T

his essay describes the rationale for the pilot phase of a computer database of text and pictures that we hope will eventually form a Corpus of ‘Lancelot-Graal’ manuscripts, will present a model for manuscript analysis in general, and will have a range of other applications beyond the field of manuscript studies. An international team of Old French specialists (Keith Busby, Elspeth Kennedy, Roger Middleton) and art historians and manuscript specialists (Susan Blackman, Martine Meuwese, Alison Stones) are collaborating with technical consultants in information science and telecommunications (Kenneth Sochats, Guoray Cai) to create and use a searchable database of primary manuscript pages and secondary commentary linked to a searchable database of images. The specialists will use the database to generate a variety of products, both in the traditional form of books and articles, and in electronic form on the World Wide Web and CD-ROMs. What we hope to learn is more about the intentions of the makers of the manuscripts and those (patrons? directors of operations?) who guided them in the choices they made. Our work to date shows that the choice, placement and composition of the illustrations varies very considerably from one manuscript to another, even among copies produced by the same scribes, decorators and artists. Certain manuscripts display at times a very surprising degree of careful attention to the nuances the text in that copy conveys. Illustrations showing the same episode in other copies will not necessarily First published in New Directions in Later Medieval Manuscript Studies, ed. Derek Pearsall (York Medieval Press), Woodbridge, 2000, pp. 167–82. Updates were published in 1999 and 2010, both reprinted in these essays, and another appeared in a special issue of Speculum in October 2017, ‘Mapping Illuminated Manuscripts: Applying GIS Concepts to Lancelot-Grail Manuscripts,’ Speculum 92/S1, S170–S189 (electronic publication available free through the University of Chicago Press).

166

present a comparably text-dependent picture. We are looking at which, where and how, in the hopes of moving a step closer to understanding why. Conceptual framework The idea for collaborative work on the manuscripts of the ‘Lancelot-Graal’ has developed from several models: the editing and interpretive work of text scholars; the Manuscrits de Chrétien de Troyes project; studies by myself and others on various aspects of the production, iconography and cultural context of the illustrated ‘Lancelot-Graal’ manuscripts and the illustration of other literary texts; the growing emergence of World Wide Web sites as important vehicles for research, learning and teaching in humanities disciplines; and the Geographic Information Systems software projects under way at the University of Pittsburgh under the direction of Kenneth Sochats. Les Manuscrits de Chrétien de Troyes grew from a session I chaired in 1988 at the 23rd International Medieval Congress at Kalamazoo, Michigan. The participants, Keith Busby, Terry Nixon, Lori Walters and I collaborated to launch and edit a research project that encompasses a catalogue of all forty-five manuscripts and fragments of the romances of Chrétien, a corpus of illustrations of the hands of all the scribes, reproductions of all the illustrations and seventeen essays by scholars from different disciplines — Old French specialists, manuscript scholars and art historians, about aspects of the presentation, transmission, interpretation, reception, decoration and illustration of the manuscripts. We published a volume containing the materials of a research dossier and a volume of selected essays demonstrating how scholars approached those materials with different questions in mind.1 The present study of the ‘Lancelot-Graal’ manuscript tradition is also a collaborative effort among art historians, text scholars, manuscript specialists, and now specialists in Information Science. Like the romances of Chrétien de Troyes, the ‘Lancelot-Graal’ is also a text in Old French to which scholars in various disciplines have devoted a good deal of attention as editors and critics.2 However, the compilation and its copies present logistical challenges that made the idea of publishing a two-volume study as we had done with the manuscripts of Chrétien seem too limited a goal, although we envisage 1 Les Manuscrits de Chrétien de Troyes, ed. K. Busby, T. Nixon, A. Stones, L. Walters; see now the additional fragment at the PRO, E122/100/13B, published by Jefferson. 2 Text editions and studies are listed in the bibliography at the end of the volume.

THE ‘LANCELOT-GRAAL’ PROJECT

167

publication in book and article form as a parts of the products we expect the project to generate. The number of surviving copies of all or part of the five-part ‘LancelotGraal’ cycle, counted by Brian Woledge at some 209 manuscripts and fragments, is more difficult to manage than the forty-five manuscripts and fragments (now forty-six) that comprise the corpus of Chrétien manuscripts.3 There are over four times as many copies as Chrétien manuscripts, to say nothing of the much higher proportion of illustrated copies among the extant manuscripts, and there are more illustrations in the illustrated copies — sometimes several hundred per volume. Describing and reproducing even a proportion of them is prohibitively expensive, either to acquire the photographs or to publish them, to say nothing of the work involved in determining which are the illustrated folios in each manuscript. In most cases this must be done from first-hand observation.4 Previous studies 5 of

See n. 1 for Chrétien manuscripts. For the Lancelot-Grail cycle see Woledge, Bibliographie and Supplément, nos. 93, 96,114; see also Micha,‘Les manuscrits du Lancelot en prose’; and especially now (2006) Middleton, ‘The Manuscripts’ in Arthur of the French. Many of the manuscripts listed separately by Woledge and Micha are actually parts of the same multi-volume set, while other partial copies may have lost their sister volumes, so that a count based simply on numbers of manuscripts is misleading. Both Woledge and Micha are also frequently inaccurate in their dating attributions. Suggested emendations to their lists are to be found in Stones, Illustrations of Lancelot, eadem, ‘The Earliest,’ pp. 42—4, and eadem, ‘Aspects of Arthur’s Death,’ pp. 87–95. Fragments are still coming to light, such as those now in the Sharon and Neil Phillips/ Phillips Family Collection, New York, which passed through Sotheby’s on 14 July 1981 as lot 30 and Christie’s on 3 April 1984 as lot 62; see Avril and Reynaud, Les manuscrits à peintures 1440–1520, p. 37, and Bruckner, ‘Reconstructing Arthurian History’, in Netzer and Reinburg, eds., Memory and the Middle Ages, pp. 68–9 and nos. 55–7,58–61. 4 There are only two facsimiles: La grant Queste del Saint Craal. Versione inedita del XIII secolo del ms. Udine, Biblioteca Arcivescovile 177 (Tricesimo, 1990 [illegally published and subsequently withdrawn from circulation]), and La Bûsqueda del santo Grial (Dijon, BM 527), ed. C. Alvar with F. Avril (Valencia, 1997). Several of the BNF manuscripts, including fr 95, fr 112, fr 113–16, fr 117–20, are now available (in 1999–2000) for consultation on microfilm in the Cabinet des manuscrits. These are now (2016) on line on Gallica, along with many more BnF manuscripts. 5 The most comprehensive study of the illustration is still Loomis, Arthurian Legends which includes reference to, and illustration of, media other than manuscripts. See also Stones, ‘Arthurian Art Since Loomis’, reprinted in these essays. Studies of the iconography of one or more branches of the cycle are: Remak-Honef, ‘Text and Image in the Estoire del saint Graal’; M. Meuwese, L’Estoire del saint Graal; Blackman, ‘The Manuscripts and Patronage of Jacques’, which includes (I, 213–43) a comparative table of the Arthurian subjects. The tables are also reproduced in eadem, ‘A Pictorial Synopsis’. 3

168

these illustrated manuscripts, including my own,6 have of necessity been limited to a consideration of just a few manuscripts or illustrations, selected either by chronology, or by subject, or for other unstated reasons. The manuscripts of the Pilot Project While it is essential to start from a general picture of what survives,7 we have decided to begin our project with a detailed study of three related manuscripts which present particularly interesting textual, illustrative and historical characteristics. They are the copy now divided between olim Amsterdam Bibliotheca Philosophica Hermetica MS 1(sold at Sotheby’s, 7.12.2010, lot 33, and now in private hands)/Manchester, The John Rylands University Library, MS Fr. 1/Oxford, Bodleian Library, MS Douce 215, and the other two copies produced by the same team of craftsmen: London, BL MSS Additional 10292–4 and Royal 14 E.III. The group has been convincingly dated in the first quarter of the fourteenth century on the basis of the inscription on the tomb depicted on f. 55v of Add 10292 (Plate 1) which includes the year 1316 (1317 ns),8 and stylistic arguments support an attribution to western Flanders or eastern Artois, on the basis of such parallels as the portrait of himself before the Virgin and Child added by Gilbert de Sainte-Aldegonde to the psalter of Tournai use which he presented in 1323 to the Chartreuse of Longuenesse near Saint-Omer (Plate 2).9 The layout of the page and the scribal characteristics are fairly consistent throughout all the volumes, but at least two scribes participated; two decorators did the champie initials, and at least three artists did the miniatures. The major painter worked in all three copies (Plates 1, 5), while hand 2 worked only in Additional 10293 (Plate 3) 6 My own focus has been on style or iconography or on a particular copy: the Morgan Lancelot (New York, Morgan Library, MS M.805–6) was included in ‘L’atelier artistique de la Vie de sainte Benoîte d’Origny. In the articles ‘Images of Temptation, Seduction and Discovery’ and ‘Illustrating Lancelot and Guinevere’, I considered the iconographical treatment of the Lancelot-Guinevere romance. In ‘The Illustrations Beinecke MS 229 and Paris, BnF fr 95, I took a manuscript now in two parts and examined questions of its iconography, style, production, and ownership. Other essays on the Lancelot-Graal in Busby’s and Walters’ edited volumes also adopt narrow foci. 7 A full analytical list, based on first-hand examination made over the last three decades, is in preparation. See now (2016) http://www.lancelot-project.pitt.edu/LG-web/Arthur-LGChronGeog.html; http://www.lancelot-project.pitt.edu/ and the list at the end of this volume. 8 Reproduced in Loomis, Arthurian Legends, fig. 248. 9 Stones, ‘Another Short Note’.

THE ‘LANCELOT-GRAAL’ PROJECT

1. London, BL Add 10292, Estoire del saint Graal, fol. 55v, Duchess Flegentine commissions the Tombs of Judgement (© British Library Board)

2. Saint-Omer, BM 279, Psalter, fol. H, Guillaume de SainteAldegonde kneels before the Virgin and Child (photo: author)

169

170 3. London, BL Add 10293, Lancelot, fol. 326, Gauvain visits King Baudemagus who lies ill, tended by two monks (© British Library Board)

4. Manchester, The John Rylands University Library, French 1, Queste del saint Graal, fol. 183v, King Arthur, Queen Guinevere and courtiers watch Galaad hold the sword drawn from the stone (© John Rylands University Library of Manchester)

THE ‘LANCELOT-GRAAL’ PROJECT

5. London, BL Add 10293, Lancelot, fol. 128v, Queen Guinevere, King Baudamagus and courtiers watch Lancelot cross the sword bridge (© British Library Board)

6. Olim Amsterdam, BPH 1, iii, Lancelot, fol. 45 Queen Guinevere and King Baudemagus watch Lancelot cross the sword bridge (photo: Lancelot-Grail Project)

171

172

and hand 3 worked only in Rylands/Douce (Plate 4). The text versions also differ, nor are the selections of illustrations the same. There are, in addition, significant variations in how comparable subjects are depicted and described in the accompanying lemmata or rubrics. For instance, Lancelot crossing the sword bridge, watched by Guinevere who has been imprisoned by King Baudemagus, was included among the original illustrations in Add 10293 (Plate 5) but omitted in the Amsterdam manuscript. Around 1400, the owner of Amsterdam (whose identity is unknown) had a miniature showing this subject added to the book in the bottom margin of the appropriate folio (Plate 6). This section is missing in Royal 14 E.III. There are many more interesting subjects and sequences of images that have never been explored comparatively. We plan a comprehensive set of comparisons among the illustrations in all three copies in order to find out what the variants in selection and treatment tell us about the relative priority of one copy over the others, the iconographical preferences to which the different copies attest, and what that tells about the circumstances of production. Other puzzles we hope to solve are questions of commissioning and ownership, using clues in the form of shields, inscriptions, and aspects of the physical make-up of the manuscripts such as water-marks on the paper flysheets, later annotations on the manuscripts themselves, and the like. Arthurian heraldry is profusely represented, with relative consistency for the first time in the entire illustrative tradition,10 but there are also some shields that may reflect a desire on the part of patrons to have themselves shown as Lancelot or as another knight. There are also shields on the borders of opening pages that seem to point to early ownership, and annotations by later owners who may be traceable. In the case of Royal 14 E.III there are indications that it was owned by Richard Roos, knight, then by Elizabeth and Cecily, daughters of King Edward IV and Elizabeth Woodville, and Lady Jane Grey,11 and watermarks on the flysheets of the Amsterdam copy indicate that the paper used in its eighteenth-century rebinding came from Amsterdam (which does not mean that the book has been there continuously). Close scrutiny of the manuscripts may provide indications that will allow us to reconstruct other aspects of their history and reception. Stones, ‘Les débuts de l’héraldique’. Noted by Ward, Catalogue of Romances, I, 341. Roger Middleton has observed (private conversation) that the leaves on which these names are written may not have formed an original component of the manuscript. 10 11

THE ‘LANCELOT-GRAAL’ PROJECT

173

We aim to create a resource in which we can see the pages in these manuscripts and their component elements, to use it ourselves to ask and answer questions prompted by the manuscripts and their physical, textual and pictorial appearance, and to make it available to others. The Web component We are creating a computer database as a means to preserve, store, analyse, compare, discuss, and eventually disseminate, the manuscript pages in these copies of the ‘Lancelot-Graal’ and our work on them. The impact of the World Wide Web is making itself felt in manuscript studies. It has become possible through the web to transmit large numbers of high quality images and to accompany them with descriptors of various kinds. The present project is one of several that draw in some measure upon pioneering efforts begun more than a two decades ago at the Bibliothèque Sainte-Geneviève in Paris and at Orléans under the auspices of the Institut de Recherche et d’Histoire des Textes of the Centre National de la Recherche Scientifique, based at that time on slides made of the illustration and decoration of manuscripts.12 The collections of microfilms of manuscripts in Austrian and other European library holdings made from the early 1970s for the Hill Monastic Manuscript Library at St John’s University in Collegeville, Minnesota, and the Vatican Film Archive at St Louis University in St Louis brought important manuscript resources to the United States and fostered early efforts to manipulate overwhelming quantities of pictorial data.13 More recently, the Biblioteca Apostolica Vaticana, the Bibliothèque Nationale de 12 Christine Baryla at the Bibliothèque Saint-Geneviève and Odile Lépinay and Gilles Kagan at the IRHT have been particularly helpful in explaining and demonstrating the scope and goals of their projects. Both initially used slides taken by François Garnier as a continuation and expansion of the wide-ranging campaigns of photography and exhibitions of manuscripts in French libraries organized in the 1950s by Jean Porcher. See Holtz and Kagan, ‘La numérisation des enluminures’. An early pioneer was Minzer; see Minzer et al., ‘Vatican Library Materials’. 13 I served on the advisory board of HMML from 1984 to 1991 and co-authored a pilot project sponsored by Control Data Corporation to computerize images from HMML’s collections: see Stones and Arthur, Manuscripts, Illumination and PARIS. HMML and the Vatican Film Library are now collaborating on an Electronic Access to Medieval Manuscripts with Columbia University and the University of California Berkeley in projects funded by the Andrew W. Mellon Foundation that will be made available for a fee to scholars (see http:// sunsite.berkeley.edu/Scriptorium and http://www2.csbsju.edu/hmml/eamms/index.html). See now (2016) www.digital-scriptorium.org (30 Nov. 2016).

174

France, the British Library and the Bodleian Library have launched internet manuscript sites and other libraries in the United States and abroad have sites in preparation.14 In addition, there are other projects concentrating on corpora of particular texts, such as the Speculum maius of Vincent of Beauvais and the Roman de la Rose.15 The creation and dissemination of World Wide Web sites is governed in large part by issues of copyright. Manuscript sites have been developed primarily by institutions that own the copyright, and generally also the manuscripts themselves. Indeed, the present project will be posted on the internet only with the full accord of the keepers concerned, with whom agreements allowing the photography and the research on the project have been reached.16 My copyrighted site of images of medieval art and architecture 14 Of particular interest is the BNF site, The Age of King Charles V (http://www.bnf.fr/ enluminures/accueil.htm), based on details (only — no full page images) of a wide variety of different illustrated texts of the fourteenth century (now (2016) part of Europeana Regia, on whose advisory committee I served (http://www.europeanaregia.eu/en/manuscripts/librarycharles-v-family); there are now many full-page images); the Beowulf project (http://www. bl.uk/diglib/beowulf ), photographed with a digitizing camera and enhanced to allow superior enlargement so that letters and words previously covered by repairs to the parchment can be read (see now 2016 http://www.bl.uk/collection-items/beowulf ); and the Bodleian Library’s demonstration pages (http://www.rsl.ox.ac.uk/imacat.html) (now 2016 http://www.bodley. ox.ac.uk/dept/scwmss/wmss/online/medieval/ and http://image.ox.ac.uk/show?collection=bo dleian&manuscript), which offer different kinds of resolutions and enlargement options. The Morgan Library’s on-line cataloguing project (text only), now in progress, can be viewed in hard copy in the Morgan Library’s Reading Room (now, 2016, on line with images, through Corsair). I thank the late Fr Leonard Boyle, O.P. (late of the BAV), Aleksandra Orlovska (BNF), Andrew Prescott (BL), and Bruce Barker-Benfield (Bodley), as well as Hope Mayo (New York), and William Voelkle, Roger Wieck and Susan L‘Engle (Morgan Library) for helpful discussion of these projects and their methods. Among the American manuscript projects now under way I found the on-line exhibition of illuminated manuscripts in Utah collections (http://www2. art.utah.edu/Paging_Through/index.html) particularly helpful and thank its author, Elizabeth Peterson, for discussing in detail her methods for scanning, enhancing and page-making. The sites mentioned in this footnote are all available free of charge. 15 The Corpus of Vincent de Beauvais Manuscripts, directed by Monique PaulmierFoucart and Marie-Christine Duchene, is at the Atelier Vincent de Beauvais, Nancy (now, 2016, under the direction of Isabelle Draelants, it is housed at the Institut de recherche et d‘histoire des textes, Paris); the Roman de la Rose project is at Providence, Rhode Island, directed by Meradith T. McMunn and William McMunn; see also (2016) romandelarose.org, a joint project of the Sheridan Libraries of Johns Hopkins University and the BnF. I thank all of them for helpful discussion of their projects. 16 For their interest and cooperation, I thank F. A. Janssens (Amsterdam, Bibliotheca Philosophica Hermetica); Bruce Barker-Benfield (Oxford, Bodleian Library); Peter McNiven (Manchester, The John Rylands University Library); Pamela Porter (London, British Library).

THE ‘LANCELOT-GRAAL’ PROJECT

175

in France and England (http://www.medart.pitt.edu) is available free on the Internet,17 while my teaching sites are accessible only to University of Pittsburgh students by course password, limited to the term in which the course is taught.18 Our approach to creating World Wide Web materials for the ‘LancelotGraal’ project is based on collaboration with specialists in information science and telecommunications at the School of Information Science at the University of Pittsburgh. Kenneth Sochats and his graduate assistant Guoray Cai have pioneered conceptual mapping projects using Geographic Information Systems (GIS) software to cross-link different levels of information across a spatially defined spectrum.19 Our project treats the manuscript page as a conceptual map whose constituent elements can be identified and defined, plotted, compared, contrasted, linked to each other in any number of possible combinations, accompanied by annotations, and associated with information objects to assist the reader or analyst with further exploring the manuscript. These information objects can be sounds, text, images or other forms of information which will provide expert commentary, guides to characters or underlying themes and other information to support 17 This site has been developed and posted since 1994 by graduate assistant Jane Vadnal (History of Art and School of Information Science) who has also contributed her copyrighted images (so acknowledged) to the project. We have been assisted by undergraduates working for academic credit and by volunteers. Funds for research assistance in 1994 and scanning assistance in 1998 were provided by the History of Art and Architecture Department. Work on medart will eventually include monuments in other countries — Italy, Spain, Greece, Germany, Austria, Poland, Turkey. A paper on this and my teaching sites, with a brief mention of the Lancelot-Graal project, is ‘Three Sites’ reprinted in these essays. The sites I have developed at the University of Pittsburgh since 1994 adhere to the University’s posted copyright restrictions, which can be consulted on http:// www.library.pitt.edu/research/copyright/. 18 Courses under development under the rubric History of Art and Architecture are: Survey of Medieval Art and Architecture; Medieval Artistic Patronage, from Justinian to Jacques Coeur; Medieval Cities: Paris and London; Medieval Iconography: What is the Grail? Course outline pages only can be seen on-line without restriction. These courses have been developed with small grants from the College of Arts and Science at the University of Pittsburgh between 1995 and 1999, with the assistance of undergraduates and graduates working for credit, and with the help of volunteers. 19 See, for instance, Hirtle and Cai, ‘Classification Structures for Cognitive Maps’; Sochats and Williams, The 1997 Pennsylvania Technology Atlas; eidem, Communications and Networking Desk Reference. Limitations of space preclude my attaching here the appendix by Guorai Cai explaining the rational for our selection of GI software which I presented at the Harvard Conference. It is to be published elsewhere. See now Cai, ‘A GIS Approach’, and ‘Visualization’.

176

the analysis and understanding of the manuscript. The attachments will underlie the image of the manuscript so as to not interfere with its pictorial impression and interpretation. They will be summoned by ‘clicking’ on parts of the manuscript’s image or on buttons located outside the manuscript’s frame. We are not aware of any similar GIS application of the type we propose in existence today. The project to date Various members of our team have presented papers using the materials of our project at the University of Pittsburgh School of Information Science and Department of History of Art and Architecture colloquia (1996 and 1998); the Leeds International Medieval Conference in July 1996; the International Arthurian Congress in Garda, 1996; the Computers and the History of Art Conference in London in October 1998; the Medieval Academy of America, April 1999; and the International Medieval Conference in Toulouse, 1999. Publications from these conferences are in preparation or in press. Taxonomy The project began at an invited conference and study-session at the University of Pittsburgh in October 1995, during which we drafted a taxonomy of information about the manuscripts, texts and pictures of the project.20 We consulted published guides such as Vocabulaire codicologique,21 Iconclass,22 Aleph23 and the Getty Thesaurus,24 none of which quite correspond to the requirements of a non-religious text and its manuscripts and illustrations, nor can they accommodate the particular text-specific subjects of the ‘LancelotGraal’ which require new identification since no published subject-lists 20 Funding came from a consortium of departments and programs: the Honors College, the Medieval and Renaissance Studies program, the Faculty of Arts and Sciences, the School of Information Science, the departments of French and Italian, History of Art and Architecture, German, to all of whom I am grateful. Subsequent funding is described in articles on CHArt and Kodikologie, both reprinted in these essays. 21 Muzerelle, Vocabulaire codicologique. 22 Van de Waal et al., Iconclass. 23 As used on-site at the Index of Christian Art, Princeton University (see https://ica. princeton.edu (30 Nov. 2016). Currently (2016) in revision. 24 Peterson, ed., Art and Architecture Thesaurus.

THE ‘LANCELOT-GRAAL’ PROJECT

177

exist.25 Our taxonomy remains a working document that is added to as we work. Imaging26 The Amsterdam manuscript was photographed at the Bibliotheca Philosophica Hermetica by Alison Stones with the assistance of Martine Meuwese in the summer of 1996, using a Nikon FTN 35 mm single-lens reflex camera on a tripod and Ektachrome 64 daylight slide film. The Rylands Library manuscript was photographed on film from the same batch by the Rylands photographer in 1997. The Bodleian and British Library manuscripts were photographed by their respective photographic departments in 1998. We consulted specialists at the Morgan Library and the J. Paul Getty Museum for recommendations about the choice of film and photographic methods, and drew from my experience photographing in European and American libraries.27 We considered that the far greater cost of Ektachrome transparencies is not justified by the results, for example, by comparing the British Library’s CD-ROM, the Medieval Realm, made from slides, as opposed to the Maps CD-ROM, made from Ektachromes. Digitizing directly from the manuscripts was not an option. Digitizing Digitizing began at the University of Pittsburgh in Spring, 1998, by graduate students working for academic credit under the guidance of Alison Stones and Kenneth Sochats.28 We used a Nikon 35mm slide scanner LA-1000 to In the late Fall of 1998, after this conference, the Berkeley-Columbia Digital Scriptorium posted its taxonomy on http://sunsite.berkeley.edu/Scriptorium. 26 Photography of the Amsterdam manuscript was funded by Alison Stones and Elspeth Kennedy. Funds to purchase new complete sets of slides of the Bodleian, Rylands and British Library manuscripts were provided by the Central Research Fund of the Office of Research, University of Pittsburgh, in 1997 and 1998. I gratefully acknowledge this support which led to major funding described in Kodikologie, reprinted in these essays. 27 My photographs of manuscripts are deposited at the Conway Library, Courtauld Institute, London, and at the Photo Study Archive of the J. Paul Getty Center in Santa Monica (now Los Angeles); my images of monuments can be consulted on http://www.medart.pitt.edu (30 Nov. 2016), developed by my technical assistant Jane Vadnal, as noted above. 28 In Spring term 1998, Nancy Bennett (School of Information Science) established scanning guidelines in consultation with Kenneth Sochats (School of Information Science) 25

178

make two scans, an archival version at 2400 dpi and a working version at 300 dpi.29 A copy of each 300 dpi scan has been manipulated to a limited degree in Adobe Photoshop, beginning with version 4 and (since 1999) in version 5. Exact records have been kept for each scanned image describing the enhancements performed. We began with pages containing miniatures or champie initials and have also scanned text pages when a member of our team requested particular passages. By the end of Spring Semester 1999 we have completed the scanning of the illustrated pages in the Amsterdam and Royal manuscripts and have made headway with Additional and Rylands. Archiving While the project is in progress, the original slides are preserved in acid-free boxes in the History of Art and Architecture Department at the University of Pittsburgh. The archival scans are preserved in a locked folder on the Department’s Visual Resources Lab NT server. The 300 dpi images and their enhanced copies are preserved in separate folders on the NT server, which is routinely backed up to tape. Slides and scans will ultimately be maintained in optimum prevailing conditions at a specialist storage site. Migration strategies will be developed for its maintenance and transfer as standards of archiving improve. Systems For preliminary access and conference presentation while the project is in progress, the 300 dpi images have been copied to ImageAXS software. In Spring Semester 1999, eight graduate students have begun to develop Excel spread sheets containing information structured in layers: (1) image index (image scan identifier, place of preservation, library, shelf number, folio); (2) text (branch of the five-part cycle, edition references, variants); (3) rubric (space, actual words); (4) script (scribal hand, erasure, gap); (5) decoration and Elizabeth Peterson (University of Utah, Department of Art). Art History graduate students Katheryn Dimitroff, Marion Dolan, Joan Gauthier, Mary Laurent, Savannah Schroll, with Claire Edwards and Jeremy Jacobs (English Department) and Jane Vadnal (School of Information Science), participated in digitizing and entering descriptors on spread sheets during parts of Fall and Spring Semesters 1999. 29 In accordance with Bodleian Library restrictions, the slides of Douce 215 were scanned at 300 dpi only.

THE ‘LANCELOT-GRAAL’ PROJECT

179

(pen-flourished initial, border, champie initial); (6) miniature (format, frame, medium (gold, silver, colour), narrative content, component elements, specific incident, textual context); (7) references in secondary literature; (8) additional commentary. These layers are a preliminary step towards the full application of GIS software. Test examples were converted to Arc View software in late Spring 1999 and work on this continues in Summer 1999 under the guidance of Kenneth Sochats. Once we have transferred the information to the GIS system, the images and all layers of accompanying description and commentary can be electronically transmitted among the project participants who together and individually can make use of and contribute to the work of analysis. Thus, the accessor of the manuscript receives not only the manuscript artifact but also a rich corpus of ideas to help him or her better understand and appreciate the manuscripts. Indeed, the GIS-based system might be viewed as a scholarly forum for presenting ideas about the manuscripts, their construction and history. Unravelling the links between and among the different types of information about the ‘Lancelot-Graal’ is our narrow aim in this project, but as an intellectual exercise our tools and methodologies are generalizable to all kinds of other areas of conceptualization and analysis beyond the limits of humanities disciplines.

VIII Teaching and Research on Medieval Art on the Web: Three Sites

T

his article describes three types of web sites which have been developed under my direction at the University of Pittsburgh. The three types of sites were created with different, but complementary, aims in mind: two of them (http://www.pitt.edu/~medart (now [2016] http://www.medart. pitt.edu), hereafter MedArt and http://vrlab.fa.pitt.edu/stones-www/ stones-haa0240 together with http://vrlab.fa.pitt.edu/stones-www/stoneshaa1210) are primarily geared towards teaching and learning medieval art and architecture (both now obsolete), while the third (http://vrlab.fa.pitt. edu/STONES-WWW/VAlanc.html now http://lancelot-project.pitt.edu) is a site for research on the manuscripts of the Lancelot-Graal. All the sites are the result of collaborative efforts. The teaching sites were created and developed by myself and an art history and information science graduate student, Jane Vadnal, assisted by a large number of undergraduate and graduate students working for academic credit, and by volunteers, at the University of Pittsburgh between 1994 and the present.1 The research site is a collaborative project in which six specialists in Old French literature, manuscripts, illumination, and the history and reception of manuscripts (listed below) are collaborating with specialists in information science and telecommunications. The technical consultants for the Lancelot-Graal project 1 This article was first published in Computing and Visual Culture: Representation and Interpretation (CHArt Conference Proceedings 1998, ed. T. Szrajber), London, 1999, pp. 111–22. Since 1999 the projects described have undergone many transformations: the teaching projects are now obsolete since my retirement in 2012 but the MedArt and LancelotGrail projects are both on-going and have developed substantially in the intervening period.

MEDIEVAL ART ON THE WEB: THREE SITES

181

are Kenneth Sochats and his graduate student Guoray Cai, Department of Information Science and Telecommunications, School of Information Science at the University of Pittsburgh.2 This project began in the Fall of 1995 with a conference at the University of Pittsburgh and is continuing with the assistance of graduate students at Pitt. Each project, and accompanying site, has become refined and defined in the course of discussion with experts each with different training and point of view. We have considerably modified our goals and expectations in the light of these discussions which have played a critical role in the evolution of our thinking at all levels. A second important consideration is that of copyright. For the teaching sites were limited to those images on which I or Jane Vadnal personally own the copyright; other images could be posted only within the fair use guidelines of the University of Pittsburgh.3 The structure and content of the sites are the personal intellectual property of myself and Jane Vadnal. For the Lancelot-Graal project, the libraries which house the manuscripts retain the copyright on the images from those manuscripts; the dissemination of the project and its site occurs only with the express permission of the libraries concerned.

We received summer research assistance funding for the ~medart site from the History of Art and Architecture Department of the University of Pittsburgh in 1994. Funding from the College of Arts and Sciences has funded limited access teaching sites, including those for the two courses, HAA 0240 and HAA 1210, which I illustrate here: we received small research assistant grants from CAS in 1995–96, 1996–97, and 1997–98. A Hewlett Packard summer grant in 1997 funded a photography trip to Italy for a new component for the survey of medieval art and architecture, HAA0050, that is under development and will eventually also become part of ~medart. The student assistants have been trained by Jane Vadnal to do flatbed and slide scanning and enhancing, and some have participated in page creation. We scan at 300 dpi. 2 Slides for the Lancelot-Graal project were puchased with a grant from the Central Research Fund of the Office of Research of the University of Pittsburgh in 1997–98. I have profited from discussing photographic, technical, and copyright questions relating to the Lancelot-Graal project with specialists elsewhere, particularly at the manuscript departments and the photographic departments of the British Library, London, the Bodleian Library, Oxford, the Pierpont Morgan Library, New York, and the J. Paul Getty Museum, Los Angeles, as well as with the other participating libraries, the John Rylands University Library of Manchester and the Bibliotheca Philosophica Hermetica, Amsterdam. 3 The University of Pittsburgh Libraries’ site maintains and monitors a copyright guidance page: http://www.library.pitt.edu/research/copyright/ and the present activities are in compliance with the reguirements stated there.

182

The first site Jane Vadnal and I developed is http://www/pitt.edu/~medart (now http://www.medart.pitt.edu), hereafter MedArt (figs. 1–2). It was created as a visual resource for teaching medieval art and architecture and is aimed particularly at an audience of American undergraduate students as supplementary material on which students can base written descriptions of monuments and conduct a comparative analysis. The rational was developed at the University of Minnesota in a seminar run by Russell Burris of the Consulting Group on Instructional Design in the late 1970s and early 1980s.4 MedArt contains images of architectural monuments of the Middle Ages which I photographed over the last 25 years or so in the ordinary course of visiting medieval sites with a camera and tripod. These are slides which I used for years in teaching medieval art, and our goal in this site has been to provide a set of place-by-place pages of images structured so as to provide, as far as possible, an idea of what the building, or program of glass or painting, looks like. We supplement actual views with site maps based on out-of-copyright books such as Shepherd’s Historical Atlas and with plans and cross-sections from Dehio-Bezold.5 Fig. 3 is our plan of Chartes Cathedral, adapted from Dehio-Bezold, to which we have attached postage-stamp sized images which are clickable so that students can “walk” around the monument, inside and out, pausing to look at the buttresses and the portal sculptures, or the vaults and elevation, and the windows, on the inside. This has been modified as per the present fig. 3 through our updated Chartres site launched in 2006 (home page, fig. 4) following a seminar taught at Pitt and on site and a photographic campaign conducted by Pitt students with the permission of the Rector of the Cathedral. We began to supplement the images with brief references to the primary sources for dating the various components of the building, and to the secondary literature [listed now as a separate bibliography]. In collaboration with the Digital Research Library at Pitt we developed a database of Chartres images (http: images.library.pitt.edu/c/ chartres) which is linked to MedArt.

4 Published by Russell Burris, Sheila McNally and Alison Stones, Research on New Approaches to Teaching Art History, Minneapolis, 1982, esp. pp. 69–100. See also Donald R. Woods, Problem-based Learning: How to Gain the Most from PBL, Waterdown, ON, 1994. 5 Shepherd, William R. (1911), Historical Atlas, New York : Henry Holt and Co; Dehio, Georg and G. von Bezold. (1887–1901), Die kirchliche Baukunst des Abendlandes, 7 vols., Stuttgart: J. G. Cotta.

MEDIEVAL ART ON THE WEB: THREE SITES

183

MedArt has been deliberately developed without reference to a particular academic curriculum, and without narrative explanatory text, with the idea that its potential use as an image site should be unrestricted. Our aim has been to provide more images, more conveniently, than books do, so that these images may form a basis for description, analysis, and comparisons done orally or in written form by students themselves as part of an active learning process. Although I had an audience of American undergraduates primarily in mind, email responses to the site have suggested that the audience ranges very considerably, from elementary and high school students, to their parents, to university students all over the world, and to the general public at large. Most of the responses are requests for permission to download images for use in paper-writing — exactly the function for which the site was created. I routinely grant permission provided that the requester agrees to acknowledge my copyright or the copyright of other contributors to the site. Other requests (for use on personal web sites, commercial use, and the like) I decline. Many requests are for further information of a bibliographical nature, which I supply if I can do so quickly — answering detailed questions could become extremely time-consuming. Some of the responses are simply thank-you notes, which I appreciate. Future plans for MedArt are open-ended. We began to conduct photographic campaigns specifically for this project: during several summers my assistant Jane Vadnal has made photographic trips to France and has made pages for the site using her slides, most notably Vézelay, Moissac, Saint-Gilles, and Saint-Denis which can be reached from the France menu in MedArt (figs. 5, 6, 7). So far the project is limited to images of monuments in France and England; we have in hand a large number of images of monuments in Germany, Italy, Spain and other European countries which could be scanned and posted to the site in future campaigns. We would also like to include more bibliographical references along the lines of what we have done for Chartres. We regard MedArt as a project that could be sustained and on-going. In 2016–7 Dr Alison Langmead, Head of the Visual Media Workshop in the History of Art and Architecture Department at the University of Pittsburgh (http://www.haa.pitt.edu/collections/visual-media-workshop) and current custodian of MedArt is Principal Investigator of a project funded by the National Endowment for the Humanities examining the sustainability of MedArt (see her poster at iConference 2015, co-authored with Aisling Quigley, “Sustaining MedArt: Assessing the Persistence and Longevity of a Pioneering Digital Humanities Project” (http://hdl.handle.net/2142/73731).

184

1. Opening page of Medart (http://www.medart.pitt.edu)

2. Opening page of France on Medart (http://www.medart.pitt.edu/)

MEDIEVAL ART ON THE WEB: THREE SITES

3. Chartres, West Façade on Medart (http:// www.medart.pitt.edu/)

4. Opening page of Chartres on http://www.images.library.pitt.edu/c/chartres)

185

186

5. Saint-Denis windows opening page on Medart (http://www.medart.pitt.edu/) 6. Saint-Denis, Infancy window, Annunciation and Nativity panels on Medart

MEDIEVAL ART ON THE WEB: THREE SITES

7. Saint-Denis, Jesse Tree window on Medart

8. Home page for on-line course materials for HAA 0240, Medieval Artistic Patronage, Spring 2007

187

188

9. Opening page of HAA 1210, Medieval Iconongraphy, Spring 2010

10. Continuation of opening page of HAA 1210, Medieval Iconongraphy, Spring 2010

MEDIEVAL ART ON THE WEB: THREE SITES

11. Continuation of opening page of HAA 1210, Medieval Iconongraphy, Spring 2010 12. Continuation of opening page of HAA 1210, Medieval Iconongraphy, Spring 2010

189

190

The second web project is not just a single site but a cluster of several teaching sites, or rather course components that have the potential to be recombined in different ways according to the goals of a particular course. I developed components for the following courses: HAA 0050 Survey of Medieval Art and Architecture (http://vrlab.fa.pitt.edu/stones-www/stoneshaa0050); HAA 0240 Medieval Artistic Patronage (http://vrlab.fa.pitt.edu/ stones-www/stones-haa0240, fig. 11); HAA 1200 Medieval Cities: Paris and London (http://vrlab.fa.pitt.edu/stones-www/stones-haa1200); HAA 1210 Medieval Iconography: What is the Grail ? (http://vrlab.fa.pitt.edu/stoneswww/stones-haa1210 (figs. 9–11). I used these sites until my retirement in 2012; they are now obsolete although components using these materials can be used in conference papers and the like. These sites were aimed at undergraduates enrolled in particular courses in different aspects of medieval art and architecture as part of their B.A. degrees at the University of Pittsburgh; the goal was again to conveniently provide students with many images, unaccompanied by narrative discourse or other commentary, which I delivered in lectures. Sometimes I lectured from the pages on the sites: the disadvantage was (in the 1990s and early 2000s) that the pages were slow to load directly from the web. We experimented with storing pages for a lecture on a zip disk for speedier loading in the classroom. Another option wass to show slides in the classroom and reserve the web pages for out-of-class review. These teaching sites were limited in access because they included images on which I did not own the copyright. They were available by protected password only to students registered in the relevant course. However, important components of these courses were drawn from and linked to MedArt on the one hand and to the Lancelot-Graal materials on the other (the latter also protected by limited access password and limited to selected images). Figs. 5–7 show components of the site on the Royal Abbey of Saint-Denis, presented in the context of http://vrlab.fa.pitt.edu/stoneswww/stones-haa0240, my course on Medieval Artistic Patronage, in a unit on Abbot Suger (d. 1151). This component of the Patronage course also forms part of the Saint-Denis site on MedArt and can also be accessed independently (now [1 Dec. 2016] http://www.medart.pitt.edu/image/ France/St-denis/sdenmain.html ). The idea for structuring a course around the Grail (figs. 9–11) was inspired my research on the Lancelot-Graal project, the third of the sites presented here. The idea in this course was to treat the Grail as a “problem”

MEDIEVAL ART ON THE WEB: THREE SITES

191

to be “solved” by the students working through a series of tangentially related topics, most of which would find a place in a traditional course in medieval iconography (elements from the Life of Christ, the illustration of the Pseudo-Gospels, the cult of relics and its artistic manifestations). The course also provided a broad introduction to the Lancelot-Graal project for students unfamiliar with this Arthurian text and its anthropological and art-historical aspects and was deliberately cross-disciplinary in its approach. I note again that copyright restrictions on manuscript material severely limited what could be offered in the public domain. Although a manuscript specialist, I have found that creating new pages of manuscript material for teaching is possible only on limited access sites or by linking to existing sites created by libraries. Consequently I am grateful to the keepers of the manuscripts in the Lancelot-Graal project for allowing research on the project to proceed. While the project has been under development, the images are not posted to the web. Since I presented this talk in September [1997], I have taught a graduate seminar on the project at the University of Pittsburgh in which students have worked out standards for scanning and enhancing slides of manuscript pages and are developing the application of descriptors to the pages.6 In conclusion I emphasize again how important to the implementation of these projects the collaborative activity and attendant discussion has been. Work on these sites has also shown the importance of the interrelations between teaching and research, and within and among different components of related material. Seeing and creating links, and showing students how to see and create them, is where the excitement and the challenge of the web are to be found. Its potential seems limitless and will continue to evolve.7

A more detailed presentation of this site was my paper for the Manuscripts Conference sponsored by Harvard University in October, 1998, reprinted in these essays. 7 The Electronics Committee of the International Center of Medieval Art, to which I was elected chair 1999–2002, is now providing a forum for notices of web projects and discussion. See http://www.medievalart.org/. 6

IX Towards a Comparative Approach to Manuscript Study on the Web: the Case of the Lancelot-Grail Romance (With Ken Sochats) 1. Introduction

T

he Lancelot-Grail is the most popular version of Arthurian romance, surviving complete or in part in some 200 manuscript copies made between c. 1220 and 1504 (Woledge 1954 71–79, 1975 50–59). These manuscripts are housed today in libraries all over Europe and the USA but are available in the original only to scholars. Whereas some can be consulted complete in digitized form on the Bibliothèque nationale de France (BnF) site Gallica or on Mandragore or Digital Scriptorium (illustrated pages only), on the French provincial libraries site Enluminures, or on the sites of individual libraries,1 many more are available only as selected illustrations in secondary literature. Some manuscripts are fully illustrated (London, BL First published in Kodikologie und Paläographie im digitalen Zeitalter, 2, ed. F. Fischer, C. Fritze, G. Vogeler, with B. Assmann, M. Rehbein, P. Sahle, Norderstedt, 2010, pp. 33–42. For a fuller analysis of the entire False Guinevere episode see now A. Stones, “Illustration et stratégie illustrative dans quelques manuscrits du Lancelot-Graal” in Quand l’image relit le texte (Actes du Colloque international 15–16 mars 2011), ed. M. Pérez-Simon et S. Hériché Pradeau, Turnhout, 2012, pp. 99–116, reprinted in these essays. 1 On Gallica: BnF fr 95, 344, 16999; Mandragore: BnF fr 105, 111, 113–116, 117–120, 122, 1422–1424, 9123, 19162, 24394); on the French provincial libraries site Enluminures: Tours, BM 951; Le Mans MM 354; Dijon BM 527: Digital Scriptorium: University of California, Berkley, Bancroft Library 106, 107; individual library sites: The John Rylands University Library of Manchester, French 1; Cologny-Genève, The Bodmer Library MS 147 (on ecodices); New Haven, Yale University, Beinecke 229. And now many BnF and BL

TOWARDS A COMPARATIVE APPROACH

193

Add 10292–4 has 748 illustrations, the most of all), while others contain a single picture at the beginning of the major textual subdivisions or Branches into which the lengthy episodic narrative is divided (L’Estoire del saint Graal, Merlin, Suite vulgate du Merlin, Lancelot (with its own subdivisions), La Queste del saint Graal, La Mort Artu). Though attributed to authors Robert de Boron (Estoire, Merlin) and Gautier Map (Queste, Mort Artu), most of the manuscripts were made by anonymous scribes, decorators and illuminators. Some were made for or acquired by famous collectors — Jean de Berry († 1416), Jacques d’Armagnac († 1477) — but most patrons are unknown. We aim to determine what kinds of people found these texts interesting, where, and what aspects of the text the patrons and makers found compelling. Why the Lancelot-Grail? In a previous research project organized by Alison Stones, a team of 15 specialists in literature, palaeography, codicology and history of art drew together an illustrated catalogue and essays on the 45 (now 46) manuscripts and fragments of the romances of Chrétien de Troyes (see bibliography). In the project described in this paper we aim to make the manuscripts of the Lancelot-Grail and our findings available not just in print but also on the web, and to use web-based technology in the analytical and presentation processes. The Lancelot-Grail Project began as a multi-institutional cross-disciplinary collaboration in the late 1990s, based at the University of Pittsburgh.2 2. Goals and Methods Our aim is to enable on-line navigation of the manuscripts of the LancelotGrail romance and their illustrations in several ways, both synchronic and diachronic. We explore how the spatial analysis based on GIS concepts can manuscripts. Photographed for the Lancelot-Grail Project: olim Amsterdam BPH 1, Brussels, BR 9246 and 9627–8, Copenhagen, KB Thott 1047, London, BL Add 10292–4 and Royal 14 E.III, Le Mans MM 354, The John Rylands University Library of Manchester, French 1, Oxford, Bodl. Douce 215, Rennes BM 255. 2 Technical collaborators: Ken Sochats (Information Science and Telecommunications, University of Pittsburgh); Guoray Cai (Information Science, Pennsylvania State University), research assistant Jane Vadnal (Pittsburgh); Medieval French: † Elspeth Kennedy (Oxford); Medieval French, History of the Book and Ownership: Roger Middleton (Nottingham); Medieval French and Codicology: Keith Busby (Wisconsin); Medieval Art History: Alison Stones (Pittsburgh), Martine Meuwese (Leiden); graduate students Katherine Dimitrova, Marion Dolan, Julia Finch, Courtney Long, Kathryn Martin, Karen Webb (Pittsburgh) and Irère Fabry (Paris–III); Palaeographical Consultant: Michael Gullick.

194

be used in non-traditional applications, treating the manuscript page as a conceptual map in which different levels of information may be overlaid, using Active Server Pages to move within an individual folio, from a folio to an episode, and from an episode to a branch — or vice-versa. In the absence of a medieval term we define an ‘episode’ as a sub-section of a ‘branch’, a sequence of text and picture which concentrates on a particular hero or event (such as the False Guinevere in the Lancelot; Maritime Adventures in Estoire and Queste). The term ‘branch’ (branke, branche) is medieval: it is used in several manuscripts to distinguish the major subdivisions of the text (Estoire, Merlin, Lancelot, Queste, Mort Artu), and sometimes to mark subdivisions within the Merlin and the Lancelot. Branch divisions have been followed by modern text editors. What can be learned is how different copies of the same text differ from each other in wording, picture, page layout and the like, prompting investigation of what those differences mean. 3. Pilot Project We selected three manuscripts from the same ‘workshop’ made in Northern France or Flanders c. 1310–1325: London, British Library Additional 10292–4 and Royal 14 E.III, and a third copy divided among olim Amsterdam, Bibliotheca Philosophica Hermetica MS 1, Oxford, Bodleian Library MS Douce 215 and Manchester, The John Rylands University Library MS French 1. We chose these because BL Add 10292–4 is the most fully illustrated of all; it is complete; it contains the date of 12 February 1316 (1317 new style) ‘carved’ on a tomb on f. 55 in BL Add. 10292; and it was the basis for Sommer’s 1907–1913 edition of the text. This phase was funded by the National Endowment for the Humanities and by Visiting Fellowships for Alison Stones at All Souls and Magdalen Colleges Oxford and Corpus Christi College Cambridge. We worked from printouts from old black and white microfilms, from first-hand study of the originals, and from new colour slides shot by Alison Stones at Amsterdam or purchased from the libraries (British Library, Bodleian Library, Rylands Library). Pittburgh students scanned the slides with a view to eventually devising a web site. At the time, digital images were not available but the scanned images gave us useful working copies, particularly for text pages. In subsequent phases we were able to acquire better images, as described below.

TOWARDS A COMPARATIVE APPROACH

195

4. Product Models To analyse the manuscript page as a conceptual map on which can be plotted picture (subject, treatment), decoration (minor and major initials, penflourished, champie or foliate initials), text (episodes and events, names of characters), layout (columns, lines) and margins (decoration, annotations, blemishes) we developed a taxonomy of descriptors based on the contents of single pages (layout, script, decoration, illustration, text, marginalia, notes and annotations, physical signs of use and wear and tear, together with subdivisions of these categories). This was presented at the New Directions in Medieval Manuscript Studies conference at Harvard organized by Derek Pearsall in 1998 (Stones 2000). Each of the mark-up instances has associated spatial and descriptive data further describing that particular mark-up. By using a standard descriptive framework, we were able to show how components of one page might be linked to similar components on other pages or in other manuscripts. We built authority lists for the subjects of the illustrations of all the branches of the text (fig. 1).3 5. Developing a Series of Navigational Options Navigating comparatively by ‘branch’ and ‘episode’ (cluster of related scenes) gives new insight in the use of the manuscripts. Table 1 compares British Library (BL) manuscript Add 10292 with olim Amsterdam, Bibliotheca Philosophica Hermetica (BPH) ms. 1, ii. BL Add 10292 will be posted on the web later through DIAMM (see below). Eventually the images will be included in the comparative table and each folio will be linked to the folio analysis. In this excerpt, BL Add 10292 gives many more illustrations than Amsterdam although the text of the latter is not abbreviated. However the pictures in Amsterdam often include more detail than those in BL Add 10292 where the action is played out over more scenes. A major difference is that BL Add. 10292 emphasizes the legal aspects of the challenge to Queen Guinevere’s legitimacy and the downfall of the lying False Guinevere and 3 See the appendices in Stones 1977, 1988, 2000, 2008, 2009, 2013; Blackman 1999, Meuwese 1999, Shailor 1984.

196

her sponsor. BL Add incorporates a significant detail borrowed from legal illustration (cf. Gratian’s Decretum) where the motif of a lighted taper is used to indicate the pronouncement of excommunication and anathema. The patron of BL Add is likely to have had a particular interest in the legal aspects of the text, unlike the patron of Amsterdam. 6. Extending the Parameters of the Project This phase was funded by the Fulbright Foundation and by a Digital Innovation Fellowship from the American Council of Learned Societies for Alison Stones. Our goal was to obtain better quality images of the Pilot Project manuscripts and to broaden the scope of the project to include more manuscripts and to plot them across time and space. High resolution scans from the British Library of the illustrated pages in BL Add 10292–4 and Royal 14 E.III were purchased. Photography was done by the Digital Image Archive of Medieval Music (DIAMM).4 We had the illustrated pages in Amsterdam BPH 1 and Manchester, The John Rylands University Library French 1, Rennes BM 255 (c. 1220) and Le Mans MM 354 (c. 1285) photographed at high resolution by DIAMM. We obtained permission from the Bibliothèque nationale de France to download for free and incorporate, or to link, Lancelot-Grail manuscripts from BnF sites Gallica and Mandragore. Downloading and resizing were done by University of Pittsburgh Art History graduate students. We obtained permission to post images from the Bibliotheca Philosophica Hermetica, Amsterdam, and from Rennes BM 155 and Le Mans MM 354. The John Rylands University Library of Manchester has posted the images shot for the Project by DIAMM on its web site. Lengthy attempts to negotiate a contract between the University of Pittsburgh and the British Library resulted in failure. We are now planning to display BL images through DIAMM under the terms of DIAMM’s copyright agreement with the BL.

Table 1. The end of the False Guinevere episode, comparing BL Add 10292 and Amsterdam BPH 1, ii (BL Royal 14 E.III lacks the Lancelot section of the text); text references are to Sommer (vol. 4); contractions have been silently expanded. 4

Cf. the chapter by Julia Craig-McFeely in this volume 307–339.

TOWARDS A COMPARATIVE APPROACH

S IV 69. 17

S IV 72. 9

S IV 75. 16

S IV 76. 37

S IV 79. 4

BL Add 10292, f. 151r. Rubric: Ensi que li rois Artus tint la roine Genieure par le main et la bailla a garder le roi Galehot par deuant lor baronie, et les barons en orent grant pitie. King Arthur entrusts Queen Guinevere to King Galehot, witnessed by the barons. L, champie initial Text: Lors prent li roys la royne et en uait a Galehot et li liure par le main... BL Add 10292, f. 152r. Rubric: Ensi que on gete sentence sour le roy Artu. The pope enjoins King Arthur to leave his new wife and take back the old one, extending towards him a lighted taper, a motif derived from the canon law indictment of excommunication and the pronouncement of anathema. O, champie initial Text: Oor [sic] dist li contes que ensi est li roys Arthus départis de sa femme par le desloiaute de lautre Genieure... BL Add 10292, f. 153r. Rubric: Ensi comme li rois Artu se fist confesser dun hermite en son hermitage. King Arthur, repentant, confessing to the hermit Amustans in his church-like hermitage. T, champie initial Text: Tant dist mesires Gauvain au roy Artus son oncle... BL Add 10292, f. 153v. Rubric: Ensi que li roys Artu et se baronie oirent messe en.j. hermitage King Arthur and his men hear mass in an elaborate, Gothic, hermitage. Q, champie initial Text: Quant li roys ot ensi parler termite si giete vn souspir... BL Add 10292, f. 154r. Rubric: Ensi que li fause Genieure gist mesele et si uo (uo expunged) vint li rois Artu parler a li. The False Guinevere, with veiled head, lying on her deathbed with Bertolai at her side holding a Ciborium with a cross on top, repeats her confession to King Arthur and the barons. D, champie initial Text: Dame vous gisies en si dolereuse carire comme cele qui a tout le pooir du cors perdu...

197

Amst. BPH 1, II, f. 227v Rubric: Chi gist malade de meselerie li fausse royne et li roys Artur le vintveir The False Guinevere, crowned, lying on her deathbed with Bertolai at her side holding a covered vessel, repeats her confession to King Arthur and the barons. O, pen-flourished initial Text: Or dist li contes que ensi est li partis de sa femme par le desoiaute de lautre Genieure...

198

7. Comparative Selection and Treatment of Branches, Episodes, Folios Fig. 2 shows a section from a comparative page of the Maritime Adventures episode from Estoire manuscripts in Rennes, Le Mans, Amsterdam and Paris (BnF). The choice of champie initials links Le Mans and Amsterdam whereas long rubrics link Amsterdam and BnF and narrative scenes link Rennes and BnF: these findings point to levels of complexity in the transmission and illustration of these four copies which can be corroborated with further comparisons in order to reconstruct overall patterns of transmission and reception. After the second phase of the Lancelot-Grail Project the collected data allowed the following conclusions: The manuscripts vary substantially in selection, placing, and treatment of illustrations. Sometimes a champie initial is substituted for an illustration. Of special interest is the treatment of Solomon’s enchanted ship and its cross: sometimes as a cross on the sail, at other times a cross held by Tout-en-Tout in the ship. These pictorial variants depend on textual variants in the respective manuscripts, indicating that illustrators (or planners) paid careful attention to textual description (Stones 2009). A selector model is under development which will allow other manuscripts, branches, and episodes to be selectively compared. 8. The LANCELOT-GRAIL Manuscript Tradition Across Time and Space Using our expert art-historical and palaeographical analyses we plotted the chronological and geographical distribution of as many Lancelot-Grail manuscripts as possible, based on intensive research at the Bibliothèque nationale de France and on scattered manuscripts in many other collections (such as New York, Yale, Berkeley, Bonn and others). On the project web site we linked manuscripts either to pages created by ourselves or to existing sites of individual libraries. The laborious task of downloading, resizing, labelling, has been done by graduate students in art history at the University of Pittsburgh. The task continues. The interest in the Lancelot-Grail romance spread from Northern France in the early 13th century to England and Italy: the map in figure 3 shows this graphically and is a sample of a series of projected maps which will chart in 50-year increments the gradual spread of interest as reflected in patterns of collecting and gift-giving between c. 1220 and 1504, the date of the latest dated manuscript.

TOWARDS A COMPARATIVE APPROACH

1. Amsterdam, BPH, MS 1, ii, f. 202r (opening of the False Guinevere episode from the Lancelot branch reproduced courtesy of the Bibliotheca Philosophica Hermetica, Amsterdam). 2. Maritime Adventures (reproduction in greyscale).

199

200

3. Sample map of distribution of Lancelot-Grail manuscripts.

The goals of this phase of the project are first to use mapping to better identify where and when the manuscripts were manufactured, and then to relate differences in manuscripts to local political, cultural, economic and other conditions. We hope our plotting of changing patterns of interest in these stories may lead to correlations with the significant improvements made in measurement, navigation and other geographic technologies. 9. Future Plans The above outline indicates some of the directions that research on the Lancelot-Grail manuscripts have taken in the previous phases of research and development. The results of each phase can now be applied and exploited in more detail both on the web and in conference papers and published articles. Our goals are to make much more of our analysis available on the web and to unite in tabular form the illustrations with the descriptive and analytical research that has been carried out so far in the project. We hope our approach will be transferable to other manuscript projects.

X Stories in Pictures and their Transmission: A Comparative Approach to the Manuscripts of the Lancelot-Grail Romance

I

t may not be Troy with 300 MSS, but the Lancelot-Grail romance was among the most popular vernacular texts of the Middle Ages, surviving today in some 200 copies in whole or in part, produced between c. 1220 and 1504. Most of the manuscripts are illustrated, yet the choice, distribution, and treatment of images varies considerably from copy to copy. No two manuscripts contain quite the same set of pictures (though some come close). What constituted a ‘fully illustrated’ Lancelot-Grail manuscript? Did such a notion even exist in the Middle Ages for vernacular illustration? In this paper I compare the nine copies that contain the entire five-branch Lancelot-Grail cycle (Estoire, Merlin, with Suite vulgate, Lancelot with five sub-divisions of its own, Queste and Mort Artu), charting the patterns of their opening illustrations in order to suggest some conclusions about what factors determined their illustrative choices, by and for whom, how the treatment of the opening illustrations evolved in time and place, as a first step towards determining whether there actually was a concept of a ‘fully illustrated’ copy. The importance of pictures in this cycle is very well known: it is through seeing images of the adulterous relationship between Lancelot and Guinevere, painted on the walls of Morgan’s prison by Lancelot himself, that Arthur comes to realize that the rumours about the affair were true (BnF fr. 116, First published in Res gestae – res pictae: Epen-Illustrationen des 13. bis 15. Jahrhunderts, ed. C. Cipollaro and M. Theisen (Codices Manuscripti et Impressi, Supplementum 9, Vienna), pp. 40–51.

202

f. 688v by the Master of fr. 114; Arthur looking and reading captions, BnF fr. 112 f. 193v, compiled and written by Michel Gonnot, 1470).1 The only image of Lancelot actually making an image is a sculpted one — a statue of Guinevere so life-like that he kisses it. Images in other stories are sometimes equally poignant: in Le Roman de la Manekine the heroine is about to be burnt at the stake along with her infant, but kind friends substitute an image so lifelike as to deceive her would-be torturers; another example is the statue of Toute Belle weeping, from Guillaume de Machaut’s Livre du Voir Dit, or Christine de Pizan in Fortune’s Sale merveilleuse in the Livre de la mutacion de fortune (Munich, BSB, ms. Gall. 11, f. 53).2 And of course the story of Pygmalion underlay all of these. Part I: the Manuscript Tradition To Brian Woledge goes the credit for drawing up what is still the most useful list of Lancelot-Grail manuscripts. Published in 1953 and revised with a Supplement in 1975,3 it is the only study to take in the entire manuscript tradition, whereas the text editors have concentrated on the respective branches they have edited and none of them gives an overview of the reception of the text as a whole. So Woledge’s list remains the starting-point for any investigation of the complete corpus, although for our present purpose several moditifations must be made. Woledge, and indeed most of the text editors, was not interested in noting which manuscripts are illustrated. For a study of the illustrative tradition, then, we are on our own. If we start with Woledge’s list of the complete copies, 9 are listed.4 However from the point of view of the illustrations, two of those 9 can be eliminated. BnF fr 98, a 15th-century copy, has only minor decoration, marking the beginnings of major branches with a foliate initial and border, and dividing paragraphs 1 Many of the manuscripts discussed here are available on the web so I keep the illustrations in this article to a minimum. Others are clustered on my web site (http://www. lancelot-project.pitt.edu). See particularly the page entitled “Chronology and Geographical Distribution.” 2 These are illustrated in Stones, “Seeing the Walls of Troy”; see also Bloch, “Lancelot the Illustrator”. 3 Woledge, Bibliographie and Supplément. 4 Bonn, LUB, 526; Geneva, Bibl. Bodmer anc. Phillipps 1046 = Bodmer 147; London, BL Add 10292–4 (hereafter Add); Paris, Arsenal 3479–80; Paris, BnF fr 98, 110, 113–116, 117–120, 344.

STORIES IN PICTURES AND THEIR TRANSMISSION

203

with champie initials only.5 All in one volume, it has been attributed to the 15th century but nothing is known about its original patron, nor indeed of its early history. It was not part of the Royal Library at Blois but may have belonged to François 1er.6 The copy in the Bodmer library at ColognyGenève, ms. 147, listed by Woledge with its ex-Phillipps collection number 1046, is a compendium that lacks the Lancelot proper and so does not actually contain a complete Lancelot-Grail cycle. It also has many interpolations as noted by Françoise Vieillard, which put it in a special category on its own.7 Similarly, BnF fr 113–116 lacks the Suite vulgate du Merlin, and Ars 3479–80 and fr 117–120 give only a single composite miniature for the opening of the Estoire del saint Graal, and those very well-known miniatures emphasize Lancelot rather than the Grail, although the Estoire text is present.8 One manuscript must be added: Woledge did not know the ex-Phillipps collection manuscript that passed through H P. Kraus in New York in 1978 and was then, until December 2010, owned by the Bibliotheca Philosophica Hermetica in Amsterdam, MS 1. It is now in a private collection whose owner has chosen to remain anonymous.9 I published it as part of the same set as Fr. 1 in The John Rylands University Library in Manchester of which another part is Douce 215 in the Bodleian Library Oxford, making another complete set, though one which always lacked the Suite vulgate du Merlin and has many missing leaves.10 How do the manuscripts that contain the entire cycle (or almost) fit the rest of the manuscript tradition? They are not among the earliest copies — partial copies antedate the surviving complete manuscripts — nor do they include the latest. First I survey the entire manuscript tradition, then I examine more closely some of the iconographie relationships among the opening illustrations of the branches in the complete manuscripts.

Reproduced in black and white on Gallica (4 Feb. 2014). I thank Marie-Pierre Laffitte for this information. 7 Fully reproduced in colour on e-codices with reference to the analysis by Vielliard, Manuscrits français, pp. 46–50. 8 Ars 3479–80, BnF fr 113–116 (but not apparently 114 or 115 as of 4 Feb. 2014), 117–120 and fr 344 are reproduced in colour on Gallica; BnF fr 112, Jacques d’ Armagnac’s abridged version, in 3 volumes is also in colour on Gallica. 9 Sotheby’s 7.XII.2010, lot 33. See also Stones, Gothic Manuscripts, Part I vol. 1, ill. 603; Part I vol. 2, pp. 354–368, Cat. no. III–75, and back cover plate. 10 Stones, “Another Short Note,” reprinted in these essays. 5 6

204

The manuscript tradition: geography and chronology The tradition extends chronologically from the beginning of the 13th century with Rennes BM 255 (containing Estoire, Merlin without the Suite vulgate, and Lancelot with lacunae) to 1504, the date of the last dated manuscript, a copy of Estoire on paper, written by Fr. J. de Rochemure, Paris, BnF fr 1427.11 By then printed editions were also in circulation, containing illustrations which are lacking in many of the fifteenth-century manuscripts — but I leave the printed copies aside here.12 The manuscripts (about 180 in whole or in part) are mostly of French manufacture, with half a dozen copies apparently made in England, and a dozen or so made in Italy. As to numbers of pictures, Add is the copy that contains more pictures than any other: a grand total of 748. That is a lot — too many for us to consider all of them here. No other manuscript approaches Add for numbers: Bonn is next with 346 and BnF fr 344 closely follows with 344 (sic); Jacques d’Armagnac’s copies, BnF fr. 112 (special version) and 113–116 have 258 and 209 respectively, the work of Evrard d’Espingues; BnF fr. 117–120 and Ars 3479–80 were sold by Jacques Raponde in 1405, one copy to Jean de Berry († 1416), the other to Philippe le Hardi († 1404) or, more likely, his son Jean de Bourgogne (1404–1419), and have 130 and 131 illustrations respectively. 11 But these totals are deceptive: for instance, British Library Royal 14 E.III, closely related stylistically to Add and Amst/Douce/Rylands, has more pictures for Estoire, Queste and Mort Artu than do the comparable branches in Add and Amst/Douce/Rylands; but Royal lacks the Lancelot so its overall total number of illustrations is only 116. If we consider manuscripts containing only parts of the cycle, one can reconstruct several pairs that were most likely parts of the same set of volumes. Examples are BnF fr 748 containing Joseph and Merlin, and 754 with Lancelot Part I, probably made early in the 13th century and probably southern; BnF fr 95 (Estoire, Merlin, Suite Vulgate, Sept sages and Pénitence Adam) and Yale 229 (Agravain, Queste, and Mort Artu), probably made in 11 The Lancelot in Rennes 255 ends at Sommer vol. IV 220.33; see Stones, “The Earliest”; Cassagnes-Brouquet, Romans de la Table Ronde. I find it convenient to subdivide the Lancelot as follows: LI = En la marche de Gaulle, L2 = Journey to Sorelois; L3 = Conte de la Charette, L4 = Suites de la Charette, L5 = Agravain. The Bonn copy was the base manuscript of the edition by Poirion et al., Le Livre du Graal.

STORIES IN PICTURES AND THEIR TRANSMISSION

205

Thérouanne c. 1300; St Petersburg NLR Fr. F.v.XV.5 (Estoire with Joseph interpolations) and Oxford Bodl. Rawl Q.b.6 (Lancelot complete, Queste and Mort Artu), made probably in Paris in the early 14th century, part of the huge wave of ‘post-Honoré’ manuscripts; BnF fr 9123 {Estoire, Merlin and Suite), and Ars 3481 (Lancelot part 1), both unusual in prefacing the text with lists of the rubrics that also accompany each miniature, both copies made in Paris in the second quarter of the 14th century and illustrated by the Sub-Fauvel Master;14 perhaps BnF fr 16999 (Lancelot Parts 1–4) and Ars 3482 (Merlin, Suite vulgate, Agravain, Queste, Mort Artu), probably Parisian, made in the mid-14th century; and the pair adapted by Guillaume de la Pierre for Jean-Louis de Savoie († 1487), Brussels, BR 9246 (Estoire) and Paris BNF fr 91 (Merlin, Suite Vulgate). All these pairs either lack a branch to complete the set or are incomplete in one way or another. At the same time there are several sets of sister volumes. Within the group of complete MSS, BnF fr 112 and 113–116, Jacques d’Armagnac’s copies, were painted in the main by Evrard d’Espingues in the 1470s; Add and Amsterdam/Douce/Rylands were produced by the same team as Royal 14 E.III; BnF fr 344 and ex-Philipps 1047 both come from the same Metz workshop;15 Tours BM 951 and BnF fr 12580 were both substantially illustrated by the Hospitaller Master, in Acre or Cyprus before 1291 or in Paris; Bonn LUB, S 526, whose colophon tells us that Arnulphus de Kayo wrote it in Amiens in 1286, gave rise from the stylistic point of view to copies associated with Thérouane or Cambrai in the following decade (BnF fr 110, BL Add 5474, BnF fr 5237); another Thérouanne cluster of at least 2 copies comprises BnF fr 749, Bodl. Ash. 828, Turin BN L III. 12 and Bologna AS b.l.bis; the pair BnF fr 19162 and 24394 are also associated with the region of Thérouanne, probably Saint-Omer because their artists are close to the Psalter of Saint-Omer, London, BL Yates Thompson 43 written in 1276.

The 1488 copies by Le Bourgeois and du Pré, and Vérard’s 1494 copies are listed in Woledge, Bibliographie p. 77 and Supplément p. 54. 13 On Jacques Raponde, see Büttner, “Jacques Raponde, ‘marchand’ de manuscrits enluminés”. On Jean de Berry and Jacques d’Armagnac’s copies see Blackman, Jacques d’Armagnac, esp. vol. 1, ch. 5; ead. “A Pictorial Synopsis of Arthurian Episodes”; Hermant, “Le Lancelot-Graal de Jean de Berry et Jacques d’Armagnac” 14 An artist emanating from the circle of the anonymous Fauvel Master, the latter so-called for his very polished tinted drawings in the Roman de Fauvel, Paris, BnF fr 146, but I call the Sub-Fauvel Master a related artist whose work is much less fine, painted in full colour but sketchily drawn. See Stones, “The Artistic Context of le Roman de Fauvel”. 12

206

Le Mans MM 354 and BnF fr 770 can I think be attributed to Douai c. 1275–85 or so, and Kathy Krause has associated 770 with the patronage of the Ponthieu family.16 For the rest, there is little to go on and I have many question marks. Part 2: Comparing pictorial cycles How did the pictorial cycle or cycles come into being and how did they evolve? A comparison of the treatment of opening miniatures in the textually complete copies can elucidate to a degree some possible interrelationships among these manuscripts. However, it quickly becomes apparent that the incomplete copies also have a contribution to make to the evolution of the picture-cycles and need also to be taken into consideration. Even the copies that lack historiation have an important part to play in establishing markers in the text that eventually became foundations for the placement of images. Most of the earliest manuscripts contain only minor decoration — pen-flourished, champie, or other small initials — but these display patterns of placement that are occupied by miniatures or historiated initials in the illustrated copies.17 The patterns are not uniform, however, and certain manuscripts, most notably Add, the most fully illustrated complete manuscript, departs from and elaborates upon patterns pioneered in earlier and contemporary manuscripts.18 The illustrations of openings in the fully illustrated and related copies will serve as representative of the structuring principles as a whole. They typify some of the complexities, similarities, and differences among the illustrations in the rest of the cycle. 1. Bonn LUB, S 526 – BnF, fr 110 – BnF fr 344 (Appendix 1) This cluster represents the three earliest complete manuscripts. Bonn LUB 526 (written in 3 columns by Arnulphus de Kayo in Amiens in 1286, as 15 I know ex-Phillipps 1047 only from the Sotheby sale catalogue of 1 July 1946, lot 10, pl. XV. It is currently in the collection of the Lebaudy family of Versailles. 16 Krause, “The Manuscript Contexts of the Fille du Comte de Pontieu”. 17 See Stones, “Un schéma d’emplacement pour l’illustration de l’Estoire del saint Graal.” Howard Bloch has rightly pointed to narrative switches as obvious places for illustrations, but in my view the situation is rather more complicated. Bloch, “Lancelot the Illustrator”. 18 I examine a sequence of illustrations where Add gives a particular pictorial emphasis to legal aspects of the Lancelot in Stones, “Quelques lecteurs”.

STORIES IN PICTURES AND THEIR TRANSMISSION

207

noted above) is the only one that is dated, and BnF fr 110, also written in 3 columns, is similar stylistically to Bonn, although its curvilinear borders would seem rather more developed than those in the Bonn manuscript and are therefore likely to be later, closer artistically to Boulogne-sur-Mer BM 192, a Guillaume d’Orange cycle written in 1295, most likely in Cambrai or Thérouanne.19 That Arnulphus says he wrote Bonn 526 in Amiens suggests that his presence there was a special circumstance and that he did not normally work there. The work of his artist, too, has more in common with books from Saint-Omer, Thérouanne and Cambrai than with Amiénois production.20 BnF fr 344 on the other hand can be attributed to Metz. Its date is problematic. It was attributed to the second quarter of the 13th century by F. Gasparri in Album de manuscrits français du XlIIe siècle: Mise en page et mise en texte, ed. M. Careri et al., Rome, 2001, 123–6, no. 31 (f. 53). If this date is right it can only apply to the writing, not to the illustration, but a date c. 1276 would be possible by analogy with the tiny illustrations in Thomas de Cantimpré’s De rerum natura, Paris, BNF lat. 523A of that date. However, the illustrations are better compared with the style of the Charter of Sainte-Glossinde de Metz, Metz, AD de la Moselle, H 4085 (5), written by Othin de Biancourt in 1293, suggesting that BnF fr 344 could fall later in the chronology of the Lancelot-Grail than the Bonn copy and substantially later than the earliest (partial) copies.21 It is written in 2 columns not 3 and it lacks rubrics which are present (and different) in Bonn and fr 110, features which align it with early manuscripts; I leave the question of rubrics aside here.22 What is most striking about the openings in these three manuscripts is their use of multicompartment miniatures to mark important text breaks, though they differ as to which breaks are so marked — calling in question the notion of what constitutes a ‘branch’ since there is no uniformity among 19 See Stones, “La production de manuscrits littéraires aux environs de 1300: entre Cambrai et Saint-Omer”; ead., Gothic Manuscripts, Part I vol. 2, Cat. III–123, pp. 547–549. 20 See ibid., Cat. Ill–121, pp. 531–543. 21 For the Charter, see ibid.,Part II vol. 1, Cat. IV–2, 8–9; for BnF lat 523A see Cat. IV–3, 9–10; for BnF fr 344, see Part I vol. 1, pp. 40, 71,120; Part I vol. 2, pp. 447, 540; Part II vol. 1, p. 23, ill. 82, 47, 218. 22 A much later case of copying of pictures and rubrics from one manuscript to another, is that of BnF fr 105 and Brussels, BR 9246, see Stones, “L’Estoire del saint Graal dans la version adaptée par Guillaume de la Pierre pour Jean-Louis de Savoie” reprinted in these essays.

208

1. Mort Artu: The Damsel of Escalot offers her sleeve to Lancelot; Lancelot fighting in the Tournament of Winchester. Bonn, UB, S 526, f. 443v

3. Suite vulgate de Merlin: Merlin sets fire to the tents of the rebel barons. London, BL, Add. 10292, f. 102r

4. Suite de la Charette: Margondas surrenders to Lancelot. London, BL, Add. 10293, f. 205v

STORIES IN PICTURES AND THEIR TRANSMISSION

2. Estoire del saint Graal: The hermit-author in his chapel. Private Collection (olim Amsterdam, Bibi. Philosophica Hermetica 1), f. lr

209

210

even the closely-related pair, Bonn and fr 110. In both, multiple miniatures distinguish several openings from the historiated initials (first quire in Bonn) and single-column miniatures used in both for the rest of the illustrations: Estoire (6 miniatures in Bonn, page missing in fr 110), Lancelot Part 1 (2 miniatures in Bonn, 3 in fr. 110), Lancelot Part 5 (Agravain) (2 miniatures in Bonn, 1 in fr 110), Queste (2 in Bonn and fr 110), and Mort Artu (2 in Bonn, 1 in fr 110). The other branches, Merlin and Suite vulgate, Lancelot Parts 2, 3, and 4, have a single-column miniature in Bonn and fr 110 (no illustration for (Charette in fr 110). There are examples among the manuscripts containing only parts of the cycle for the use of multiple miniatures: Le Mans MM 354 and its sister manuscript, BnF fr 770 both have clusters of scenes (though neither the same number nor the same subjects) at the opening of their Estoire. Particularly problematic here and in the manuscripts compared below is the question of the Suite de la Charette, whose text opening is not even marked decoratively: the manuscripts make different choices about what to depict towards the end of Charette and the beginning of its Suite. In fr 344 on the other hand the opening illustration of Estoire is a historiated initial like those in the body of the text, while a four-part miniature and a historiated initial open the Merlin, preceded by a miniature at the end of Estoire, and a six-part miniature and historiated initial open Lancelot Part 1, preceded by a six-part miniature at the end of the Suite vulgate. The other branches of Lancelot have only a historiated initial, showing that they were not considered important break points. Queste opens in fr 344 with a large four-part initial A, while the opening of Mort Artu is missing. If we assume that a pictorial emphasis through the use of multiple scenes is a means to express the importance of a particular branch, then these manuscripts differ as to what was considered important — Merlin, Lancelot Part I and Queste in fr 344; Estoire, Lancelot Parts 1 and 5, Queste and Mort Artu in Bonn, and Lancelot Part 1 and Queste in fr 110. But lacunae in fr 110 and fr 344 mean the picture is incomplete. And a later owner (14th or 15th c.?) disagreed with the structure offered by the multiple miniatures in Bonn, adding his own division into 9 Books by the letter .L. and a number, and a title, in the top margins, of ff. 1 (Estoire = De Ioseph de Arimathie .L. .I.), 60 (Merlin = lci comence de Merlin .L. .II), 82 (Suite vulgate = Ici comence des premiers faiz le roy Artu .L. .III.), 171 (Lancelot Part 1 = Ici comence de la marche de Gaulle .L. .IIII), 259 (Lancelot Part 2 = Ici fine de la marche de Gaulle. Et comence de Galahot .L. .V.), 307 (Lancelot Part 4 [Suite de la charette] = Ici fenist de Galahot Et commence la premere partie de la queste Lancelot .L. .VI.), 335

STORIES IN PICTURES AND THEIR TRANSMISSION

211

(Lancelot Part 5 [Agravain] = Ici comence la seconde partie de la queste Lancelot .L. .VII.), 406 (Queste = Ici commence dou saint graal .L. .VIII.), 443v (Mort Artu = Ici comence la mort dou Roy Artu et des autres .L. .IX.). Despite the difference in format, several subjects are similar across all three copies, notably Agravain, with Agravain riding to the tent with bier and grieving damsel, and Queste in Bonn and fr 110 with the damsel summoning Lancelot; but fr 344 has the Death of King Mordrain, the only manuscript to include this scene as an opening so far as I know. There is ambiguity about the Suite vulgate and its opening, reflected in the choice of a different subject in each manuscript, and this will also characterize the illustration of this branch in the other copies. Striking is the emphasis in Bonn for Mort Artu on the Demoiselle d’Escalot, without parallel anywhere in the illustrative tradition so far as I know (fig. 1). 2. Add – Amst/Douce/Rylands (Appendix 2) These in turn can be cross-compared with the three earlier manuscripts. Like the pair Bonn – fr 110, Add and Amst/Douce/Rylands are close in terms of production: and argument could be made for both to be illustrated by the same artist at different points in his career, or if not that, then closely related artists were responsible, perhaps at the same time. And the incomplete related copy Royal 14 E. III is also part of the same equation for the three branches it contains.23 What emerges is a much closer grouping, lacking the use of multi-compartment miniatures apart from the opening miniature in Estoire in Add and the subdivided miniature for Merlin where both compartments occupy the width of a single column. Nevertheless there are similarities and differences in pictorial emphasis within this cluster of three copies: for Estoire, the opening miniature in Add groups two scenes: the hermit-author before his altar blessed by the Hand of God, followed by Christ in a cloud handing the book to the hermit-author. In Royal these episodes are both on the opening page but are single-column miniatures in the first and third column, while in Amst the hermit-author is shown in his chapel. The miniature is The major painter also did the frontispiece in Saint-Omer BM 270, a Psalter of Tournai Use offered by Guilbert de Sainte-Aldegonde to the Chartreuse of Longuenese in 1323, see Stones, “Another Short Note”, reprinted in these essays, and ead., Gothic Manuscripts, Part I, vol. 2, Cat. III–76; the secondary artist of Add 10292–4 also did the Psalter-Hours of SaintOmer Use, London, BL, Add 36684 and New York, The Pierpont Morgan Library, M.754, written after 1318. See Stones, Gothic Manuscripts, Part I, vol. 2, Cat. Ill–130. 23

212

badly defaced so it is unclear whether a figure (Christ ?) was present in the opening on the left side of the chapel (fig. 2). And the figure of Christ in a cloud giving the book is placed on the following page. But the choice of scenes is the same, despite the greater attention given to architectural detail in Amst. The Merlin opening is identical in Add and Amst, but Add gives a colourful and rare subject for the Suite vulgate, depicting Merlin setting fire to the tents of the rebel barons, the only manuscript to include this subject rather than the more common generic scene of the rebel barons uniting against Arthur (fig. 3).24 Lancelot parts 2, 3 and 5 are identical in Add and Amst, and part 5 is also similar to what is depicted in the Bonn, fr 110 and fr 344 trio, showing Agravain riding to the tent and grieving damsel, while for Lancelot Part 4, Add is once again the only manuscript to include an illustration of the surrender of Margondas (fig 4). The summoning of Lancelot at King Arthur’s table for Queste is common to all copies (assuming it was once in fr 344 and Douce/Rylands). The king commanding the scribe at the opening of Mort Artu is ambiguous, no doubt deliberately so, in Add Rylands and Royal. 3. Ars 3479–80, BnF fr 117–120, 113–116 (Appendix 3) The structure of these copies is complicated by their later re-binding in multiple volumes, not the same as the original structure. I leave these issues aside here and impose on these volumes the branch divisions of the first two groups of manuscripts. It is clear from the start that none of these copies treats the opening of Estoire with subjects that pertain to its story. All three copies (the first volume of BnF fr 112 is not extant) give a set of subjects — substantially the same in Ars and fr 117, both sold by Jacques Raponde and made by the same craftsmen — which treat the story of Lancelot, his birth, eduction in the Lake, his knightly prowess, and his failure to see the Grail. Three of the four subjects at the opening of BnF fr 113 also depict Lancelot, his birth, knighting, and fighting Wild Men, preceded by Walter Map handing his book to King Henry — not Henry as Arthur ordering the writing, as often shown for the opening of Mort Artu (cf. Add, Rylands and Royal) but a finished volume handed over. This makes it more likely the subject is Map and Henry rather than Arthur and his scribe. For Merlin and Suite vulgate see Fabry-Tehranchi, Texte et images des manuscrits du Merlin et de la Suite Vulgate, esp. 148–170. 24

STORIES IN PICTURES AND THEIR TRANSMISSION

213

For Merlin the three copies all give the Harrowing of Hell, two of them, Ars and fr 117 including the devils. There is no pictorial interest in the transition to the Suite vulgate: Ars 3479 and BnF fr 117 both end the Merlin with Arthur’s withdrawing the sword from the stone, to which episode Add 10292 devotes several miniatures — but at the end of the Merlin rather than at the opening of Suite vulgate.25 Lancelot opens with portraits, not just the two kings Ban and Bohort as in Add, but the Queens and sons as well, emphasizing the lineage of Lancelot. Here BnF fr 112 offers a different opening, the birth of Lancelot, similar to the Estoire opening in the other three manuscripts, and pairing it with the escape from Trèbes of King Ban, Queen Elaine and their squire balancing baby Lancelot in his cradle on the neck of his horse. None of the four copies has an illustration for Lanelot Part 2, the Journey to Sorelois, but the Charette is marked in three copies by Lancelot in the cart, not illustrated however in BnF fr 112. The Suite de la charette is marked in all four copies by a scene of Lancelot and Meleagant or Meleagant’s half-sister; Lancelot defeating Meleagant is in BnF fr 112 and in part of the miniature of BnF fr 119, the other half depicting Lancelot rescued by Meleagant’s half-sister, the preferred subject in Ars. BnF fr 115 shows the sequel to the rescue, Lancelot disguised as a woman riding with Meleagant’s half-sister. Lancelot part 5 shows Lancelot riding up to the tent with the bier, with differences as to who is at the bier in Ars, fr 119 and fr 115 (not illustrated in fr 112). For the last two branches, Queste and Mort Artu, there are striking diffences between the Jacques Raponde books and those of Jacques d’Armagnac: the former give a simple Crucifixion with Joseph for Queste, and only a foliate initial A for Mort Artu, while the manuscripts of Jacques d’Armagnac present four-part miniatures for both branches, an approach similar to that of the opening of Estoire (at least in fr 113), serving to round out the cycle as a whole but also to give emphasis to the last two branches. The Crucifixion is also selected for fr 116 and 112 but as part of a four-part miniature with the other three subjects, similar but not identical, depicting the calling of Lancelot to knight Galaad, his journey to the convent and his meeting with Galaad. Mort Artu is similarly emphasized with a composite miniature in fr 116 and with an extra-large miniature of a single scene in fr 112. Overall, then, what is striking is the lack of uniformity, or to put it more positively, the variety of choice of subjects and emphases among the 25

Ead., Texte et images, p. 147.

214

three groups and their component manuscripts. With Jacques Raponde and Jacques d’Armagnac we know a little about the circumstances and the eventual owners. Did Raponde think his royal clients would not appreciate a pictorial emphasis on Arthur‘s death and the destruction of his kingdom? And would they not have appreciated the calling of Galaad to the Grail Quest ? These were precisely the branches emphasised in the illustrations of the d’Armagnac books, suggesting a deliberate preference in favour of Queste and Mort Artu. As to the other copies, we do not know who commissioned them and why they, or their makers, chose to emphasize the subjects they selected. It is important however to look at these choices comparatively: only then can we say what is similar or different and begin to pinpoint what is special in a particular copy or what is commonplace. And eventually an examination based on openings should encompass entire cycles and whole manuscripts, in an approach that is both sychronic and diachronic. We shall probably never know the complete answers but at least through comparisons we can establish what individual choices were made and we can show how they fit, or reject, the patterns present in the rest of the corpus.

STORIES IN PICTURES AND THEIR TRANSMISSION

215

Appendix 1. Text openings in Bonn S 526, BnF fr. 110, BnF fr. 344 Bonn S 526

BnF fr 110

BnF fr 344

f. 1 1. Christ breathes on the hermit-author who lies on the ground. 2. Christ hands the hermit-author a book. 3. Hermit-author led by a composite creature to a religious whom he blesses. 4. Hermit-author writes. 5. Joseph of Arimathea asks Pilate for the body of Christ (still on the cross). 6. Entombment of Christ: Joseph collects Christ’s blood in a bowl.

missing

f. 1 C, Christ hands a book to the hermit-author who lies in bed before an altar.

Merlin

f. 60 Conception of Merlin.

f. 45v Harrowing of Hell.

Suite vulgate du Merlin

f. 73v f. 82 King Arthur receives Kings King Arthur Ban and Boort. unhorses Tradelmant King of Norgales.

f. 101v C, King Arthur enthroned amidst his barons.

Lancelot Part 1 En la marche de Gaulle

f. 171 f. 164 1. King Ban unhorses King 1. Siege of Claudas. Trebes.

f. 184 6 scenes of debate and attack, C, knights riding

Estoire

f. 81v (end of Estoire) Hermit at the fountain with King Lancelot’s head and lions guarding his tomb, with onlookers. f. 81v 1. Harrowing of Hell. 2. Council of Devils. 3. Devil approaches Damsel. 4. Devil destroys family of Damsel. M, Devil and Damsel.

216 2. King Ban and King Claudas debate; King Ban’s seneschal betrays his master.

2. King Ban, Queen Elaine with baby Lancelot and a squire ride away from Trebes. 3. King, Queen and knight debate; Lady of the Lake steps into the lake with Lancelot.

Lancelot Part 2 The Journey to Sorelois

f. 259 Galehot, Lancelot and a squire ride to Sorelois.

f. 254v no illustration

f. 243 C, Lancelot and Galehot drawn in a litter.

Lancelot Part 3 La Charette

f. 296 Lancelot in the cart.

no illustration

f. 321 O, Lancelot in the cart.

Lancelot Part 4 Suite de la Charette

f. 307 Lancelot defeats Meleagant.

f. 321 no illustration f. 333 no illustration

Lancelot Part 5 Agravain

f. 335 Agravain rides away from the damsel and her knight in a tent. 2. A huntsman rides by a knight with a severed head on the ground.

f. 328 Agravain rides away from the damsel and her knight in a tent.

f. 386 C, Agravain rides away from the damsel and her knight in a tent.

Queste

f. 406 1. At Arthur’s table, Damsel summons Lancelot. 2. Lancelot knights Galaad at the convent.

f. 405 1. Damsel rides to Arthur’s table to summon Lancelot. 2. Galaad draws the sword from the stone.

f. 476 A, 1 and 2 torn away. 3. Knight rides to hermitage. Death of King Mordrain in the arms of Galaad.

Mort Artu

f. 443v 1. The Damsel of Escalot offers her sleeve to Lancelot. 2. He wears it at the Tournament of Winchester.

f. 441 Boort returns to Arthur’s court.

f. 517v–518 missing

STORIES IN PICTURES AND THEIR TRANSMISSION

217

Appendix 2. Table of Openings in Add, Amst/Douce/Rylands and Royal 14 E.III

Amst

Amst/Douce/ Rylands

Royal 14 E.III

Estoire

Add 10292 f. 1 Hermit-author saying the night office at an altar with a chalice on it; blessed by Hand of God; Christ in a cloud gives book to priest narrator in bed.

Amst, vol. I f. 3 Hermit-author saying the night office at an altar with a chalice on it; blessed by Hand of God f. 3v Christ in a cloud gives book to priest narrator in bed.

f. 3 col. 1 Hermitauthor saying the night office at an altar with a chalice on it, blessed by Hand of God; col. 3 Christ in a cloud gives book to priest narrator in bed.

Merlin

f. 76 Harrowing of Hell; Devils plotting to overthrow Christ.

Amst, vol. I f. 118 Harrowing of Hell; Devils plotting to overthrow Christ.

absent

Lancelot Part 1 En la marche de Gaulle

Add 10293 f. 1 Ban, King of Benoic, and his brother Bohort, King of Gaunes, talking together.

Amst, vol. II f. 37 absent The Destruction of Trebes: 1. King Claudas and his men, encamped outside, attack King Ban’s castle. 2. King Ban and Queen Elaine escape on horseback from their castle, King Ban looking back in sorrow (wringing his hands), Queen Elaine holding Baby Lancelot.

Lancelot Part 2 The Journey to Sorelois

f. 129 Lancelot lifts up King Galehot, who had fallen swooning from his horse.

f. 199 Lancelot lifts up King Galehot, who had fallen swooning from his horse.

absent

218 Lancelot Part 3 La Charette

f. 180v The Lady of the Lake finds Lancelot, who has gone mad, in the wood.

Amst, vol. III f. 29 The Lady of the Lake finds Lancelot, who has gone mad, in the wood.

absent

Lancelot Part 4 Suite de la Charette

f. 205v Margondas surrenders to Lancelot.

no illustration

absent

Lancelot Part 5 Agravain

f. 251 Agravain approaches a tent where a man and a damsel mourn at a bier.

Douce 215 f. 1 Agravain approaches a tent where a man and a damsel mourn at a bier.

absent

Queste

Add. 10294 f. 1 Banquet at Camelot; damsel summons Lancelot to knight Galaad.

missing

f. 89 Banquet at Camelot; damsel summons Lancelot to knight Galaad.

Mort Artu

f. 53 King Arthur asks Gawain how many knights he has killed in the quest, in the presence of his knights OR Boort relates his adventures to King Arthur, in the presence of his knights.

Rylands Fr. 1 f. 212 A, King Henry II, standing, commands Walter Map to write the Mort Artu; OR King Arthur commanding his scribe to write the adventures of the Queste.

f. 140 King Henry II commands Walter Map to write the Mort Artu OR King Arthur commanding his scribe to write the adventures of the Queste.

STORIES IN PICTURES AND THEIR TRANSMISSION

219

Appendix 3. Text Openings in Ars. 3479–80, BnF fr. 117–1220, BnF fr. 113–116 and BnF fr. 112 Ars 3479–80

BnF fr 117– 120

BnF fr 113– 116

BnF fr 112

Estoire

Ars 3479 p. 1 1. Birth of Lancelot. 2. Lancelot in the Lake. 3. Lancelot wins the tournament of Winchester. 4. Lancelot failing to see the Grail.

BnF fr 117 f. 1 1. Birth of Lancelot. 2. Lancelot in the Lake. 3. Lancelot wins the tournament of Winchester. 4. Lancelot failing to see the Grail.

BnF fr 113 f. 1 vol. I not extant 1. Walter Map presenting his book to King Henry II. 2. Birth of Lancelot. 3. King Arthur knighting Lancelot. 4. Lancelot fighting Wild Men at the Dolorous Gard.

Merlin

p. 109 Devils at the Harrowing of Hell.

f. 50v Devils at the Harrowing of Hell.

f. 117 Harrowing of Hell (without devils).

Not extant

Lancelot Part 1 En la marche de Gaulle

p. 339 Kings Ban and Boort with their Queens and sons.

BnF fr 118 f. 155 Kings Ban and Boort with their Queens and sons.

f. 150v Kings Ban and Boort with their Queens and sons.

BnF fr 112 vol. II f. 1 1. Birth of Lancelot. 2. King Ban, Queen Elaine, and a squire with baby Lancelot in a cradle, ride away from Trebes.

Lancelot Part 3 La Charette

Ars 3480 p. 59 Lancelot in the cart.

BnF fr 119 f. 321v Lancelot in the cart.

BnF fr 115 f. 355 Lancelot in the cart.

not illustrated

220 Lancelot Part 4 Suite de la Charette

Ars 3480 p. 101 Lancelot rescued from prison by Meleagant’s half-sister.

BnF fr 119 f. 333 Lancelot fighting Meleagant; Lancelot rescued by Meleagant’s halfsister.

BnF fr 115 f. 375v Meleagant’s half-sister and Lancelot in disguise ride away from the prison.

BnF fr 112 vol. III f. 3 Lancelot defeats Meleagant.

Lancelot Part 5 Agravain

Ars 3480 p. 180 Agravain rides up to the tent with damsel, squire, and bier.

BnF fr 119 f. 369v Agravain rides up to the tent with two hooded figures and bier.

BnF fr 115 f. 425v Agravain rides up to the tent with clerics and bier; he kills Drias.

not illustrated

Queste

p. 483 Crucifixion with Mary, John, and Joseph at the foot of the cross.

BnF fr 120 f. 520 Crucifixion with Mary, John, and Joseph at the foot of the cross.

BnF fr 116 f. 607 1. Crucifixion with Mary, John, and angels collecting Christ’s blood from the Wounds. 2. The damsel before King Arthur and his court, summoning Lancelot. 3. Lancelot and companions riding. 4. Lancelot greeting Galaad at the convent.

BnF fr 112 vol. IV f. 1 1. Crucifixion with Mary and John, no angels. 2. The damsel before King Arthur and his court, summoning Lancelot. 3. Lancelot greeting Galaad at the convent. 4. Lancelot knighting Galaad at King Arthur’s court.

Mort Artu

p. 591 A, foliate initial.

f. 565 A, foliate initial.

f. 678 1. Boort telling King Arthur and his court about the Queste. 2. Gauvain admitting to having killed 18 questers.

f. 182 Boort telling King Arthur and his court about the Queste.

XI ‘Mise en page’ in the French Lancelot-Grail: the First One Hundred and Fifty Years of the Illustrative Tradition

I

f the number of surviving manuscripts is a reliable guide to what was read in the Middle Ages, the five-part Lancelot-Grail Cycle (or Vulgate Cycle) and its derivatives and followers ranked with Geoffrey of Monmouth’s Historia regum britanniae and the Pseudo-Turpin Chronicle of Charlemagne’s exploits in Spain as one of the most popular vernacular texts.1 Surviving in over a hundred manuscripts, most of which are densely illustrated, it provides a formidable and somewhat unwieldy corpus from which to explore questions about the format and layout of text and illustration in the period between the composition of the texts c. 1220 and the decline of interest in vernacular manuscripts towards the end of the fifteenth century. This is the period that witnesses the origin and development of the illustration of vernacular texts in France, to which the Lancelot-Grail in its various forms, both non-cyclic and cyclic, is a significant contributor. I survey here what patterns of textual and pictorial layout the manuscripts of the first hundred and fifty years of the Lancelot-Grail present. I focus on the appearance of the illustrated page: how its text is laid out and what format is used for the illustrations. The kinds of questions I ask are these: What kinds of variation occur in the layout of text and picture? Do changes in text layout and picture This essay was first given as a paper at the Oxford ‘History of the Book’ Conference in 1992. I thank Linda Brownrigg for her helpful comments. First published in A Companion to the Lancelot-Grail Cycle, ed. C.R. Dover, Woodbridge, UK and Rochester, NY, pp. 125–44. I include an updated the list of manuscripts at the end of the present volumes. 1 The manuscripts of the Lancelot-Grail Cycle are listed in Woledge, Bibilographie and Supplément. For more images of openings, see “Stories in Pictures” in this volume.

222

formats vary together? How important are regional preferences? Is there a chronological progression? How close can we come, in this period of sparse documentation, to determining who made the choices, and on what basis? The attributions I propose for dating and placing these manuscripts are based on a study of the style of their illustrations; it is on this basis, too, that I reconstruct ‘groups’ of related manuscripts (to use a more neutral term than ‘atelier’ or ‘workshop’). Almost all the Lancelot-Grail manuscripts can be shown to have been made by teams of craftsmen whose activities were not limited to the production of romances, but included liturgical and devotional works as well as those made for knowledge or for pleasure. Nor are these books necessarily of inferior artistic quality to their more expensive relatives, although there are instances where qualitative distinctions exist. In most cases I have argued the justification of the stylistic and chronological sequence in more detail elsewhere, and further collaborative work on these questions is in progress. 2 The present investigation is limited to those books, made before 1350, which contain one or more branches of the Cycle and whose pictorial decoration is substantial, consisting of more than a single opening illustration for each branch of the Cycle. There are approximately fifty such manuscripts or sets of volumes. The others, which either have decoration or are unillustrated, I reserve for separate treatment at another time. I also exclude from consideration here the several manuscripts that were made in England and Italy.3 FINDINGS The patterns of layout in thirteenth- and early-fourteenth-century LancelotGrail manuscripts are far from straightforward. There is no chronological progression from one kind of layout to another, nor is one format preferred in any one region. Patterns that match on one dimension, such as the number of columns of text, may differ on another, such as the format of the miniatures or historiated initials. I summarize, under headings, some of the relationships that have emerged. See the collaborative project htp://www.lancelot-project.pitt.edu, discussed elsewhere in these essays. 3 For the fifteenth-century manuscripts, see Blackman, ‘The Manuscripts and Patronage of Jacques d’Armagnac,’ and ead.,‘A Pictorial Synopsis’; and now Hermant, ‘Jacques d’Armagnac’ and Avril and Reynaud, Quand la peinture était dans les livres. 2

MISE EN PAGE IN THE FRENCH LANCELOT-GRAIL

223

Overall format These books are not uniform and it is not clear whether it was the rule or the exception for all five branches of the Cycle to be copied as a complete Lancelot-Grail Cycle. Several more or less complete sets are preserved: Bonn 526 (Amiens and Thérouanne or Cambrai, 1286, fig. 4); BnF fr 110 (same provenance about a decade later); probably BnF fr 95 and Yale 229 (Thérouanne or its region in the 1290s); BnF fr 344 (Lorraine soon after 1300); BL Add 10292–4 (Flanders or Artois, 1316); olim Amsterdam / Rylands / Douce 215 (same approximate date and provenance); BnF fr 105, or more likely BnF fr. 9123, and Ars 3481 (fig. 6) with a lost third volume (Paris, second quarter of the fourteenth century). Only three of the surviving complete sets (Bonn 526, BnF fr 110, BnF fr 344) are still bound as a single volume. Those three, as we shall see, do not have the same text layout, or choice of format for illumination. We do not know whether the others were originally bound in single volumes, for I know of no LancelotGrail manuscript that preserves an original binding. The other manuscripts all include one or more branches of the Cycle and it is unusual for other texts to be included in the same volume: the manuscripts that do so are Modena E 39 (second quarter of the thirteenth century, of uncertain provenance); BnF fr 95; Berkeley, UCB 106 (c. 1250, Paris?); UCB 107 (c. 1250, Paris?); BnF fr 770 (Douai, c. 1285); BL Add 5474 (same artists as BnF fr 110); Cologny-Geneva, Bodmer 147 (c. 1300, of uncertain provenance). Of these, Modena E 39, Berkeley UCB 106 and 107, BnF fr 95, and BnF fr 770 are books in which the additional texts occur in separate quires that, although contemporary and in some cases (BnF fr 95, BnF fr 770) certainly by the same scribes and painters, need not originally have interfered with the textual structure of a set of Lancelot-Grail volumes. Perhaps they were purchased in the same lot and stayed together. They might have been originally bound together too, for Additional 5474 and Cologny-Geneva, Bodmer 147 show that other material could even be interpolated into or excised from LancelotGrail text. Whether these last two books were originally part of complete Lancelot-Grail cycles, we do not know. The few mentions of Lancelot-Grail manuscripts in the contemporary inventories suggest that an owner might have a copy of one of the branches without the others: at the death in 1304 of Jean d’Avesnes, count of Hainaut, the item at the top of his inventory was ‘uns grans roumans a

224

rouges couvertures, ki parolle de Nasciien de Mellin et de Lancelot dou Lach.’4 Extant manuscripts containing these branches and made before 1304 are: Rennes 255 (fig. 1); UCB 106; Tours 951 (assuming the presence of the Lancelot announced on the last folio but now lost); BnF fr 748/754 (assuming they were originally bound as one volume). The inventory made at the death of Robert de Béthune, count of Flanders from 1305 to1322, listed ‘un livre de Merlin,’ as did the 1303 will of Jean Cole, bourgeois of Tournai.5 None of the extant copies has Merlin alone, but if the documented volumes actually included Estoire or Joseph as well, then BL Add 38117 might have belonged to Robert of Béthune. It is probably a little too late in date to be considered as Jean Cole’s book. In the manuscripts themselves, however, there is no positive evidence of ownership. The list of ten ‘roumans qui sont monseignieur’ — written by a thirteenth- or early-fourteenth-century hand on the back flyleaf (fol. 269) of BnF fr 12569 (Chevalier au Cygne and Chanson d’Antioche) — includes a ‘Lancelot du Lac.’ We do not know who ‘monseigneur’ was, or whether one of the extant copies of Lancelot is meant, but the wording makes it clear that someone on the staff of the household was in charge of the books and kept records.6 Text columns It is known from numerous manuscripts whose pictures were never finished that in the process of making the book, the copying of the text preceded the drawing and painting. Spaces for pictures had to be anticipated at appropriate places on the written page and left blank by the scribe. Considerable attention has focused recently on the various intermediaries that were used in this process: notes, sketches, lists, and combinations of all three.7 Just how this worked is still a question on which there is much research to be done, Dehaisnes, Documents et extraits, I,156. I have commented upon this reference in Stones, ‘Beinecke 229 and fr 95, n. 144. 5 For Robert de Béthune see Stones, ‘Yale 229 and fr. 95,’ p. 240; for Jean Cole see de la Grange, ‘Choix de testaments tournaisiens,’ 38 and Stones, ‘Yale 229 and fr. 95,’ nn. 143 and 147. 6 The manuscript is described in Duparc-Quioc, La Chanson d’Antioche, and ascribed to Arras c.1300. See also Stones, ‘Yale 229 and fr. 95,’ n. 144. The book-list was edited by G. Paris in Romania, 17, 1888. For the manuscript see Stones, Gothic Manuscripts, Part I, vol. 2, Cat. III–8. 7 See Alexander, Medieval Illuminators. 4

MISE EN PAGE IN THE FRENCH LANCELOT-GRAIL

225

including the placement of the minor initials in the unillustrated books. Two aspects of the problem of direct concern here are: the number of columns in which the text is written and whether this correlates with the format selected for the illumination. Another dimension of the question of text/picture relationships is that of the rubrics, which I touch on briefly below. Of the fifty or so manuscripts under consideration here, thirteen are written in three columns, one manuscript alone in one column, and the rest in two columns. There are no manuscripts in four columns.8 Those in three columns are: Rennes 255 (Paris, c. 1220, fig. 1); BnF fr 770 (Douai, c. 1285); Bonn 526 (Amiens and Thérouanne or Cambrai, 1286 [fig. 4]); BnF fr 110 (Thérouanne or Cambrai, c. 1295); New York, M.805–6 (Amiens or Laon, c. 1312); BL Royal 14 E.III (Flanders or Artois, c. 1315); BL Add 10292–4 (Flanders or Artois, 1316); and the Parisian-made books of the 1320s–40s: BnF fr 105, BnF fr 9123, Ars 3481 [fig. 6], BnF fr 333, Ars 3482, and BnF fr 16999. Parisian products account for six of the thirteen copies; they represent a wide chronological range, from the earliest, Rennes 255, c. 1220 [fig. 1],9 to the cluster of five books made at the end of our period, between c. 1320 and 1340, for and perhaps by, the libraires Thomas de Maubeuge, Geoffroy de Saint-Ligier, and Richard de Montbaston.10 Generally, the three-column text layout is common in books made by this team [fig. 6] in the second quarter of the fourteenth century, for the styles in the books they made show clear evidence of collaboration, yet it is not the only layout used for Parisian Lancelots of this period: Oxford, Douce 199 (unique among Lancelot-Grail manuscripts) is written in one column, and Bodl. Rawlinson Q.b.6 and St.Petersburg Fr.F.v.XV.5, are in two columns. Parisian Lancelot-Grail manuscripts of the mid-thirteenth century do not continue the three-column layout, the preference is for two columns, a format which can already be found perhaps in the second quarter of This layout is rarely used for texts in prose, but exceptions include parts of the verse and prose compilation from Saint-Omer, Ars 3516, which can be dated by its computus table to 1268 and localized by its calendar. BnF fr 1553, a comparable miscellany of verse and prose texts composed soon after 1284, is smaller in format and has all its texts in two columns. 9 For the date and provenance, see Stones, ‘The Earliest.’ 10 Their roles in the Paris book trade have been recently addressed in several of the articles in Fauvel Studies. See now Rouse and Rouse, Manuscripts and their Makers. I am sceptical about attributions to the Montbaston husband-and-wife-team as illuminators because there is not a single manuscript that is signed by them in that capacity. 8

226

the century (?) in Modena E 39 (of uncertain provenance). The sparsely illustrated manuscript BnF fr 768, the base manuscript of Kennedy’s edition,11 which contains a single historiated initial, is another early example of a Lancelot manuscript written in two columns, and there are unillustrated early examples. The other six three-column Lancelot-Grail manuscripts include two pairs associated with the diocese of Thérouanne: Bonn 526 [fig. 4] and BnF fr 110 (c. 1285–95), and a second pair, BL Royal 14 E.III and Add 10292–4 (made a generation later, c. 1315–25). But both groups also include a third LancelotGrail manuscript, or part of one, written in two columns: olim Amsterdam / Rylands / Douce in the latter case and BL Add 5474 in the former. The two-column text layout was also the rule in the earlier Thérouanne products, BnF fr 748/754, and the pair of Estoire / Merlin manuscripts, BnF fr 19162 and BnF fr 24394. Of the remaining books with a three-column text layout, BnF fr 770 was made in Douai c. 1275 and associated with three two-column Lancelot-Grail manuscripts: BnF fr 342, written in 1274; Le Mans 354 [fig. 3], and Bodl. Digby 223.12 No other Lancelot-Grail manuscripts can be connected at the moment with New York, M.805–6 (made in Amiens or Laon c. 1315), nor do any of its associates have a three-column layout.13 Overall, the three-column books tend to be larger — approximately 400x300mm — than those written in two columns and the one written in one column, but there are exceptions. Several two-column volumes are comparable in size: BnF fr 95/Yale 229 is 470 x 330/40mm, olim Amsterdam / Rylands / Douce is 405/411 x 290/292mm, Oxford, Bodl. Rawl.Q.b.6 is 412 x 262mm; BnF fr 122 is 408 x 305mm, and some of the three-column volumes are smaller: BnF fr 770 is only 318 x 232mm; New York, M.805–6 (admittedly heavily cropped) is 346 x 254mm and BNF fr 333 is 380 x 280mm.

Cited in n. 1. The artistic context of the manuscript is discussed by Stirnemann, ‘Some Champenois Vernacular Manuscripts,’ p. 207 and fig. 24. 12 The localization is based on Valenciennes, BM 838, a martyrology of Notre-Dame des Prés, (O. Cist.), Douai, and the psalter-hours Brussels, BR 9391, of the use of Saint-Amé (O.S.A.), Douai. Le Mans 354 was written by Walterus de Kayo who, as Terry Nixon has noted, wrote another manuscript in 1285, see Stones, ‘The Illustrated Chrétien Manuscripts,’ pp. 252–3. 13 For the related material, see Stones, ‘Vie de sainte Benoîte.’ 11

MISE EN PAGE IN THE FRENCH LANCELOT-GRAIL

227

Format of the illustrations The range of format used to enclose the narrative scenes covers historiated initials, miniatures in one, two, or three text-columns, and multiple images consisting of a single large miniature across several columns or single miniatures juxtaposed on the same page. Most manuscripts include borders with figures, hybrids, or animals, particularly on opening pages. Superficially, they seem rarely to fulfill a function intrinsic to the narrative, and appear to reflect designs invented by workshops or individuals for other contexts. Only in the Ars 5218 Queste manuscript, written, illuminated, and bound by Pierart dou Thielt in 1351, do the marginal figures support the kind of parallel action that can be said to present a coherently understandable commentary on the main illumination.14 Further work on the other manuscripts may yield different results. Opening illustrations Most manuscripts adopt one format for the main body of the illumination and another for the opening pages of the five major branches, as Emmanuèle Baumgartner has observed for Tristan manuscripts,15 and Lori Walters for those of Chrétien de Troyes.16 Although the Suite Vulgate and subdivisions of Lancelot are not usually singled out for special openings, the opening of Joseph is given special treatment when it follows an illuminated Estoire. What that special treatment consists of depends partly on what is used for the illustrations throughout: so if historiated initials are the rule, the opening one is bigger, with a more emphatic border (Rennes 255; BnF fr 748/754), or a miniature (BR 9627–8; Tours 951; Bodl. Rawl.Q.b.6); BL Add 38117; BnF fr 1422–4); or a cluster of miniatures (BnF fr 344, Yale 227). If miniatures are the rule, then the opening illustration may be a historiated initial (BnF fr. 95/Yale 229; BnF fr. 122 for Mort Artu). More usual, however, is a cluster of miniatures or a large subdivided miniature to mark the opening (BnF fr 342; Le Mans 354; BnF fr 770; BnF fr 19162; Bonn 14 Analysis and references are in Walters, ‘Wonders and Illuminations.’ See now Moore Hunt, Illuminating the Borders, for analysis of the motifs in BnF fr 95. 15 Baumgartner, ‘La “première page”’. 16 Walters, ‘Paris, BN, fr 1433’ and ead.,‘Multi-Compartment Opening Miniatures’.

228

526 [fig. 4]; BnF fr 110; Cologny-Geneva, Bodmer 147; BnF fr 749; New York, M.805–6; olim Amsterdam/Rylands/Douce; BL Royal 14 E.III; Add 10292–4; St. Petersburg Fr.F.v.XV.5; BnF fr 105; BnF fr 9123; Ars 3481 [fig. 6]; Douce 199; Ars 3482). This is the largest sub-group, comprising eighteen manuscripts. Some simply include a larger miniature at the opening: BnF fr. 333; or a large miniature at the opening and others sporadically in the text: UCB 107; Oxford, Bodl. Ash. 828; BL Royal 20 D.IV; Ars 5218. The opening pages are given no special treatment in Modena E 39; BnF fr. 339, BnF fr. 24394; BL Add 5474; BnF fr 12573; BnF fr 12582. What precisely were the criteria that guided the selection of subjects to form part of the opening miniature or initial? This is an area that needs more work. I have shown elsewhere that the selection of the subject in Mort Artu manuscripts can have important implications for how the ending of the story was interpreted, and the formal treatment of the subject selected can also play a key role in determining a subject’s effect.17 At other times the format, and even the selection of the subjects, may be more a matter of workshop practice, particularly in the environment of mass-production that characterizes Parisian output in the second quarter of the fourteenth century. Historiated initials Historiated initials are the standard format for illustrations at the beginning of the Lancelot-Grail illustrative tradition (Rennes 255 [fig.1], Modena E 39, BnF fr 339, BR 9627–8), as also in the earliest illustrated verse romances in Latin, for example the Virgil, BnF lat 7936 of c. 1200,18 or in French, the Roman de Troie, Ars 3340, written in 1237.19 At the end of our period historiated initials are still current, if no longer exclusive (BnF fr 1422–4, c. 1320–40),20 but their use no longer correlates with the three-column text format: if Rennes 255 has three columns [fig. 1], Modena E 39 and Stones, ‘Aspects of Arthur’s Death’, reprinted in these essays. Avril, ‘Un manuscrit d’auteurs classiques.’ 19 Reproduced in Samaran and Marichal, Manuscsrits datés, I, 159, pl. 12, but is still neglected in the literature because it was omitted in Buchthal, Historia Troiana. 20 Its illuminator also collaborated with Pierart dou Thielt in part of the French verse Alexander, Oxford, Bodl., Bod. 264 (facsimile by James, The Romance of Alexander) and he can be traced in several other mss, including a Pontifical of Cambrai in Wrocław. For a summary, with previous literature, see Stones, ‘Fauvel, with a Note on Fauvain.’ Another unnoticed ms that can be added to the oeuvre of Pierart dou Thielt is the psalter Chartres, 17 18

MISE EN PAGE IN THE FRENCH LANCELOT-GRAIL

229

Paris, BnF fr 1422–4 have two. The use of historiated initials does however appear to be related to geographical distribution, for they are the rule rather than the exception in Parisian products of the thirteenth century (Rennes 255 [fig. 1], BnF fr 339, BR 9627–8, UCB 106, Tours 951, Oxford Bodl. Rawl.Q.b.6), no doubt under the impetus of the small, illuminated academic study-Bibles that dominated the Paris book trade.21 But other Parisian-made books, such as the law books in Latin,22 and William of Tyre’s Outremer in French, prefer single-column miniatures.23 Perhaps UCB 107, a Parisian Lancelot with miniatures, was made under their influence. By c. 1310 (St Petersburg Fr.F.v.XV.5) the extensively-illustrated Lancelot-Grail had also come to be produced in Paris with single-column miniatures.24 With the mass-production enterprises associated with the Maubeuge / Saint-Ligier / Montbaston libraires in the 1320s–40s,25 the change to miniatures in one or two columns had become decisive; it would become the dominant format for the fifteenth century. Yet the thirteenth-century Parisian preference for historiated initials was not limited to Paris. They are also the exclusive or primary format for the illustration of manuscripts made in Flanders, Artois, and Lorraine: BnF fr 748/754 is probably from Thérouanne,26 BnF fr 344 and its twin, Sotheby’s 1.7.46 (ex-Phillipps 1047), are from Lorraine, probably Metz (the former including some miniatures as well [fig. 5]);27 and BL Additional 38117 (Huth Merlin) is probably from Arras,28 and BnF fr 1422–4 from Tournai. It is worth noting that the Tournai manuscript can be considered a lesser product from an artistic environment where the major products had full-page miniatures.29 BM 549, destroyed in WWII, a photograph of which is in the Fonds Porcher at the BnF. See now Bibliographie des manuscrits de la Bibliothèque municipale de Chartres (Médiathèque l’Apostrophe) par l’IRHT, sub numero. 21 Numerous examples are reproduced in Branner, Manuscript Painting. 22 See Melnikas, Decretum Gratiani. 23 See Folda, Saint-Jean d’Acre. 24 Laborde, Manuscrits à peintures conservés dans l’ancienne Bibliothèque impériale publique de Saint-Pétersbourg, I, 7, no. 6, and I, 31, no. 27. 25 See Stones, ‘The Artistic Content.’ 26 My attribution, based on similarities with Ars 3516, the literary miscellany having a calendar of Saint-Omer (diocese of Thérouanne). 27 BNF fr 344 was mentioned by Vitzthum, Pariser Miniaturmalerei, p. 123. 28 Localization based on linguistic analysis by Dees, Atlas des formes linguistiques, p. 522: ‘Pas-de-Calais, sud-est.’ 29 The major example is the Roman d’Alexandre, Oxford, Bodl. Bod. 264.

230

Miniatures The most visually impressive Lancelot-Grail manuscripts are the ones illustrated predominantly with miniatures. Even one-column miniatures are bigger on average than historiated initials, which normally occupy less than the width of a text column. The miniature format allows more picture space, and the multiple-column miniature increases it. No doubt the cost factor rose proportionately as well, a question I have discussed elsewhere.30 On the whole, miniatures, rather than historiated initials, are the preferred format for Lancelot-Grail manuscripts made in the northern provinces.31 Again, there are several exceptions — the occasional Parisian book with miniatures (UCB 107 and the fourteenth-century ones), books with mainly historiated initials but with miniatures for the openings of branches, as noted above. This might suggest that, in a scale of relative hierarchies, the miniature was more highly regarded, but evidence to the contrary comes from the books (cited above) that are predominantly illustrated with miniatures but open with a historiated initial. Further evidence comes from manuscripts in which substantial numbers of both miniatures and historiated initials are scattered throughout the book: Le Mans 354 (4 miniatures [fig. 3], 11 historiated initials); BnF fr 770 (102 miniatures, 52 historiated initials), Yale 229 (77 large miniatures, 51 small miniatures, 36 historiated initials — but note the different proportions in BnF fr 95, which has 99 miniatures, 28 small miniatures, 9 historiated initials), New York, M.805 (38 miniatures, 157 historiated initials). Then there are the instances where the illumination is predominantly miniatures, but the occasional historiated initial appears: Bonn 526; Cologny-Geneva, Bodmer 147; BL Add 10292–4; olim Amsterdam / Rylands / Douce. Most striking among the manuscripts with miniatures are the multiplecolumn pictures that spread horizontally across the page, allowing sequential episodes to be represented or more supporting landscape or architectural elements to be included than in the single-column miniature or historiated Stones, ‘Yale 229 and fr 95,’ pp. 80–83. The possibility of influences coming to this region from the Empire might be a factor in the preference for miniatures over historiated initials: the Heisterbach Bible, for instance, attributed by Swarzenski to ‘Cologne?’ in the mid-thirteenth century, presents some striking formal and iconographical parallels with Bonn 526. See Swarzenski, Rhein, Main und Donau, pp. 91–5, no. 10. 30 31

MISE EN PAGE IN THE FRENCH LANCELOT-GRAIL

1. Rennes, BM 255, Estoire del saint Graal, fol. 32, Tout-en-Tout and King Mordrain (© Rennes, BM).

2. Brussels, BR 9627–8, Mort Artu, fol. 139v, Mordret and his army lay siege to the Tower of London (© Brussels, BR)

231

232

3. Le Mans, MM 354, Estoire del saint Graal, fol. 49v, King Tholomer routed by Evalach; the White Knight intervenes in support of Evalach (© Le Mans, MM)

MISE EN PAGE IN THE FRENCH LANCELOT-GRAIL

4. Bonn, Landes-und-Universitätsbibliothek 526, Lancelot-Grail, fol. 1, Opening of L’Estoire del saint Graal, the Hermit-Author and Joseph of Arimathea (© Bonn, LUB).

233

234

5. Paris, BnF fr. 344, Lancelot-Grail, fol. 184, End of Suite vulgate and beginning of Lancelot (© BnF).

MISE EN PAGE IN THE FRENCH LANCELOT-GRAIL

6. Paris, Bibliothèque de l’Arsenal 3481, Lancelot, fol. 3, Destruction of Trèbes, Death of King Ban, Abduction of Lancelot, Queen Elaine received by nuns (© BnF)

235

236

initial. To what extent do the subjects change when these long miniatures are used instead of the compressed space of the historiated initial or singlecolumn miniature? This is another area where much is still to be done. Of the books written in three columns, only New York, M.805–6 exploits the full width of the page, with three-column miniatures scattered throughout the book as well as on opening pages. In contrast, BnF fr 770, although written in three columns, limits the width of its large miniatures to two columns — was this because it is related to BnF fr 342, which has text and miniatures in two columns? Yet a third Lancelot-Grail manuscript, Le Mans 354, which is related to both, has a two-column text with only singlecolumn miniatures or historiated initials [fig. 3]. The rectangular miniature spanning the full width of the written space seems not to have occurred in Lancelot-Grail manuscripts produced in Paris. While the possibility of a lost example should not be discounted because there are numerous parallels in other texts, the multiple-column miniatures are generally found on opening pages rather than in the body of the illustration.32 By c. 1320–40 the threecolumn books BnF fr. 105, Ars 3481, and BnF fr 333 have miniatures in two columns (apart from openings [fig. 6]), while BnF fr 9123 and Ars 3482 are also written in three columns but have miniatures in a single column (apart from openings). The lack of uniformity is all the more striking because the stylistic links between BnF fr 9123 (main hand), BnF fr 105 and Ars 3481 are so close that they must be by the same painter (Geoffroi de Saint-Ligier?).33 Both BnF fr 342 and New York, M.805–6 are from the north, if not from the same place (most likely Douai and Amiens or Laon, respectively). The three other books with two-column miniatures across the width of the page are BL Roy. 20 D.IV, BnF fr 122 (written in 1345ns), and Ars 5218 (written, illuminated and bound by Pierart dou Thielt in 1351). There is no doubt that Ars 5218 was made at Tournai, and the other two books were probably produced there too: both Roy.20 D.IV and BnF fr Several examples are reproduced by Avril in the exhibition catalogue Les Rois maudits. See Stones, ‘Fauvel.’ There is a numbered rubric-table at the beginning of BnF fr. 9123 which omits one of the miniatures, and the miniatures in the manuscript are numbered. The number next to the miniature that is missed in the rubric-table repeats the previous figure: so the rubric-list and the numbers followed the miniatures and rubrics present in the text, not vice-versa. The rubric-list and the rubrics themselves repeat verbatim. The presence of a similar list in Ars 3481 but not in BnF fr. 105, another copy of Estoire, suggests that BNF fr 9123 and Ars 3481 were the first two parts of a three-part set (of which the last has not survived). 32 33

MISE EN PAGE IN THE FRENCH LANCELOT-GRAIL

237

122 have a two-column miniature illustrating the same subject (Lancelot and the enchanted dance) at the same place in the text, and textual lacunae account for the absence of correlation between the others.34 Yet even here it is worth noting the close textual relationship between Roy.20 D.IV and BnF fr 1422–4, probably also from Tournai a decade or so later, and the corresponding absence of links in the format of the illustrations: BnF fr 1422–4, as mentioned above, has historiated initials. The absence of full-page miniatures in Lancelot-Grail manuscripts is probably the most surprising gap in this complex web of possible choices that producers, patrons, and illustrators faced. Even in the manuscripts with multiple miniatures, whether these occur in isolation or clustered in their columns on opening pages, the pictures always have lines of text beneath them, and usually a good half-page of it. In the manuscripts under consideration here, there is no extant parallel (and perhaps none ever existed) for the fullpage scenes found in the Roman de Troie, BnF fr 1610 or the Chevalier au Cygne and Chanson d’Antioche manuscripts BnF fr 12558 and BnF fr 12569. Crusading tales, with their obvious parallels in the iconography of the Bible (the Old Testament Picture Bible, New York M.638, is the most likely source, particularly for BnF fr 12558), are not alone in preserving full-page miniatures. If the Roman de la Poire, BnF fr 2186, is something of a special case whose circumstances are still not altogether clear,35 there are also the groups of full-page miniatures in the French prose Alexanders of c.1300,36 and the Yvain / Atre périlleux in BnF fr 1433, perhaps made in Tournai c. 1320–30.37 There are also numerous examples of the painters of Lancelot-Grail manuscripts working on another book that does have full-page pictures: the best example is New York, M.805–6. Its main artist also painted the Vie de sainte Benoîte, Berlin, Staatliche Museen Preussische Kulturbesitz, Kupferstichkabinett 78.B.16; two detached leaves from a Missal in the Vienna Kunstgewerbemuseum; and the medical treatise BL Sloane 1977, all with full-page miniatures that are highly idiosyncratic in treatment.38 See Stones, ‘Yale 229 and fr. 95.’ Its illustrations are fully reproduced in Marchello-Nizia, ed., Le Roman de la Poire. For a date c. 1270–80 for BnF fr. 2186, see Keller, ‘La structure du Roman de la Poire’. 36 Stones, ‘Three Illustrated Alexander Manuscripts,’ and for a fourth Alexander belonging to the same group, see ead. and Ross, ‘The Roman d’Alexandre, another illustrated manuscript’. 37 See Walters, ‘Super Romance.’ 38 See Stones, ‘Vie de sainte Benoîte.’ 34 35

238

That painter was clearly fully capable of handling the full-page format if required. Finally, there is the question of the density of the illumination, whatever the format. Textual lacunae account for some of the differences in numbers of illustrations, and full tabulation of the distribution of illustrations in each branch would be required to determine exactly where the differences occur. The numbers listed in the Appendix (at the end of these volumes) show the overall totals of illustrations for each manuscript. Among those that preserve the complete Lancelot-Grail Cycle, for example, there are densities of illumination that range from 325 pictures in BnF fr 95 / Yale 229 (lacking the first two parts of Lancelot, which might add another 100–120 images) to 344 in BnF fr 344 (though Queste and Mort Artu are incomplete), to 349 in Bonn 526, to 748 in BL Add 10292–4. Was the sheer number of pictures a factor in determining the format? One could argue that, for speed of production, the single-column miniature can be made most quickly as there is no need to waste time deciding which initial letter is required (though guide letters usually indicate it), whereas multiple-column miniatures would clearly take longer to paint. Bonn 526 and Add 10292–4 do indeed have single-column miniatures. If one conclusion emerges from this survey, it is that the choices were not altogether governed by the easily-detected criteria of type of text, date, or regional preference. If inherited tradition and workshop practice were to some extent significant factors, the results were neither predictable nor uniform. More comparative work on the iconography of these lengthy cycles would help clarify what the dependency patterns are and establish more clearly which aspects of the texts were considered most important through pictorial emphasis. Overall, the matrix of determinants is a highly complex one that must include substantial unknowns and missing links.

XII Short Note on Manuscripts Rylands French 1 and Douce 215*

M

anuscript French 1 in the John Rylands Library, Manchester,1 is an early fourteenth century illuminated copy of the second part of the prose Vulgate cycle of old French romances, containing the texts of Agravain (f. 1–181v), Queste (f. 182–211v), and Mort Artu (f. 212–257v). The first part of the complete cycle, containing the Estoire del Graal, Merlin, and the first part of Lancelot, is lost so far as is known, and there are several gaps in the remaining parts of the manuscript.2 Apart from the removal of several leaves whose miniatures have been cut out,3 there are the following lacunae in the Rylands manuscript : 1. At the beginning, the text of Agravain is incomplete.4 2. The opening of Queste lacks the first leaf.5 3. A section of about 100 pages is missing between f. 204 and 205.6 4. After f. 205 there is a gap of two leaves.7

* The research for this article was facilitated by a grant from the Central Research Fund of the University of London. I should like to thank the Librarian of the John Rylands Library, Manchester for permission to reproduce ff. 73v, 188 and 226 of Rylands French 1, and the Keeper of Western Manuscripts at the Bodleian Library, Oxford for permission to reproduce ff. 14, 32 and 35 of Douce 215. First published in Scriptorium, 22, 1968, pp. 41–45. 1 Tyson, Hand list of the Collections. 2 These are described in detail by Pickford in “An Arthurian Manuscript”. See now “Another Short Note” where these missing volumes are recovered. 3 Pickford, op. cit., p. 8–9, 15. 4 Pickford, op. cit., p. 8. 5 Pickford, op. cit., p. 11. 6 Pickford, op. cit., p. 12. 7 Pickford, op. cit., p. 13.

240

5. At the end of Queste, the last folio is missing after f. 211v.8 6. The last gathering is missing from the end of Mort Artu, after f. 257v.9 Several of these missing sections are to be found in a manuscript in the Bodleian Library, Oxford, Douce 215,10 which contains seven fragments of one or more leaves from the texts of Agravain, Queste, and Mort Artu, written, like the Rylands texts, in Picard dialect, and containing illumination. A. Micha has established the fact that the first eight leaves of the Douce manuscript form the gathering missing from the beginning of the Rylands manuscript containing the opening pages of Agravain.11 It is clear from an examination of the illumination that the two manuscripts belong together and the other sections of the Oxford manuscript can be shown to fit textually into the gaps left in the Manchester one as do the first eight leaves. Of the six remaining sections of Douce 215, five are from Queste. The first of these consists of thirteen leaves, v. 9–21v of the present foliation. F. 9 begins “jusca eure de none. Mais si tost comme chele eure fu passee enporta li lyons le lionchel a son col et a son repare...” 12 This comes from the passage about the temptations of Perceval, and it continues the sentence which breaks off at the end of f. 204v in the Rylands manuscript, “Tout le iour demoura laiens Percheval...”. 13 The Oxford section ends on f. 21v with the words “avant ier le jour de pentecouste prisent li chestiel et li chevalier terrien.i. tournoiement ensemble chest adire qu’il comenchierent ...” 14 One leaf is missing between this and the third section of Douce 215, f. 22–25v. F. 22 begins “soloient trouver et par che lui annuia plus la queste. Me sire Gauvains chevaucha de la pentecouste dusques ale magdalaine sans aventure trouver...”15 and f. 25v ends “...Sire fait me sires Gauvain, se iour eusse loisir...” 16 Three or four leaves are missing between this and the next four leaves of Douce 215, f. 26–29v. F. 26 begins “Ha chevaliers jou te conjur Pickford, op. cit., p. 13. Pickford, op. cit., p. 15. 10 Madan, Summary Catalogue, vol. IV, no. 21789. Pächt and Alexander, Illuminated Manuscripts, I, no. 571. 11 Micha, Les manuscrits du ‘Lancelot’ en prose, pp. 479–80, 486. 12 Passages are quoted in diplomatic transcript, with abbreviations silently expanded, proper names capitalized and punctuation modernized. 13 This passage corresponds to p. 68, line 30 of the text, as published by Sommer, Vulgate Version, vol. VI. 14 The equivalent of Sommer, op. cit., p. 102, line 35. 15 The equivalent of Sommer, op. cit., p. 105, line 16. 16 The equivalent of Sommer, op. cit., p. 116, line 4. 8 9

SHORT NOTE

1a. Oxford, Bodl. Douce 215, Queste del saint Graal, fol. 35, The Grail knights on Solomon’s ship read the explanatory letter (© Bodleian Library) 1b. Manchester, The John Rylands University Library, French 1, Mort Artu, fol. 226, the boat bringing the body of the Demoiselle d’Escalot arrives at Camelot (© The John Rylands University Library of Manchester)

241

242

2a. Oxford Bodl. Douce 215, Queste del saint Graal, fol. 14, Lancelot visits a dendrite hermit (© Bodleian Library) 2b. Manchester, The John Rylands University Library, French 1, Queste del saint Graal, fol. 188, The White Knight defeats King Baudemagus in mounted combat (©The John Rylands University Library of Manchester)

SHORT NOTE

243

3a. Oxford, Bodl. Douce 215, Queste del saint Graal, fol. 32, Adam and Eve expelled from Paradise (© Bodleian Library)

3b. Manchester, The John Rylands University Library, French 1, Agravain, fol. 73, Lancelot breaks the bars of Morgan’s prison to pick a rose (©The John Rylands University Library of Manchester)

244

par la foi que tu dois a chelui que liges hons tu ies...» 17 and f. 29v ends “... quant li preudons voit Lyonnel ki a Bohort voloit...” 18 Two leaves are again missing between this and f. 30 “...quil le fendi lui et le chevai iusquen treteus fu li primiers caus de eheste espee.” 19 The end of this section on f. 39v fits onto the beginning of Rylands f. 205. F. 39v ends “...car il ne troevent laiens home ne feme ains sont tout mort. Si chierkent amont et aval et dient que molt a chi grant piert de gent. Quand il vindrent au...”and Rylands f. 205 begins “maistre palais ...” 20 All these fragments therefore form a fairly continuous section which fits into the lacuna between f. 204v and f. 205 in the Queste text of Rylands French 1 leaving comparatively few gaps in it. One more folio of Douce 215 belongs to the Queste, this is f. 45, the missing final leaf of the Rylands Queste section. The text of the latter ends on f. 211v with the words “et la merveille du graal et le pooir que Diex i ot mis et chil ert desloiaus et crueus comme chil ki tous ert estrais de malvais lig...” and continues on f. 45 of the Douce manuscript “...nie de paiiens si ne crut riens de quan qu’il disent ains dist quii estoient aucun traiteur mauvais...”21. The last five leaves of Douce 215 are from Mort Artu and form a continuous part of the last gathering that is missing from Rylands 1. The Rylands text ends f. 257v “... il sorent tantost que chestoit aucune haute personne...”.22 The Douce text begins on f. 40, one page later “...ent molt bien et ochistrent grant partie des gens le roi Aguiscant...” 23 and ends 44v ”... si tout contre mont et sentier tant quii est venus au lieu qui asses estoit povre et y avoit une petite chappelle ancienne et il descent alentiee et oste son heaume. Et quant il...” .24 As these last four lines of the Oxford text are in a later hand than the rest of the manuscript it is likely that the text of Mori Artu was originally left incomplete or was lost at an early date. There can be no doubt that all the sections of Douce 215 were once part of the same book as Rylands French 1. Proof based on the interrelation of the text fragments is confirmed by the fact that the manuscripts are identical The equivalent of Sommer, op. cit., p. 126, line 11/12. The equivalent of Sommer, op. cit., p. 136, line 20. 19 The equivalent of Sommer, op. cit., p. 141, line 3/4. 20 The equivalent of Sommer, op. cit., p. 173, line 18. 21 The equivalent of Sommer, op. cit., p. 196, line 14. 22 The equivalent of Sommer, op. cit., p. 368, line 26. 23 The equivalent of Sommer, op. cit., p. 371, line 28/29. 24 The equivalent of Sommer, op. cit., p. 386, line 30. 17 18

SHORT NOTE

245

in format, 25 script,26 decoration and illumination. Both parts of the book contain two kinds of decorative initials, one done in red or blue with pen scroll-work in blue or red, the other larger initials in gold against a coloured background of blue and pink. More interesting is the illumination, which was carried out by three artists, two contemporary, the third somewhat later than the other two. Both Douce 215 and Rylands 1 contain miniatures by all three. The first hand was responsible for most of the illumination of the book as a whole, including the three remaining leaves on which the illumination extends as a border round the margins enclosing the text. The first of these is at the beginning of Agravain (Douce 215, f. 1), the second is Rylands f. 82 which must once have come at the beginning of a volume, 27 the third is Rylands f. 212 at the opening of Mort Artu. The first page of the Queste text, which belongs between Rylands f. 181v and f. 182 and ought also to be illuminated in this way, is still missing. The same hand was also responsible for the miniatures contained in the text columns from the beginning of Agravain to the end of this text,28 in which there are 27 miniatures, 1 in Douce 215, 26 in Rylands 1.29 This artist again did the miniatures from the middle of Queste to the end of Mort Artu, 30 three miniatures from the Douce manuscript, 31 nineteen from the Rylands section. 32 A comparison between two illuminations, one from each part of the book, makes it clear that the same artist did both. (Fig. 1.) The second hand was responsible for the miniatures of the intervening section of the book, from the beginning of the text of Queste (Rylands MS. ff. 182–204v) to the middle of the Oxford section (Douce MS. ff. 9v–21v), a total of 11 illuminations, one in Douce 215, ten in Rylands 1. 33 Again it is 411 x 292 mm, Micha, op. cit., p. 479, 486. 411 x 292 mm, Micha, op. cit., p. 479, 486. 27 Pickford, op. cit., p. 3. 28 Douce 215, ff. 1–8v; Rylands 1, ff. 1–181v. 29 Douce 215, f. 1; Rylands 1, ff. 16v, 24v, 46v, 69, 77, 82, 84v, 90, 94v, 96, 101v, 105v, 109v, 114v, 118, 121, 122, 133, 139v, 140v, 143v, 148v, 158v, 167, 172v, 179. 30 Douce 215, ff. 31v–39v; Rylands 1, ff. 205–257v. The last five leaves of Mort Artu, Douce 215, ff. 40–44v, contain no illumination. 31 On ff. 31v, 35, 39v. 32 On ff. 208, 211v, 212, 216v, 220v, 223, 224v, 226, 228, 230, 231, 240v, 242, 244v, 252, 252v, 253v, 254, 255. 33 Rylands 1, f. 182, 182v, 183v, 184v, 188, 188v, 190, 192v, 195v, 198v. Douce 215, f. 14. 25 26

246

clear that the same artist was responsible for all. (Fig. 2.) As these miniatures interrupt the sequence of those by the first hand yet do not correspond to any logical division of the text, it is likely that they are the work of a contemporary of the first hand who may even been employed in the same workshop. This is an interesting hypothesis as the two hands are so clearly different in style that it is unlikely that the two artists received the same training. 34 The miniatures by the third hand were done after the text and the illumination by the other two hands were completed, as is shown by the fact that they are always found in the bottom margin, never in spaces in the text like the others. On stylistic grounds they belong to the end of the fourteenth century and have been ascribed to the school of Boucicaut. 35 Two miniatures by this hand are in the Douce sections of the book and eighteen in the Rylands section. 36 Again it is clear from a comparison that they are by the same hand. (Fig. 3.) Unfortunately it does not seem possible to trace back the history of the manuscripts to the time before which they became separated. The Rylands manuscript was in the Sunderland Collection where it was No. 670 in the sale catalogue of 1881. The Catalogue entry shows that it was at this date incorrectly bound. 37 It is not known by which member of the Sunderland family it was acquired nor from what source. The Douce manuscript was No. 4006 in the La Vallière sale catalogue of 1783. 38 It is there described39 as containing forty-five leaves and seven miniatures as it does at present, and there is no clue to its previous condition or ownership.

Pickford, op. cit., p. 14 ff., draws attention to the differences between the three hands. Pächt and Alexander, op. cit., p. 44. 36 Douce 215, f. 3v, 32. Rylands 1, f. 13v, 15, 24, 25v, 33, 38, 43v, 48v, 53v, 66v, 72, 73, 191, 193v, 194v, 205, 223v, 239v. 37 Pickford, op. cit., p. 6. 38 Madan, op. cit., p. 558. Pächt and Alexander, op. cit., p. 44. 39 Catalogue de M. le Duc de la Vallière, vol. 2, Paris 1783. 34 35

XIII Another Short Note on Rylands French 11

I

n 1979, H. P. Kraus published as no. 31 in his Catalogue 159 an interesting set of copiously illustrated volumes containing sections of the text of the prose version of Arthurian Romances known as the Prose Lancelot or the Vulgate Cycle.2 The MS now (to 2010) belongs to the Bibliotheca Philosophica Hermetica in Amsterdam, where it is housed without press mark.3 The set was attributed in the Kraus catalogue to the ‘Artois-Picardie school’ and compared with a MS in the Bibliothèque nationale, Paris, that contains some of the same texts, fr 95. It was dated c. 1290, a little later than the date proposed for fr 95. I suggest that it is possible to come a good deal closer in attributing this manuscript: it most probably forms part of the missing sections of the set that now comprises MS French 1 in The John Rylands University Library, Manchester, and MS Douce 215 in the Bodleian Library, Oxford, and it can be much more precisely attributed to a workshop that produced, in the second and third decades of the fourteenth century, two other copies of the same French text as well as illustrated vernacular texts in Flemish, and a number of devotional books for patrons whose homes were in St-Omer, Thérouanne, Tournai, Cambrai, Ghent, and further east in the dioceses of Liège and Utrecht. In 1968, while a student at the Courtauld Institute, I published my first article, in which George Zarnecki took great interest (‘Short Note). It is a pleasure to dedicate this sequel to him in honour of his seventieth birthday (12 September, 1985). Published as ‘Another Short Note on Rylands French 1,’ in Romanesque and Gothic, Essays for George Zarnecki, ed. N. Stratford, Bury St.Edmunds, 1987, pp. 185–192. 2 See also Kraus, Cimelia, pp. 12–15, pl. I thank Mr Kraus for kindly allowing me to study the MS and photograph it in the summer of 1980. I also wish to record my gratitude to the late Neil R. Ker who first drew the Kraus volumes to my attention and who was most generous in helping me to piece together their chequered history. 1

248

The Amsterdam MS consists of three volumes. The first contains the first branch of the cycle, the Estoire del Saint Graal, corresponding to Sommer’s edition vols. I and II.4 It ends on f. 118v with the rubric for the miniature that opens the second branch, Merlin: ‘Chi (co)menche le branke de Mellin’.5 The second volume begins with the opening of Merlin, which occupies the first 37 folios, and continues with the third branch of the cycle, the Lancelot proper, ending on f. 233 with the rubric for the illustration at Sommer, IV 89/5, ‘Chi entra li dus de Clarence en un chastel.’ The third volume follows directly from the second, and ends with rubric for the Agravain section of the Lancelot proper (Sommer IV, 362), ‘Chi (co)menche li livres de la banque Agravain’, indicating that, at one time, there were more volumes to complete the set.6 The text is written in two columns of 44 lines, and decorated with pen-flourished initials and borders, champie initials, and single-column miniatures, accompanied on the opening folios of each branch of the text by foliage borders with human figures, hybrids, and apes. The history of the Amsterdam volumes is difficult to reconstruct. The three volumes were reunited in the collection of Sir Thomas Phillipps, who seems to have recognized that they belonged together, although he purchased them from two different sources. The first volume, which erroneously says Chronique d’Angleterre on the spine, has the printed number 1045 glued to the spine, with the 5 changed in ink to 7.7 I refer to it hereafter as Amsterdam I thank Dr F. A. Janssen, Director of the Bibliotheca Philosophica Hermetica for his kindly interest in my continued work on the MS and for granting permission to reproduce the photographs. The manuscript was given the shelf number ‘1’; it was sold at Sotheby’s on 7.12.2010, as lot 33 and is now in private hands. The present owner has chosen to remain anonymous so it is convenient to refer to it as ‘olim Amsterdam BPH 1’. 4 Sommer,Vulgate Version. See also n. 5. below. 5 For the Merlin see also Robert de Boron, Merlin, ed. Micha. This MS is Micha‘s N, and is closest to BnF fr 113. For the Prose Lancelot see alo id., Lancelot, Roman en prose. 6 They may have all been bound together initially, as are a number of the extant sets of this text, including some which are contemporary with or earlier than this set: BnF fr 344, BnF fr 110, Bonn 526 (written in 1286). For the most complete list see Woledge, Bibliographie and Supplément; some revisions to the dates suggested by Woledge are given in Stones, ‘The Earliest ?’. See also my unpublished Ph.D. thesis, Illustrations of Lancelot. 7 There are some problems here, as Neil Ker’s preliminary work on this MS made clear. The MS is said to have come from the collection of the Ducs de la Rochefoucauld, whose arms are on f. l, and to have been no. 856 in the sale of Charles Yarnold on 6.6.1825; but there was no such number in that Yarnold sale. Nor does the MS correspond to the descriptions of Phillipps MSS 1045, 1046 or 1047, which were sold at Sotheby’s on 1.7.46 as lots 14, 8, and 10 respectively: 1045 (written 1357) is now MS 227 at Yale (see Shailor, Catalogue 3

ANOTHER SHORT NOTE

249

Phillipps 1045/7. The other two volumes contain Phillipps numbers 3630 i and 3630 ii, and came from the Robert Lang sale (17.11.1828, no. 1305). It is most likely that Phillipps was correct about the three volumes belonging together. Not only do they contain clearly complimentary parts of the same text, but they are closely similar in dimensions, script, decoration, and illumination (figs. 1 and 2).8 Not that they are identical in these respects: different scribes seem to have written each one of the three volumes, although they are clearly contemporary; and there are two kinds of champie initials in the first volume (1045/7), distributed by gathering, and another two in the second volume (3630 i), and none at all in the third volume (3630 ii). The small rectangular miniatures, which occur in all the volumes, have silver bands around the frames in 1045/7 (or are they tin?), and gold in 3630 i and ii. There also seems to be a change of hand in the illumination in the second half of 3630 i, although the general stylistic features remain the same. Of enormous interest is the fact that all three volumes have erased notes on of Medieval and Renaissance Manuscripts at Yale, I, pp. 318–20); 1046 is Bodmer MS 147 in Cologny-Geneva (see Vieillard, Manuscrits de la Fondation M. Bodmer, p. 46); 1047 was in the possession of Scheler in 1954 and in 1980 in the hands of an anonymous collector (Micha, op. cit., xix) but it is definitely not this MS; for one thing, the texts it contains also include the Merlin and the Suite Vulgate as indicated by Micha (‘Les manuscrits du Merlin en prose,’p. 93). In addition, there is a photo of the opening miniature of Phillipps 1047 on pi. XV of the Sotheby catalogue of 1.7.46, which is certain proof that it is not the Amsterdam MS. The historiated initial shown has distinctive foliage types and knots in the stems that make up the borders, and a cusped spiral termination with gold balls and bell-shaped flowers (convolvulus?); it is clearly by the same hand as the illustrations in BnF fr 344 (see Stones, Illustrations of Lancelot, p. 149), to which there is also a very close similarity in the script type. This painter is one of the lesser artists in the large workshop operating in Metz in the early years of the fourteenth century, and whose major patron was Renaud de Bar, bishop of Metz (1302–16). The most likely explanation for the number change on the Amsterdam volume is that Phillipps reassigned this manuscript’s number after buying the other two volumes and realizing their affinity; but this does not explain why 1045 was changed to 1047 in ink and not altered again. There is a note inside the front cover of 1045/7 in Phillipps’ hand, saying, ‘Bought I believe in Yarnold’s sale or of Longman and Co. It is the first volume of the body of this work of which I bought the other two volumes in Lang’s sale. See 3630 in my catalogue . . .’. See also Munby, The Phillipps Manuscripts, p. 47. For MSS 1045, 1046 and 1047 see id., p.13. 8 Their basic specifications are as follows: Amsterdam Phillipps 1045/7: 118 ff., 3630 i: 233 ff., 3630 ii: 104 ff., each 405 x 290 mm. Douce 215: 45 ff., 411 x2 92 mm; Rylands Fr.l i: 129 ff., ii: 128 ff., each 411 x 292 mm. All have the text in 2 columns of 44 lines. There are 109 illustrations in the Amsterdam volumes, 6 in the Oxford volume and 74 in the Rylands volumes.

250

almost all the illuminated pages, some of which are fairly clear, others of which are almost totally erased. Enough remains to indicate that two sets of notes were generally put in: one for the rubricator, the wording of which is exactly the same as that of the rubric above the miniature, only written in cursive instead of textualis, and the other for the illuminator, with a description of the scene written usually in general terms (a chapel, a bishop), rather than giving the names of the particular characters: figs. 3 and 4).9 Francis Douce was the first to realize that there was a connection between these volumes and the Rylands/Douce volumes (although he knew only the Douce part). On the inside cover of Douce 215 is a note in his hand indicating that he had seen Phillipps 3630 when it was in the Lang collection and bid on it in the Lang sale, but not enough.10 On the basis of the reproductions in Kraus Cat. 153, the late Neil Ker thought that the scribe of 3630, f.108 was the same as the main hand of Rylands French 1/ Douce 215.11 The Rylands/Douce MS also has traces of the erased notes, although they are much better preserved in the Amsterdam set. The most distinctive feature about the volumes, and the one that bears clearest witness to their common origin, is the illuminations. The main painter of the Amsterdam volumes is the same as that of the Rylands/Douce main hand (figs. 1 and 2). He adopts a similar arrangement for the openings of the branches of the cycle, with a border around three sides of the page and a single column miniature, and single column miniatures with a geometrical (usually twisted thread) motif on the frame, with an outer band of gold (plates 5 and 6). Not only is the main painter the same in all volumes, but the variations in minor decoration that occur between the volumes of the Kraus set also occur in the Rylands/Douce set, so that one can match the different types of champie initials as well (figs. 7, 8 and 9). If there was indeed a second painter involved in the illustration of the second half of 3630 i, he must have been a very close associate as the differences are slight. A further proof that these volumes all belonged together originally is the presence on f. 43 of 3630 ii of a miniature by the later hand that also inserted some miniatures in the borders of the Rylands/Douce volumes (figs. 7, 8 9 Most have been erased to the point that it is possible to read only enough to realize that one set is worded exactly as the captions above the miniature, while the other is a general description of the scene. See my forthcoming article, ‘Indications écrites et modèles picturaux’, reprinted in these essays. 10 I owe this information to the kindness of the late Neil R. Ker. 11 Reproduced in Kraus Catalogue 153, no. 31 as the lower of the two illustrations.

ANOTHER SHORT NOTE

251

and 10). The scene in 3630 ii is the famous episode of Lancelot crossing the sword bridge, clearly felt by the fourteenth-century owner to have been a very serious omission! All that is lacking in the Amsterdam volumes is the participation of the second hand of the Rylands/Douce set, presumably a near contemporary and close associate of the main hand, who did a few miniatures in the middle of the work of the main hand, and who uses a different set of frames around his miniatures, and draws figures on a slightly larger scale than the first hand, and with more firmly drawn outlines; but in other respects, such as details of physiognomy, his style is closely related to that of the main hand. His participation in the Rylands/Douce volumes is important for the structure of the workshop as a whole, since he seems to have worked independently on a number of other books which cannot be ignored in a discussion of the structure of the shop, the geographical distribution of its products, and the fundamental question as to whether it was the books, the patrons, or the artists who travelled, or a combination of the same.12 Possibly this second painter appeared on the scene only after the painting in the Amsterdam volumes, which contain the first parts of the text, had been completed. Was he, in fact, a visitor to the shop, or perhaps an assistant who then moved elsewhere and set up shop on his own? The rest of his output and some of its implications are outlined below. First I consider the main hand. As has been widely recognized, the main painter of the Rylands/Douce Prose Lancelot (more usually now called the Lancelot-Grail) made two other illustrated copies of the same text, in addition to the Amsterdam volumes: London, British Library, Add10292–4 and Royal 14 E.III.13 Many features are common to all these MSS: again, the script and minor decoration is all of the same type, the main style of illumination is the same in all, although there are considerable variations of For two illustrations by this hand see Stones, ‘Short Note’, pl. 7, further discussed in my article cited in note 9 above. I use the term ‘workshop’ as a conceptual framework within which to group illuminations in different styles (and therefore presumably by different painters) provided that the different styles occur together, and preferably intermingled, within at least one book. Such a book I see as a pivot within the structure of the workshop. Two books provide pivots here: Amsterdam/Douce/Rylands, and BL Add10292–4, on which see below; but the same argument can also be made in relation to BL Add 28784, where geographical factors may suggest a somewhat different explanation, see below. The case of the Michiel van der Borch books, discussed below, is also problematical in so far as his work has not so far been found in the same book as that of the other painters discussed here. 13 See Vitzthum, Die Pariser Miniaturmalerei, pp. 133 ff., for BL Royal 14 E.III and Add 10292–4, and Loomis, Arthurian Legends, pp. 97–8, 123, for Rylands as well. 12

252

the kind noted within the Amsterdam volumes: whether gold or silver (or tin) is used for the outer frames, whether there is an ink-drawn squiggle at the corners of the miniatures, different motifs on the coloured borders, etc., which may or may not be characteristic of different hands working within the same style. I assume for the sake of simplicity that these are the kinds of variations that would be due to the different phases in an individual’s career, while leaving the question open for future discussion. The iconographical cycles, in so far as they have been studied in detail, are also remarkably close.14 Roy.14 E.III, like Rylands/Douce, is incomplete, but the Amsterdam volumes are not likely to have been part of it, because the first branch of the cycle, Estoire, is present in Roy. 14 E.III, and also because its text, like that of Add 10292–4, is written in three columns and not in two like the Douce/ Rylands volumes. The date of the Amsterdam/Douce/Rylands volumes can be roughly established by comparison with other products of the shop. Two dates are particularly relevant: one, noted by Loomis, is the date of 1316 (1317 ns), ‘carved’ on the tomb slabs of Nabor and le sire de Karabel in Add 10292, f. 55v.15 The other date of interest is 1323, when Gilbert de Sainte-Aldegonde presented his psalter, St-Omer 270, to the Chartreuse of Longuenesse near St-Omer.16 The psalter is an earlier MS, a psalter that has been attributed, if not altogether convincingly, to the use of Tournai, and to which a donor portrait was added together with some prayers and the record, in French, of the donation.17 The portrait is by the main painter of the Lancelots (fig. 11). The Amsterdam/ Douce/Rylands MS is likely to have been made at some 14 For an analysis of Queste and Mort Artu, the fourth and fifth branches of the cycle, see Stones, Illustrations of Lancelot, pp. 307–84. 15 Reproduced in Loomis, Arthurian Legends, fig. 248 and as fig. 1 in my essay The Lancelot-Grail Project in these essays. The inscription says ‘le douzime iour de feuier/l’an de grace.m.ccc.et xvi’ which fell before Easter, date on which the year changed in this period. By modern reckoning the date is 1317 (1317 ns for New Style). The same scene occurs in Royal 14 E.III but without a date on the tombs. In Amsterdam Phillipps 1045/7 this episode is not illustrated. 16 See Leroquais, Psautiers manuscrits, II, 203, pl. 17 Leroquais attributes the use to Tournai on the basis of the prominence of Piat in the litany, although Piat is less prominent than, for instance, in Tournai Bibl. de la Ville 31bis, and Eleutherius is absent, which in my view makes the attribution not altogether convincing; and the calendar has been adapted for Longuenesse and is not useful in determining the origin of the book. Stylistically the early part is characteristic of illumination in Bruges and Ghent, see Carlvant, Thirteenth-Century Illumination in Bruges and Ghent, passim. But the liturgical indications are closer to Tournai than to Bruges or Ghent.

ANOTHER SHORT NOTE

1. Manchester, The John Rylands University Library, French 1, Agravain, fol. 82, Guinevere’s messenger to the Lady of the Lake talks to King Claudas who has intercepted her (©The John Rylands University Library of Manchester)

253

254

2. olim Amsterdam, BPH 1ii (ex-Phillipps 3630 i), Merlin, fol. 1 Harrowing of Hell and the Devils plotting (© Lancelot-Grail Project)

ANOTHER SHORT NOTE

3. olim Amsterdam, BPH 1i (ex-Phillipps 1045/7), L’Estoire del saint Graal, fol. 114v, Bishop Josephé on his deathbed hands the Grail to Alain le Gros (© Lancelot-Grail Project) 4. olim Amsterdam, BPH 1i (ex-Phillipps 1045/7), L’Estoire del saint Graal, fol. 114v, Pen-flourished decoration, marginal notes in leadpoint, catchword in ink (© Lancelot-Grail Project)

255

256

5. olim Amsterdam, BPH 1i (ex-Phillipps 1045/7), L’Estoire del saint Graal, fol. 89v, The Christian sinners enter the boat and sail to England (© Lancelot-Grail Project) 6. Oxford, Bodl. Douce 215, Queste del saint Graal, fol. 35, The Grail knights on Solomon’s ship read the explanatory letter (© Bodleian Library)

ANOTHER SHORT NOTE

257

7. Oxford, Bodl. Douce 215, Queste del saint Graal, fol. 32, Adam and Eve expelled from Paradise (© Bodleian Library)

8. Manchester, The John Rylands University Library, French 1, Agravain, fol. 73, Lancelot breaks the bars of Morgan’s prison to pick a rose (©The John Rylands University Library of Manchester)

258

9. olim Amsterdam, BPH 1i (ex-Phillipps 1045/7), L’Estoire del saint Graal, fol. 95v, Champie initials (© Lancelot-Grail Project) 10. olim Amsterdam, BPH 1iii (ex-Phillipps 3630 ii), Charette, fol. 45 Lancelot crosses the Sword Bridge (© Lancelot-Grail Project)

ANOTHER SHORT NOTE

11. Saint-Omer, BM 270, Psalter, fol. H, Gilbert de Sainte-Audegonde before the Virgin and Child (photo: author)

259

260

time between the two dates, perhaps slightly closer to 1323 because, when the minutiae of style are compared, some of its foliage types, for instance, are closer to those in St-Omer 270, and the combination of Gothic arch and geometrical frame is particularly close in both (cf. figs. 2 and 11). It is not known where Gilbert de Sainte-Aldegonde commissioned the painting he added to his MS. His activities can be associated with St-Omer on the one hand (the Chartreuse of Longuenesse is nearby) and perhaps Tournai on the other (if his psalter was really made for use at Tournai), but the geographical distribution of the other clients to which this workshop catered appears to extend substantially further east, crossing not only political borders but ecclesiastical ones as well, and including a substantial amount of illustrated Flemish texts as well as French ones. Did these clients commission their books from a shop based in St-Omer or possibly Tournai, or was it the artists who moved? This is not the place for a complete analysis of every dimension of this question, but I present here a brief outline of the stylistic breakdown of the shop’s products and the geographcial anchors they suggest, in order to better situate the Amsterdam/Douce/Rylands MS and as another step towards the clarification of the operations of this complex shop.18 Vitzthum noted stylistic links between Add10292–4, Royal 14 E.III and Stowe 17, a book of hours of the diocese of Liège; and the main hand of Add 28784 can also be cited here.19 The second hand in Add 28784 worked on a further corpus of MSS which are largely Liège-related but also have links, and perhaps origins, in Cologne books like those illustrated by Johann von Valkenburg in 1299.20 Can one therefore assume that this second hand 18 See also Stones Illustrations of Lancelot, ch. 6 for a preliminary analysis of this shop, and ead., ‘Three Illuminated Alexander Manuscripts,’ especially notes 11, 31 and 49. A fourth Alexander from the same workshop as the three discussed in the 1982 article has surfaced in a private collection in San Francisco, to which James Marrow kindly drew my attention; it contains scenes that occur nowhere else in the Old French tradition. See the forthcoming study by D. J. A. Ross (published with A. Stones as ‘The Roman d’Alexandre in French Prose: another illustrated manuscript’). 19 Vitzthum, Pariser Miniaturmalerei, p. 133. Add 28784 is reproduced in Randall, Images in the Margins, where many of the other MSS discussed above and below are also illustrated (Add 10292–4, Add 36684, Douce 5–6, GKS 3384 8o, M.754, Stowe 17, Rylands fr. 1, Trinity B.11.22, Walters 82) but without comment about their stylistic links. See Stones, Illustrations of Lancelot, ch. 6. More has been done with Add 28784 by Oliver, see The Lambert-le-Bègue Psalters; ead., ‘The Crise Bénédictine’ and ‘Reconstruction of a Liège Psalter-Hours’; but links with the Amsterdam/Douce/Rylands group are not made explicit.

ANOTHER SHORT NOTE

261

represents ‘native’ Liège activity, whose sources are in the Rhine, while the main hand of Add 28784 and that of Stowe 17 are an importation from the west? Or is the main hand of Amsterdam/Douce/Rylands, Add 10292–4, Royal 14 E.III, Stowe 17 and Add 28784 also a Liège-based painter with sources in the Rhine, who moves west?21 I return to this question below, in the context of an examination of the third style of the Lancelot group and its origins. The second painter of the Douce/Rylands set does not reappear in the other two Lancelots, but Add 10292–4 contains, in the middle, the work of its own second painter, whose work I discuss below. The second hand in Douce/Rylands does reappear in a Cambrai book, Baltimore, Walters 82, whose calendar points to a Benedictine house dedicated to St Lawrence in the diocese of Cambrai;22 but other books by this painter were made for patrons in Ghent: Oxford, Bodl. Douce 5–6, a psalter made for the abbey of 20 See Vitzthum, Pariser Miniaturmalerei, passim, and Oliver in Wallraf-Richartz Jahrbuch, 32, 1979. 21 Important here is the role of the Breviary of St-Jacques, Liège, Darmstadt 394, which in many ways is close to the Lancelot group but has much more sophisticated modelling of drapery and a palette in which pale mauve is dominant. See Oliver, ‘The Crise Benedictine’. It looks like a later phase in the activity of the same artist, but if so it is hard to reconcile with the idea of him moving west, an argument for suggesting that several individuals painted in this style. Also important to the question of artists’ movements are the close links, recognized by Vitzthum, Pariser Miniaturmalerei, p. 216, with the Breviary of Prüm (dioc. Trier), Berlin, SBPK, theol. 2o271. The answer to the question of how the links work with MSS made for use in the diocese of Trier may be found in the Prague section of the Metz Pontifical made for Renaud de Bar, bishop of Metz, between 1302 and 1316, where there is a painter whose work, to judge from Kvet’s reproduction, is remarkably close to that of one of the Amsterdam/ Douce/Rylands hands. Not having seen the book I am not sure whether the first or second Amsterdam/Douce/Rylands painter is closer. See now (having seen the manuscript), Stones, Gothic Manuscripts, Part II, vol. 1, Cat. IV–9 where I compare the second Amsterdam/ Douce/Rylands artist. 22 See Randall, Images in the Margins, passim; Stones, Illustrations of Lancelot, ch. 6. My thanks to Christopher Hohler for the liturgical attribution. The script of this book is very distinctively spurred and does not reappear elsewhere in the MSS discussed here. It warrants further investigation. My list of illuminated MSS in this group in 1970 also included Douce 5–6, Copenhagen GKS 3384 8o, Trinity B.11.22, and the Weigel antiphonary cuttings (sold at Christie‘s on 16.11.2011 as lot 13). Kerstin Carlvant has added Add 29253, Bruges GS 77/ 98 and The Hague K A XX to this artist‘s output, see, in addition to her dissertation cited in n. 17 above, ‘Collaboration in a Fourteenth-Century Psalter’. Carlvant points here to another subdivision of hands in the illumination of the Copenhagen Psalter, where the second hand is an assistant. See now Stones, Gothic Manuscripts, Part I, vol. 2, Cat. nos. III–71–79 for this group.

262

St Peter, Blandin, Ghent, and a breviary also of St Peter’s, London, BL Add 29253; there is also a ritual for the Cisterican house Ter Doest, not far from Bruges, and now in the Groote Seminaar, Bruges 77/98; and two books for Franciscan use, the hours in Cambridge, Trinity B. 11.22 and the psalter, Copenhagen Royal Lib.GKS 3384 8o. The Franciscan books have no precise indication of provenance. Also by this hand are some antiphonary cuttings formerly in the Weigel Collection (formerly Martin le Roy), of unknown use or provenance, and the Spieghel Historiael of Jacob van Maerlant, The Hague, KA XX. What has not so far been stressed is the importance of the presence of this painter in the Amsterdam/Douce/Rylands MS and its implications for the structure of the shop: not only does the geographical range now encompass Cambrai and Ghent, in addition to its St-Omer and possibly Tournai connections, through Amsterdam/Douce/Rylands, but the scope of the illumination also includes books in the vernacular written in Flemish as well as in French, in addition to books made for Cistercians, for Benedictines, and for clients with Franciscan interests. This group raises similar questions to those posed by the Liège links of the first painter of Amsterdam/Douce/Rylands: is this work done by a branch of the main shop, does the second painter of Amsterdam/Douce/ Rylands set up on his own, in Ghent or Cambrai, or are the patrons simply coming from further away? A Faits des Romains MS in Paris, BnF fr 295, published by F. Avril,23 sheds interesting light on this question: one of its artists is closely similar to the second hand of Amsterdam/Douce/Rylands, and this artist worked in Italy in collaboration with an Italian painter. Avril links him with the painter ‘Jean d’Ypres’ mentioned in Neapolitan documents of the early fourteenth century. A Moralia in Job in Grenoble may also be by this painter and is also closely related to the Amsterdam/Douce/Rylands second hand.24 If the work of the second painter extends the range of the shop not only as far east as Liège but also south of the Alps, there seems to be yet another geographical spread both east and south through another painter whose work is closely connected to the same atelier, though who does not seem to have collaborated directly with the two illuminators of Amsterdam/Douce/ Rylands discussed so far. Basically the style is very similar to both those hands but the drawing is looser and more fluid, the figures are presented Avril, ‘Trois manuscrits napolitains’. See Avril, who most kindly drew my attention to this MS and the question of the itinerant artists working in this style, ‘Un atelier “picard”’. 23 24

ANOTHER SHORT NOTE

263

on a larger scale, the foliage motifs are bigger too and are accompanied by a particular repertoire of white linear decorative motifs. Here there is a second named artist: Michiel van der Borch, who is named in 1332 as the illustrator of The Hague, M–W 10.B.21, a Rijmbijbel by Jacob van Maerlant.25 To him I would also attribute a slightly earlier work in French: The Tresor of Brunetto Latini, Vat. Reg. lat. 1320, written before 132726 and a number of devotional books: Stonyhurst 62,27 BnF nal 3178,28 and The Hague 135.E.15.29 These devotional books all contain rubrics in Dutch and indications of use in the diocese of Utrecht; while the Brunetto Latini text, though written in French, was already by 1327 in the possession of Henry of Ventimiglia.30 Stylistically, there is the same level of variation in detail among the Michiel van der Borch MSS, as observed above in the work of the main painter of Amsterdam/Douce/Rylands, and again it is always possible to explain such differences in terms of the participation of assistants; I assume, again, that they are rather the expression of the different stages in Michiel van der Borch’s career, which the date of before 1327 for the Vatican MS suggests was not substantially later than those of the first or second painters of Amsterdam/Douce/Rylands. If Michiel’s devotional books show that his clientèle had shifted somewhat to the east, the inclusion of a French work among his products suggests that his activities were based 25 See most recently Ekkart, De Rijmbijbel van Jacob van Maerlant. Comparisons are drawn with Douce 5–6, Copenhagen 3384 8o and The Hague KA XX, ibid., pp. 39–40. See now van Oostrom, Maerlants Wereld; Biesheuvel, Maerlants Werk; Meuwese, Beeldend Vertellen. 26 Stones, Illustrations of Lancelot, ch. 6. 27 Alexander and Crossley, Medieval and Early Renaissance Treasures no. 22 (not illustrated: photographs at the Warburg Institute, University of London). 28 I thank F. Avril for kindly drawing my attention to this MS. 29 See Carlvant, ‘Collaboration in a Fourteenth-Century Psalter’, p. 162, n. 52, where the link with The Hague MW 10.B.21 is made. Another MS, BL Add 24681, reproduced without comment in Beer, ‘Pariser Buchmalerei’, pl. 10, seems to fall in between the Michiel van der Borch MSS and the work of the second painter in Douce/Rylands and related MSS: in general the figure style is closer to the latter, while certain motifs, particularly the roses that are so prominent in the borders, are closer to Michiel van der Borch’s work. At the same time there is a second painter involved in the painting of the figures (for instance, the Virgin of the Annunciation in the full page miniature reproduced in Beer‘s article — and also the Virgin Annunciate in the historiated initial?) who would otherwise not strike one as remotely related to the workshop as a whole. Could this also be the hand of Princeton 44–18, a book of hours for the use of Chartres? 30 It is not by the same hand as BnF fr 295 or the Grenoble MS. There is an Italian participant, but his work is a later addition.

264

not too far from the orbit of the Francophone world; what is more difficult to explain is the early acquisition of the Vatican MS by its Italian owner since one cannot be sure whether it was the patron, the painter, or the book which travelled. Finally there is the second hand in Add 10292–4, whose work occurs in the middle volume, Add 10293, and who was, I therefore assume, an important collaborator of the main hand.31 This artist’s other independent activities bring one back once more to the orbit of St-Omer and Thérouanne since he also painted the Hours of Thérouanne or St-Omer, London, BL Add 36684 and New York, Morgan Library M.754 (after 1318 because of the inclusion of an indulgence of that year),32 as well as Boulogne-surMer 130, a Vincent of Beauvais Speculum historiale that is directly copied from Boulogne 131, made in 1297 for Eustache Gomer de Lille, abbot of St-Bertin.33 As with the other painters discussed here, the vernacular texts include not only the popular Arthurian romances in French, but at least one important Flemish MS as well: this time it is another of Jacob van Maerlant’s texts, The Hague KA XVI, Der naturen bloeme.34 For the work of this painter, the indications point more consistently to St- Omer or Thérouanne and there are no suggestions, as with the MSS painted by the other hands, that patrons were based elsewhere as well. Furthermore, there are elements of stylistic continuity between the Add 10293 painter and MSS made in the same region in the previous decades, of which the copying of the illustrations of Boulogne 131 in Boulogne 130 is symptomatic.35 If the strong St-Omer and Thérouanne connections apply to two of the painters in the Lancelot group, these places are probably more compelling as For my preliminary list, see Stones, Illustrations of Lancelot, chapter 6, passim; see now Stones, Gothic Manuscripts, Part I, vol. 2, Cat. no. III–130. 32 Facsimile of the New York section edited by Madame Th. Belin. For the date see P. Gerson’s unpublished paper on deposit in the file at the Morgan Library. 33 See L’Art du Moyen Age en Artois, ed. Porcher, no. 66, and Stones, Illustrations of Lancelot, pp. 449–50 for the idea that it is Boulogne 130 that depends on Boulogne 131 and not vice-versa as Porcher thought; see now Stones, Gothic Manuscripts, Part II, vol. 2, pp. 258–9. 34 This MS is normally dated c. 1350 but its connections with Add 10292–4 seem to have gone unnoticed and must surely suggest that the MS is earlier than has been thought. See Deschamps, Middelnederlandse Handschriften, no. 35. 35 The MSS most closely associated with Boulogne 131 (written in 1297) are Walters 90 (after 1297), Marseille 111, BnF fr 95, Bruges Stadb. 45/144. See Stones, Illustrations of 31

ANOTHER SHORT NOTE

265

the suggested provenance of the Amsterdam/Douce/Rylands MS than the alternatives: Ghent, Cambrai, Tournai, or the dioceses of Liège and Utrecht, or still further afield. It is probably more convincing to postulate St-Omer or Thérouanne as the location of a workshop whose products include both Flemish and French works than a base in predominantly Flemish-speaking Ghent or largely French-speaking Cambrai or Tournai, although of course any location along the linguistic frontier would cater for both languages; Ypres, because of the reference to Jean d’Ypres, is also a possibility, but one which is unsupported by other positive evidence.36 Nothing specific is known about the patron of the Amsterdam/Douce/ Rylands volumes. Neither they nor Add 10292–4 nor Royal 14 E.III can be directly connected with the patronage of Gilbert de Sainte-Aldegonde since they, unlike St-Omer 270, contain neither his shield nor his name, and he seems not to have left an inventory from which such a MS could be identified. It has likewise not so far been possible to connect these books to items in the inventories that do survive. These volumes are a little too late to have been the ones mentioned in the inventory of Jean d‘Avesnes, count of Hainaut and Holland from 1299 to 1304, ‘Premiers, uns grans roumans a rouges couvertures, ki parolle de Nasciien de Mellin et de Lancelot du Lach’.37 Similarly the will of Jean Cole of Tournai, dated 1303, included a Merlin among six or seven literary texts and compendia.38 It is hoped that

Lancelot, Chapter 5 and ead., ‘Three illuminated Alexander manuscripts’, nn. 11, 31, 49. K. Carlvant added MS Morgan 929 to this group, see her report to the Trustees of the Pierpont Morgan Library on file at the Morgan Library. An important transitional manuscript between this group and the works attributable to the Add 10293 painter is the Sellers Hours, in a private collection in Dallas, Texas (now at the Bridwell Library, Southern Methodist University, MS 13), published by Greenhill, ‘A Fourteenth-Century Workshop.’ Professor Greenhill relates the Sellers Hours to the hours BnF lat 1394 and the Bible historiale BnF fr 152; her other comparisons, M.805 and its group, seem to me unrelated. See now McQuillan, ‘Who was Thomas of Lancaster ? and The Sellers Hours’ and Stones, Gothic Manuscripts, Part II vol. 2, pp. 206, 209 n. 43. 36 Dehaisnes shows that at St-Omer, which was located well within the Flemish-speaking area, documents relating to such items as marriage contracts and land transfers were written in French from the early thirteenth century, although the official use of both languages persisted well into the sixteenth; he observed that Flemish was spoken in the outlying suburbs in the late nineteenth century. See Dehaisnes, ‘Delimitations du français et du flamand’. 37 Dehaisnes, Documents et extraits, pp. 143–8, discussed in Stones, Illustrations of Lancelot, pp. 115–27. 38 See de la Grange, ‘Choix de testaments tournaisiens’, esp. p. 38.

266

future work on the unpublished inventories of the regions associated with the painters of these volumes will help to clarify more fully the circumstances of production of these popular romances and the books related to them.

XIV Les débuts de l’héraldique dans l’illustration des romans arthuriens

A

u quinzième siècle, comme l’ont bien démontré Gerard J. Brault et Michel Pastoureau,1 le roi Arthur et les chevaliers de sa Table Ronde sont pourvus d’armoiries qui se retrouvent avec une certaine continuité d’un recueil héraldique à un autre, ainsi que dans les manuscrits arthuriens de l’époque. C’est pour aider le spectateur à bien reconnaître Lancelot que les peintres de deux manuscrits ayant appartenu à Jacques d’Armagnac ont représenté le héros dans la charrette portant un écu d’argent à trois bandes de gueules.2 En fait, selon le texte, Lancelot est déguisé à ce moment-là et ses amis ne le reconnaissent pas.3 Évidemment ces armoiries, qui deviennent de rigueur pour Lancelot, dépendent de l’épisode, dans le Lancelot propre, où la demoiselle de la Dolereuse Garde lui remet successivement trois écus d’argent à une, deux et trois bandes de gueules qui lui donneront la force d’un, de deux ou de trois hommes.4 On ne s’étonne donc pas de retrouver Lancelot portant Publié dans Cahiers du Léopard d’or, 8, 1997, pp. 395–420. 1 Brault, Early Blazon, (c. r. par Stones, Speculum 49, 1974, p. 319–320) ; Pastoureau, Armorial des chevaliers de la Table Ronde. 2 BnF, fr 119, f. 312v°, manuscrit que Jacques d’Armagnac hérita de son arrière-grandpère Jean de Berry, et dans lequel il fit faire plusieurs changements par ses propres peintres, y compris les armoiries, et BnF, fr 115, f. 355v°, commandé par le duc d’Armagnac lui-même (Micha, éd., Lancelot, t. Il, frontispice). L’épisode de la charrette n’est pas illustré dans le recueil arthurien qu’il avait aussi commandé, BnF, fr 112. L’étude iconographique de ses manuscrits a été faite par Blackman, The Manuscripts and Patronage of Jacques d’Armagnac et ead.,« A Pictorial Synposis ». 3 Pour ces mêmes armoiries utilisées comme déguisement dans la Mort Artu, voir infra et note 18. 4 Brault, Early Blazon, pp. 46–47; Pastoureau, Armorial, p. 82, citant Sommer, Vulgate Version, t. III, p. 147–151; voir aussi Micha, Lancelot, t. VII, p. 320, par. XXIVa 14; Ken-

268

d’argent à trois bandes de gueules dans le recueil héraldique intitulé Noms et blasons des chevaliers de la table ronde, Milan, Trivulziana, cod. N 1395, datant de la fin du XVe siècle, ainsi que dans celui d’Oxford, Bodl., e Mus. 78, écrit au XVIe siècle par le héraut français Montjoie, et dans un recueil du XVIIe siècle, Bruxelles, KBR, ms. 19132–66.5 Par contre, les débuts de l’héraldique dans les manuscrits arthuriens sont loin de présenter un tableau unifié des armoiries de ces héros. Si mes collègues Brault et Pastoureau ont eu tendance à rechercher les plus anciens exemples des armoiries qui vont par la suite devenir de rigueur pour les chevaliers, et cherché à démontrer ainsi le développement positiviste du phénomène, mon examen des origines des armoiries des chevaliers de la Table Ronde présentera un tableau quelque peu révisionniste : il me semble que les débuts de l’héraldique arthurienne sont en fait assez chaotiques et que la stabilité dont témoignent les armoriaux tardifs a mis longtemps à s’établir. Dans la période formative, entre 1225 environ et 1315, l’emploi de l’héraldique dans les manuscrits arthuriens témoigne souvent d’attitudes individuelles, particulières et créatrices, dictées non pas par le désir de se conformer à des normes déjà établies – car il y a encore énormément de variété dans l’emploi et le traitement des armoiries des chevaliers arthuriens à cette époque – mais par un sens de l’expérimentation qui se fonde aussi bien sur l’exploitation et le transfert de motifs visuels que sur l’interprétation des textes. C’est dans le ms. Londres, BL, Add 10292–4, daté de 1317 (ns) et fait à notre avis à Gand, Tournai ou Saint-Omer, que G. Brault a constaté la présence de plusieurs armoiries qui anticipent celles des armoriaux de la fin du Moyen Âge.6 En fait, ce manuscrit ainsi que ses proches parents, Londres, BL. Royal 14 E.III, et la copie divisée entre la Rylands Library de Manchester, ms. Fr. 1, la Bodléienne d’Oxford, Douce 215, et (anciennement) la Biblioteca Philosophica Hermetica d’Amsterdam, ms. 1, font une sorte de coupure entre la période formative des cycles d’illustration de ces textes (vers

nedy, Lancelot do Lac, t. I, p. 187–188. Une indispensable table de concordance entre les trois éditions du texte se trouve dans Kennedy, Lancelot and the Grail, Appendix II, pp. 321–346. 5 Brault, Early Blazon, pp. 40–41, signale le ms. de Milan dans la note 1, p. 41; Pastoureau, Armorial, pp. 35–38. 6 Brault, Early Blazon, p. 47. Le tableau complet des armoiries des chevaliers dans ce manuscrit, ainsi que dans les deux copies apparentées, reste à faire. Pour l’étude stylistique de ce groupe, voir Stones, « Another Short Note », repris dans ces essais.

LES DÉBUTS DE L’HÉRALDIQUE

269

1225–1300) et l’établissement d’iconographie et d’héraldique relativement stables, dont on trouve encore le reflet dans les codifications héraldiques de la fin du Moyen Âge. Nous tenons pourtant à apporter quelques modifications à ce résumé des développements, pour indiquer que l’héraldique arthurienne du XIVe siècle est souvent nettement moins stable que dans le groupe Add., Roy., Ryl./Oxford/Amsterdam. Les exemples que nous présentons ici se situent dans le cadre d’une étude approfondie sur l’iconographie des cinq branches du Cycle de la Vulgate (Lancelot en prose, Lancelot-Graal) dont nous tirons quelques extraits. Nous nous trouvons en effet devant un vaste corpus d’illustrations dans environ cent-quatre-vingts manuscrits, dont certains contiennent jusqu’à quatre cents miniatures. Ils s’étalent dans une fourchette chronologique des environs de 1225 jusqu’à la fin du XVe siècle,7 et leur distribution géographique est concentrée dans le Nord de la France et les Flandres, avec quelques manuscrits copiés et peints en Angleterre et en Italie.8 L’étude exhaustive des armoiries partirait de l’analyse interne de chaque manuscrit, pour se poursuivre comparativement, héros par héros, à travers les diverses branches du texte et les nombreux manuscrits. Nous nous limiterons ici à quelques exemples des armoiries de Lancelot et de Galaad, choisis pour illustrer diverses attitudes envers la fonction et le traitement de l’écu dans la première moitié du XIVe siècle, et au XIIIe siècle. Deux exemples des armoiries de Lancelot dans l’enluminure des manuscrits du Lancelot propre datant de la première moitié du XIVe siècle offrent des points de contraste avec Londres, BL, Add 10292–4. Dans le Lancelot de la Bibliothèque Pierpont Morgan, New York, ms. M.805–6,9

7 Nous rappelons que de nombreux manuscrits illustrés du Lancelot en prose sont antérieurs aux plus anciens manuscrits illustrés subsistant des œuvres de Chrétien de Troyes, où il est parfois également question d’héraldique arthurienne. Voir Pastoureau, « Les armoiries arthuriennes », ainsi que les appendices iconographiques, avec descriptions héraldiques et illustrations, par Stones et Busby, dans Les Manuscrits de Chrétien de Troyes, pp. 266–303. 8 Esquisse préliminaire dans notre thèse, Illustrations of Lancelot. Malgré les très nombreuses erreurs de datation et l’absence de mention de la présence de miniatures, la meilleure liste de manuscrits est toujours celle de Woledge, Bibliographie et Supplément. Nous avons apporté quelques corrections à la datation de certains manuscrits, voir Stones, « The Earliest », aux p. 42–44, et ead., «Arthur’s Death » aux p. 87–101. 9 Daté du XVe siècle par Micha et Kennedy, le manuscrit se situe en fait aux environs de 1315 et a été produit peut-être à Saint-Quentin, comme nous l’avons démontré ailleurs, voir Stones, « La Vie de sainte Benoîte d’Origny ».

270

Lancelot porte neuf écus différents.10 L’un est en effet d’argent à deux bandes de gueules, qui apparaît une fois, alors que l’on rencontre d’argent à une bande de gueules vingt-trois fois. Les autres écus sont d’argent sans figure, suivant le texte (dix fois) ; de gueules sans figure (six fois) ; de gueules au lion d’argent (une fois); de gueules à une bande d’argent (trois fois); d’argent à cinq couronnes d’azur (une fois);11 d’azur au lion d’argent ou d’or – remplacés par la couleur écrue – (une fois); de sable à « menues gouttes » – besants ? – d’argent (une fois, au fol. 60, fig. 5).12 Cette grande variété contribue à l’effet décoratif du manuscrit et reflète une attitude déjà très développée au XIIIe siècle où, comme nous allons le voir, les alternances de couleurs et de figures sont très importantes – plus importantes, parfois, que l’idée de l’écu comme signe d’identification d’un chevalier particulier. Mais quelques observations techniques s’imposent. Si le bleu de l’écu de Lancelot dans M.805–6, ainsi que le vert porté par d’autres chevaliers, font partie de la gamme normale des couleurs, l’absence du rouge vermillon et du noir en dehors de ce contexte héraldique se fait remarquer. Ce rouge vermillon est réservé pour les initiales filigranées du texte, alors que dans les enluminures les peintres n’emploient qu’un rouge-orangé ou un rouge-grenat. Le noir n’apparaît que pour les traits du dessin. S. Cockerell avait indiqué, à juste titre, que certaines miniatures avaient été repeintes et il nous semble que certains écus, surtout ceux où apparaissent le rouge et le noir, ont aussi fait partie d’une campagne de repeints.13 D’autres écus sont incontestablement authentiques, notamment ceux qui sont peints en bleu, vert, ou blanc, ainsi que l’écu fendu offert à Guenièvre par la demoiselle du lac, symbole de l’amour des deux amants. L’écu redeviendra entier lors de la consommation de l’amour et il sera utilisé pour guérir Lancelot de la folie. Dans le manuscrit de la Bibliothèque Pierpont Morgan, l’écu porte la double image d’un chevalier et d’une dame qui se baisent : séparés l’un de 10 Voir la description faite par Cockerell dans A Descriptive Catalogue of Twenty Illuminated Manuscripts, n° LXXXVIII, pp. 94–116, liste des armoiries aux pp. 98–99; il importe de se méfier car de nombreux écus ont été repeints, voir n. 13 infra. 11 Variante des armes de Galehot au fol. 155, d’argent à trois couronnes d’azur, dont les couleurs sont indiquées dans une note marginale, voir Stones, « Secular Manuscript Illumination » repris dans ces essais, à la fig. 2. 12 Ainsi décrit dans le manuscrit, pour Hector qui, lui, porte ce même écu onze fois : Sommer, Vulgate Version, t. IIl, p. 278; Micha, Lancelot, t. VIII, p. 148, par. LVa 6; Kennedy, Lancelot do lac, 1.1, p. 367. Le motif de l’écu s’explique, selon le texte, par la douleur que ressentit Hector d’avoir promis de ne pas tuer le chevalier qui avait insulté son amie. 13 Cockerell avait remarqué les réparations, très évidentes aux fol. 1 et 99. Il nous semble que de très nombreuses retouches ont été faites, qui visent surtout l’héraldique.

LES DÉBUTS DE L’HÉRALDIQUE

271

l’autre de chaque côté de l’écu, ils se retrouvent unis et l’écu reconstitué, dès que Lancelot et Guenièvre ont passé une première nuit d’amour.14 Il n’est nullement question de coloris dans la description de cet écu symbolique; en effet, le peintre l’a laissé en blanc. Les ecus de Lancelot dans le ms. Paris, BnF, fr 122, f. 1 (fig. 6), fait probablement à Tournai en 1345, en font un cas compliqué aussi par la question d’authenticité d’une part, et de circonstances particulières d’autre part. La représentation de Lancelot, portant un écu d’or à la fasce de sable chargé de trois besants d’argent, témoigne d’une certaine stabilité des armoiries du héros, car il porte le même écu dans seize autres miniatures du manuscrit; mais cet écu ne se retrouve pas, à notre connaissance, dans le répertoire des armoiries de Lancelot au Moyen Âge.15 Peut-on y voir les armes d’un commanditaire, ou d’un propriétaire (dont on ne connaît toujours pas l’identité), qui a voulu faire adopter son écu par le héros? Un curieux pendant se trouve dans la Bible historiale de New York, Bibl. Pierpont Morgan, ms. M.322–323, fait selon toute apparence dans les années 1340 à Paris, où un membre de la famille Rochechouart a doté de son écu, à une époque incertaine, l’antihéros Goliath.16 De plus, il y a dans le ms. BnF, fr 122, des indications suggérant que les armoiries actuelles font également partie d’un remaniement. Sur certains folios, on aperçoit les traces d’une couche de peinture plus ancienne sous les teintes actuelles de l’écu; l’existence d’armoiries antérieures pour Lancelot semble être confirmée par un unique exemple, au fol. 188 où celui-ci paraît avec une ailette de gueules ? (peinte en brun clair) à un sautoir recroisetté d’argent, indication, peut-être, de ce qu’ont été ses armoiries originales.17 On s’étonne de ne pas y retrouver d’argent à 14 Sommer, Vulgate Version, t. III, pp. 304–305, 409–410, 414–419; Micha, Lancelot, t. VIII, p. 204–208, par. LVIIIa 13–16, p. 444, par. LXXa 31–34, pp. 453–464, par. LXXIa 1–17; Kennedy, Lancelot do lac, 1.1, pp. 401–403, 547, 552–558. Voir l’analyse de quatre images montrant la demoiselle du lac apportant l’écu dans : Oxford, Bodl., Ashmole 828, f. 98v, Londres, BL, Add 10293, f. 90v, Paris, BnF, fr 344, f. 25, et BnF, fr 16999, f. 102 – le dernier étant en or sans figure –, par Carol Dover dans Word and Image, ed. Busby. Nous remerçions le Dr. Dover de nous avoir communiqué son texte avant sa publication. 15 Il n’y a aucune correspondance non plus dans les tables de L. Jéquier dans Cahiers d’héraldique, I, [1975]. 16 New York, Bibliothèque J. Pierpont Morgan, ms. M.323, f. lv, reproduit dans Stones, « Fauvel, with a Note on Fauvain », pl. 4. 17 L’identité d’un éventuel propriétaire n’est pas connue. Cf. de gueules au sautoir d’argent dans l’Armorial Wijnbergen, n° 1193, sans propriétaire (Adam-Even et Jéquier, « Wijnberghen ».

272

trois bandes de gueules et on se demande si cette ailette se rapporte, à son tour, aux armoiries personnelles d’un propriétaire. Ainsi l’emploi de l’héraldique dans BnF, fr 122, et dans M.805–6, s’écarte des normes établies dans Add 10292–4 et qui se retrouvent aussi bien dans les manuscrits de Jacques d’Armagnac au XVe siècle que dans les armoriaux, plus tardifs encore. Dans le texte de la Mort Artu on n’apprend pas quelles étaient les vraies armes de Lancelot. Seulement lorsqu’il doit participer au tournoi de Kamaalot et défendre la reine Guenièvre contre Mordret, il est obligé de se déguiser ; c’est à ce moment-là qu’il prend « un escu blanc a trois bendes de bellic vermeilles » puis « unes armes blanches et un escu a une bende de bellic », qualifié plus tard comme « unes armes blanches et en son escu une bende de bellic de synople », grâce auxquels il ne sera reconnu que par Hector et Boort.18 La tradition illustrative n’offre pas moins de trente-quatre autres possibilités pour les armoiries de Lancelot, y compris « d’argent à trois bandes de gueules », et en comptant aussi le sinople (vert) pour le champ de son écu dans Bruxelles, KBR, ms. 9627 (vers 1250), et dans Paris, Bibl. Arsenal, ms. 3482 (v. 1340).19 Seul, de nouveau, BL, Add s’approche le plus du texte ; mais là aussi il y a plusieurs bandes sur l’écu, non une seule; et les mêmes armoiries à plusieurs bandes se retrouvent dans Roy. 14 E.III. Pourtant, dans la Mort Artu, comme dans la Queste, les données textuelles indiquent que l’identification d’un chevalier par ses armoiries était toute naturelle et attendue. Les participants au tournoi de Wincestre dans la Mort Artu vont être ainsi reconnus, c’est pourquoi Lancelot devra emprunter des armes pour se déguiser;20 en effet, Gauvain y reconnaît Hector par ses armes, sans que celles-ci soient décrites dans le texte.21 Au cours de son ultime combat contre Mordret, Lancelot identifie de même le plus jeune fils de Mordret, « si encontra le fil Mordet le plus juenne et le connut bien as armeûres, car il portoit autiex armes comme ses pere souloit fere »,22 bien que, de nouveau, celles-ci ne soient pas décrites. La mention de couleurs dans la Queste et la Mort Artu se limite strictement à «vermeille », « noir », « sinople » et « blanc », couleurs qui non seulement se Frappier, Mort Artu, pp. 79, 97,103. Table comparative jusqu’en 1340 dans Stones, Illustrations of Lancelot, pp. 359–384. Notons que, jusqu’au XIVe siècle, le terme « sinople » signifie encore «gueules», voir Brault, Early Blazon, p. 275; Pastoureau, Traité d’héraldique, p. 103. 20 Frappier, Mort Artu, p. 8. 21 ibid., p. 15. 22 ibid., p. 256. 18 19

LES DÉBUTS DE L’HÉRALDIQUE

273

repèrent dans les armoiries de certains chevaliers, mais reflètent aussi une dimension de symbolisme s’étendant entre les pôles opposés du blanc qui indique le bien, et du noir qui symbolise le mal. Les mauvais chevaliers du tournoi symbolique dans la Queste portent le noir,23 ainsi que celui qui tente de faire pécher Lancelot.24 Dans la Mort Artu, Morgain raconte à Arthur que Lancelot a porté la couleur noire au moment où son fils Galaad s’est fait chevalier;25 en fait, les armoiries de Lancelot pour cette occasion ne sont pas décrites et il faut en conclure que Morgain veut simplement présenter de lui une mauvaise image. Les bons chevaliers portent le blanc : celui qui, dans la Queste, empêche le roi Baudemagus de s’emparer de l’écu destiné à Galaad,26 celui qui fournit un cheval à Galaad,27 ainsi que les bons chevaliers du tournoi symbolique.28 Elyezer, le fils du roi Pellés, porte le blanc29 comme les nombreux ermites et moines qui apparaissent de temps en temps dans la Queste.30 Le blanc est également la couleur du cerf symbolique protégé par quatre lions.31 Le blanc, le vert et le rouge sont les couleurs de l’arbre de la vie, blanc après la chute de l’homme, vert après la naissance d’Abel, rouge après le meurtre d’Abel par Caïn.32 Du bois de l’arbre de vie ont été fabriqués à chaque étape les trois « fuissiaux » – petites pièces de bois – qui se trouvent sur le bateau sans aile et sans gouvernail, « de naturel color blanc et vert et vermeil sanz peinture nule ».33 Le vert est la couleur de la souffrance comme Pauphilet, éd., La Queste del saint Graal, p. 143. ibid., p. 146. 25 Frappier, Mort Artu, p. 63. 26 Pauphilet, La Queste del saint Graal, p. 29. 27 ibid., p. 252. 28 ibid., p. 140. 29 ibid., p. 143. 30 La mention fréquente du blanc a été interprétée comme une allusion aux cisterciens; mais aucune référence précise à cet ordre monastique n’apparaît dans le texte. De plus, la couleur de l’habit cistercien au treizième siècle est loin d’être uniquement le blanc. Voir le catalogue d’exposition, Saint Bernard et le monde cistercien, éd. Pressouyre et Kinder. Un symbolisme très général est évident, par exemple, dans l’emploi du blanc pour le linge d’autel et pour les sous-vêtements liturgiques dès l’époque paléochrétienne, qui se transfère tout naturellement à l’apparence du Graal dont l’iconographie (à la différence des traitements textuels) est calquée en grande partie sur celle des sacrements de l’eucharistie. Cf. Pauphilet, La Queste del saint Graal, p. 78. Dans L’Estoire, par contre (Sommer, Vulgate Version, I, p. 126), le blanc signifie la virginité et l’humilité. 31 Pauphilet, La Queste del saint Graal, p. 234. 32 ibid., p. 218–219. Voir aussi notre article « Signs and Symbols », repris dans ces essais. 33 ibid., p. 226. 23 24

274

l’émeraude,34 tandis que le rouge s’associe au feu.35 À la fin de la Mort Artu, c’est une enseigne rouge que fait monter Lancelot sur le château de Joyeuse Garde pour indiquer à Boort qu’il est temps de commencer, par derrière, l’attaque contre l’armée d’Arthur.36 Au début du récit, Lancelot emprunte au fils de son hôte, pour ne pas être reconnu au tournoi de Wincestre, un écu et des « couvertures » vermeils ! 37 une « armes pleines » de jeune chevalier, le plain étant utilisé dans d’autres textes comme couleur de déguisement, ainsi que l’a démontré G. Brault.38 Lorsque, dans la Queste, Galaad arrive à la cour du roi Arthur pour devenir chevalier, il porte « unes armes vermeilles, sans espee et sans escu » – donc un surcot vermeil plutôt qu’un écu. L’écu qu’il reçoit est « blanc a une croiz vermeille »;39 c’est celui fourni par Josephé, fils de Joseph d’Arimathie, pour être donné au roi Eualach dans l’Estoire.40 La croix est la seule figure héraldique décrite dans le texte de la Queste, pendant, de ce point de vue, des armes de Lancelot dans la Mort Artu. En même temps, les armoiries de Galaad, héros de la Queste, sont liées au symbolisme des couleurs de l’arbre de vie. L’épisode des trois compagnons sur le bateau, avec le lit et les « fuissiaux », lie souvent ces deux éléments, en représentant et les « fuissiaux » dans les trois couleurs, et les écus des héros.41 Cette combinaison de motifs 34 ibid., p. 124. Cette interprétation de la symbolique de l’émeraude diffère, par exemple, de celle de Bède, PL 93, 198, qui appuie sur la fraîcheur, dans le sens littéral et spirituel, de la foi des fidèles; voir Friess, Edelsteine im Mittelalter, pp. 174–175. La souffrance doit être prise, peut-être, dans le sens du renoncement des biens terrestres par les fidèles, suivant l’interprétation de l’émeraude par Alexandre Neckam, De naturis rerum libri duo, chap. 50–55, cité par Friess, pp. 178–179. Dans le ms. 164 de la Bibliothèque Méjanes d’Aix-enProvence (Nord de la France, 3e quart du XIIIe siècle) il est précisé au f. 76v que l’émeraude signifie « la vert foi de la trinité... et la grand verdure de la grant foi qui toiorz dure... ». 35 Pauphilet, La Queste del saint Graal, p. 78. Dans L’Estoire, par contre (Sommer, Vulgate Version, t. I, p. 129), le rouge signifie le sang. 36 Frappier, Mort Artu, pp. 140, 145. 37 ibid., p. 8, 22. 38 Brault, « Plain arms in Arthurian Literature ». Brault cite également le Lancelot propre, « et li escuz estoit toz blans comme nois si com a cel tans estoit costume que chevaliers noviaus portoit escu d’un sol taint le premier an que il esroit » (Sommer, Vulgate Version, t. Ill, p. 299; voir aussi Micha, op. cit., t. VIII, p. 194, par. LVIIa 7, et Kennedy, Lancelot do lac, 1.1, p. 395); Brault mentionne aussi des exemples tirés du Tristan en prose et d’Escanor. 39 Pauphilet, La Queste del saint Graal, p. 7, 28. 40 Sommer, Vulgate Version, 1.1, p. 48. 41 Pauphilet, La Queste del saint Graal, p. 226. Une variante de cet épisode est incluse dans 11 manuscrits illustrés de la Queste faits entre 1250 et 1340 environ, voir Stones, Illus-

LES DÉBUTS DE L’HÉRALDIQUE

275

offre par ailleurs des échos visuels qui se rapportent aux antécédents dans l’iconographie de l’Estoire. Les manuscrits apparentés de près à Add 10292–4, Roy. 14 E.III et Rylands/Oxford/Amsterdam incluent tous une image de cet épisode dans la Queste, avec de légères différences de configuration des personnages et de l’héraldique. Le bateau avec les trois « fuissiaux » peints en blanc, vert et rouge est présent dans chacune des trois images.42 Dans les mss. Add f. 44 (fig. 1) et Roy., les armoiries de deux des chevaliers sont les mêmes : Galaad porte d’argent à une croix recroisettée de gueules, Perceval de gueules plain (un rouge-orangé) ; dans le ms. Add, Boort est doté d’argent (gris) au lion de gueules et dans Roy. il n’a pas de marque héraldique. On se souviendra que c’est dans le ms. Add que G. Brault a puisé un grand nombre de ses exemples d’un emploi d’armoiries stables.43 Dans la copie divisée entre Rylands/Oxford/Amsterdam par contre, Perceval (c’est lui qui lit la lettre placée dans l’aumônière que les chevaliers trouvent sur le lit) ne porte pas d’armoiries, sauf si l’on tient compte de la couleur rose que l’on voit sur le revers de son ailette ; et les deux autres chevaliers portent un écu d’argent au canton de gueules et un écu de gueules (rose) à la bordure d’argent. Dans ces trois exemples, l’identification héraldique est réalisée soit par l’écu, soit par l’ailette, soit par le surcot, le manque de figure sur ce dernier laissant ambiguë l’identification des chevaliers dans le manuscrit d’Oxford par rapport aux deux autres copies. Il en est de même dans un manuscrit légèrement antérieur de l’université de Yale à New Haven, Beinecke Library, ms. 229, f. 257v (fig. 7).45 L’épisode est divisé en deux scènes superposées. Dans le registre supérieur apparaît Perceval, lisant la lettre, vêtu d’un surcot rosâtre avec une ailette au fascé de rose foncé et de rose clair, alors que les deux autres chevaliers portent trations of Lancelot, pp. 318–319; ce sujet n’a été inclus ni dans Oxford, Bodl., Digby 223, ni dans Paris, BnF, fr.110, ni dans Bonn, LUB 526; Oxford, Bodl., Douce 199 a une lacune à cet endroit. 42 Londres, BL Add 10294, f. 44; Roy. 14 E.III, f. 130v; Oxford, Bodl., Douce 215, f. 35. 43 Brault, Early Blazon, p. 47, n. 2 ; p. 48, n. 2 ; p. 50, n. 3. 44 On reconnaît cet écu ailleurs dans le manuscrit, comme dans Bodl., Digby 223, fait une génération auparavant, ainsi que dans les traditions tardives, comme étant les armes de Gauvain; voir Stones, Illustrations of Lancelot, chap. 3. Pour les armes de Gauvain voir Stones, « Note on the Heraldry of a Very Special Gauvain, » repris dans ces essais. 45 Voir Shailor, Catalogue of the Beinecke Rare Book and Manuscript Library, 1, n° 229; pour le contexte stylistique, voir Stones, Illustrations of Lancelot, chap. 4, et ead. «Illustrations of BN fr. 95 and Yale 229 ». Le manuscrit fait pendant à celui de Paris, BnF, fr. 95.

276

respectivement un surcot de gueules (rouge-orangé) et un écu d’argent (gris), et un surcot d’argent (gris) et une ailette au fascé de gueules foncé et de gueules clair (rouge-orangé de deux tons). De plus, le dessin des écus et des ailettes a été souligné par un trait blanc à l’intérieur du trait noir, qui pourrait représenter une bordure ou bien simplement avoir été ajouté pour faire mieux ressortir les couleurs. En outre, sur le registre inférieur, l’on voit un chevalier portant un surcot rouge-orangé sans écu, un second un surcot gris sans écu, et un troisième, un surcot rouge-orangé et une ailette au fascé de gueules foncé et de gueules clair (rouge-orangé), ce qui ne correspond même pas à la distribution des couleurs des vêtements au registre supérieur. Le peintre de Yale, ms. 229 n’a pas inclus les trois « fuissiaux ». En fait, le lit est très élégant et détaillé, peint en gris avec des touches finales en or sur les extrémités et au milieu des côtés longs (fig. 7). Dans les trois autres manuscrits, l’emplacement des « fuissiaux », ainsi que l’emploi des couleurs, le blanc, le vert, et le rouge, évoque en partie le modèle inventé dès les débuts de la tradition d’illustration du Cycle de la Vulgate pour la première branche du texte, L’Estoire. Dans la plus ancienne copie, Rennes, BM, ms. 255 (fig. 2), faite dès la composition des textes aux environs de 1220 comme nous l’avons montré ailleurs,46 les « fuissiaux » font partie de la structure même du lit (« dont le lit était environné et clos », fol. 46),47 détail qui, à notre connaissance, est peu repris par la suite.48 Dans L’Estoire du ms. Bonn, LUB, Voir Stones, « The Earliest ». La tradition illustrative de la Queste n’a pas conservé d’exemplaire plus ancien que celui-ci. Toutefois, les dates des manuscrits ne reflètent pas la chronologie de la composition des textes, selon laquelle L’Estoire aurait été composée après la Queste. 47 Sommer, Vulgate Version, t. 1.1, p. 136. 48 Selon le texte, Nascien est seul, alors que le ms. de Rennes, BM 255, le montre accompagné d’un groupe de cinq personnes sur la rive à côté du bateau. L’on peut se demander si cette image a été copiée d’après une image inventée pour l’épisode similaire, dans la Queste, où la découverte du bateau magique est vue par plusieurs spectateurs (cf. BnF, fr 1424, voir infra, note 50). Toutefois la fin du ms. de Rennes manque, le texte conservé s’arrête à Sommer, Vulgate Version, t. IV, p. 220, et aucun manuscrit apparenté n’a été conservé. Cette image suggère qu’à l’origine le ms. de Rennes faisait partie d’un cycle plus complet. Dans Roy. 14 E.III, fol. 51 (3e image), Nascien regarde le bateau sur lequel se trouvent le lit avec les « fuissiaux » blanc, vert et rouge, et une épée et une couronne sur le lit. Dans Add 10292, fol. 31 (1ère image), version très proche de celle de Roy., Nascien est sur le bateau (à voile et gouvernail) sur lequel se retrouvent aussi le lit et les mêmes éléments. Cet épisode manque dans la copie olim Amsterdam/Rylands/Oxford. Dans le ms. BnF, fr 95 (vers 1290–1300) – qui fait pendant à Yale, 229, au f. 52v – les « fuissiaux » ne sont pas inclus, l’épée et la couronne sont absents, et Nascien a été transformé en chevalier armé d’un écu rose chargé d’une rose. 46

LES DÉBUTS DE L’HÉRALDIQUE

277

1. Londres, BL, Additional 10294, f. 44 (Photo : British Library). Queste : Perceval lisant la lettre, accompagné de Boort, de Galaad et de sa sœur, sur le bateau à lit et « fuissiaux ».

2. Rennes, BM, 255, f. 46 (Photo : Lancelot-Grail Project). Estoire : ses compagnons, sur la rive, regardent Nascien sur le bateau à lit et « fuissiaux ».

278

3. Bonn, LUB, 526, f. 413 (Photo : Landes-und-Universitätsbibliothek Bonn). Queste : Galaad combat Perceval et Lancelot.

4. Paris, BnF, fr 95, f. 44v° (Photo : BnF). Estoire : Nascien prie devant la Main de Dieu, son écu à côté de lui; tenant son écu, il s’apprête à monter sur le bateau à lit.

LES DÉBUTS DE L’HÉRALDIQUE

279

5. New York, Pierpont Morgan Library, M.805, f. 60 (Photo : Morgan Library). Lancelot : Lancelot, monté sur un des chevaux de Gauvain que le valet de ce dernier vient d’amener, s’attend à recevoir un message de la part de la reine, qui, accompagnée de la dame de Malehaut et de ses demoiselles, lui fait signe du haut d’une tour ; Lancelot, accompagné de deux chevaliers, désarçonne à coups de lance un des chevaliers de Galehot. 6. Paris, BnF, fr 122, f. 1 (Photo : BnF). Lancelot : Lancelot traverse le pont de l’épée; il combat deux lions et un léopard, puis, à cheval, se bat à coups de lance contre Meleagant, fils du roi Baudemagus, devant le château de ce dernier.

280

7. New Haven, Yale University, Beinecke Rare Book and Manuscript Library, 229, f. 257v (Photo : Beinecke Library, Yale University). Queste: Perceval et Boort lisent la banderole d’explication; sur la rive, la sœur de Perceval et Galaad tiennent l’épée. Ils partent sur leur bateau à voile jusqu’au château Carcelois.

8. Bonn, LUB, 526, f. 27 (Photo : Landes-und-Universitätsbibliothek Bonn). Estoire : le bateau à lit et « fuissiaux ».

LES DÉBUTS DE L’HÉRALDIQUE

281

9. Le Mans, M. M., 354, fol. 59 (Photo : LancelotGrail Project). Estoire : Séraphe est converti en voyant la guérison d’un homme estropié qui touche l’écu d’Eualach tenu par Josephé.

10. Bonn, U. B., 526, fol. 57v° (Photo : Landes-und-Universitätsbibliothek Bonn). Estoire : le roi Mordrain tenant son écu, sur lequel Josephé fait une croix de son sang.

282

11. Amsterdam, B. P. H., 1, fol. 30 (Photo : Lancelot-Grail Project). Estoire : le blanc chevalier, représenté sous les traits de saint Georges, intervient pour blesser mortellement le roi Tholomer contre qui luttent le roi Eualach et son armée.

12. Manchester, The John Rylands University Library, Fr. 1, fol. 188v°. (Photo : The John Rylands University Library of Manchester). Queste : le blanc chevalier, représenté sous les traits de saint Georges, explique à Galaad, accompagné de son valet, la signification de son écu.

LES DÉBUTS DE L’HÉRALDIQUE

283

526, écrit par Arnaud de Keu ou de Cayeux (Kayo) en 1286,49 le bateau et le lit sont représentés avec les « fuissiaux » aux trois couleurs devant le lit, au f. 27 (fig. 8), sans Nascien, alors que dans la Queste du même manuscrit, la découverte du bateau au lit et à l’épée, par les trois compagnons et la soeur de Perceval, fait défaut.50 Par ailleurs, les armoiries de Galaad dans la Queste apparaissent dans les miniatures dès le milieu du XIIIe siècle, dans les manuscrits de Bruxelles, KBR 9627, f. 56,51 et Paris, BnF, fr 339, f. 233v, lors du premier combat de Galaad contre Perceval et Lancelot.52 Dans Bonn, LUB, 526, f. 413 (fig. 3), par contre, l’écu de Galaad est d’azur à une croix d’argent, tandis que l’on L’étude de ce motif dans l’illustration de L’Estoire reste encore à faire. Voir Ponceau, Étude de la tradition manuscrite de L’Estoire del saint Graal; Remak, Text and Image in the « Estoire del Saint Graal; Meuwese, L’Estoire del saint Graal. Pour la Queste, voir Stones, Illustrations of Lancelot, chap. 10. 49 Discussion sur le cadre artistique dans Stones, Illustrations of Lancelot, chap. 5. 50 Dans les plus anciens manuscrits, produits vers 1250, Paris, BnF, fr. 339, f. 257, et Bruxelles, KBR ms. 9627–8, f. 56, ainsi que dans Paris, BnF, fr 342, f. 132, écrit peut-être dans les environs de Douai par une femme en 1274, les « fuissiaux » sont absents. Le colophon de ce dernier ms. invite le lecteur à prier pour « ce li » qui l’écrit, cf. Pope, From Latin to Modern French, p. 325, par. 845, cité par Stones, Illustrations of Lancelot, pp. 50, 410. Pour la discussion sur le contexte stylistique, dont nous citons quelques exemples plus bas, voir ibid., et Stones, « Prolegomena to a Corpus of Vincent of Beauvais Illustrations »; ead., « Illustrated Chrétien Manuscripts », pp. 252–253 (voir supra, note 7) ; in pace Braem, « Ein Buchmalereiatelier in Arras um 1274 » (liste très incomplète de manuscrits comparatifs). Dans Paris, BnF, fr 344, fait probablement à Metz vers 1300, il y a au f. 510, les « fuissiaux » mais pas d’écus; dans ce ms. une deuxième initiale au f. 512v montre, exceptionnellement, la couronne sur le lit à côté de l’épée. Dans Bodl., Rawl. Q. b. 6 à Oxford, fait vraisemblablement à Paris vers 1300–1310, la perspective a changé pour mettre en valeur l’épée; avec les trois chevaliers non identifiés, on aperçoit deux « fuissiaux » incolores posés sur le lit. Dans les deux manuscrits des environs de 1340, Paris, Bibi. Ars 3482, f. 494v, et Paris, BnF, fr 1424, f. 352, l’un parisien, l’autre provenant vraisemblablement de Tournai, il n’y a ni « fuissiaux » ni écus et le dernier manuscrit change le point de vue en mettant les trois chevaliers sur le bord de l’eau et pas sur le bateau (cf. Rennes, BM 255, f. 46, supra note 48). Les « fuissiaux » et les écus font défaut aussi dans les manuscrits de Jacques d’Armagnac, Paris, BnF, fr 116, f. 661, et BnF, fr 112, f. 170 (manquant dans BnF, fr 120) ainsi que dans BnF, fr 111, f. 262v. 51 Miniature très abîmée où l’écu blanc est évident, mais on voit mal la croix de gueules. Les armoiries de Galaad, d’argent à une croix de gueules, sont représentées aussi aux ff. 7v, 11,15v. 52 Pauphilet, La Queste del saint Graal, p. 56. Treize manuscrits contiennent une enluminure montrant ce combat, dont seulement six dotent Galaad de l’écu qu’il lui faut : Bruxelles, KBR, 9627, f. 15v, BnF, fr 339, f. 236v, Yale, 229, f. 204, Rawl. Q. b. 6, f. 324, Add 10294, f. 11v, Roy. 14 E.III, f. 98v; Ryl. fr. 1, f. 194v, n’a qu’une initiale en or à cet endroit, une main postérieure a ajouté une miniature dans la marge inférieure.

284

suppose que c’est Perceval qui porte de gueules à la bande d’argent et deux étoiles du même, et que Lancelot est sans écu. Comme nous l’avons montré ailleurs, le peintre du ms. de Bonn témoigne d’une ignorance extraordinaire des armes de Galaad lorsque, au début de la Queste, il fait pendre à une croix vermeille l’écu blanc, alors que la croix de gueules devait être figurée sur l’écu même.53 Évidemment, l’écu de Galaad dans la Queste est le même que celui du roi Eualach dans la première branche du cycle, l’Estoire. Sur cet écu, Josephé, fils de Joseph d’Arimathie, avait placé une croix de « cendal » pour protéger Eualach dans la bataille contre Tholomer – incident qui n’est pas représenté, à notre connaissance, dans la tradition illustrée de l’Estoire.54 Au moment de la conversion d’Eualach et de Séraphe, Josephé fait venir l’écu sur lequel on voit une représentation du Christ crucifié grâce à laquelle le bras d’un homme estropié est guéri, épisode représenté dans Le Mans, MM, ms. 354, f. 59 (fig. 9).55 Juste avant sa mort, Josephé fait une croix sur le même écu avec son 53 L’erreur dépend peut-être d’une rubrique – ou d’une note d’instruction – mal copiée, voir la discussion dans Stones, « Indications écrites et modèles picturaux », aux pp. 330–331 (repris dans ces essais). 54 Pour le texte, voir Sommer, Vulgate Version, t.1, p. 48. La même explication est donnée à Galaad dans la Queste, Pauphilet, La Queste del saint Graal, p. 32. Un sondage préliminaire indique que cet épisode n’est pas illustré dans Rennes, BM 255, Paris, BnF, fr 770, fr 95, fr 344, Bonn, LUB 526, Paris, BnF, fr 110, fr 19162, fr 24394, olim Amsterdam BPH 1, Londres, BL Roy. 14 E .III, Add 10292. Le geste de poser la croix de sandal (ou cendal) sur l’écu rappelle les croix en soie offertes aux croisés par le pape Urbain II en 1095, et qui devinrent leur emblème. Pour le discours d’Urbain II à Clermont, voir la chronique de Foucher de Chartres, citée infra, note 60. 55 Sommer, Vulgate Version, t.1, p. 74 avec variantes; pour l’édition du ms. Le Mans, MM 354, voir Hucher, Le Saint Graal. Ce manuscrit, écrit par Gautier de Cayeux ou Keu (Kayo – un parent d’Arnaud de Kayo, copiste de Bonn, LUB 526, en 1286? Un copiste nommé Gautier de Kayo a écrit une Image du monde, BnF, fr 14964, en 1282), fait partie des manuscrits groupés autour de Paris, BnF, fr 342, auxquels appartiennent les mss. arthuriens Oxford, Bodl., Digby 223 et Douce 303, Londres, BL Harley 1629, Paris, BnF, fr 770; les fragments, Archives d’État à Modène; les chroniques, Paris, BnF, fr 12203; les romans de Marques, Laurin et les Sept Sages, Paris, Bibl. Ars 3355 et Cambridge, Fitzwilliam Museum, McClean 179; les mss. du Trésor de Brunetto Latini, Arras, BM 1060, Bruxelles, KBR, 10228, Paris, BnF, fr 1110, Vatican, Vat. lat. 3203; le Speculum naturale de Vincent de Beauvais, Bruges, Stadsbibl., 251; le Légendaire, St. Pétersbourg, Bibliothèque de l’Académie des sciences, F 403; le martyrologe de Saint-Amé de Douai, Valenciennes, BM 838, le psautier-livre d’heures à l’usage de Saint-Amé de Douai, Bruxelles, KBR, 9391, et les heures du psautierlivre d’heures à l’usage d’Amiens, Philadelphie, Free Library, ms. Widener 9 (références dans Stones, Illustrations of Lancelot, chap. 3, ead., « Prolegomena », ead., « Illustrated Chrétien Manuscripts », voir supra note 50) . On peut ajouter à cette liste les le Recueil littéraire, Paris,

LES DÉBUTS DE L’HÉRALDIQUE

285

sang – épisode correctement illustré par le même peintre qui avait introduit l’erreur dans Bonn, ULB 526, f. 57v (fig. 10).56 Et pendant la bataille contre Tholomer un blanc chevalier vient au secours d’Eualach, portant ses armes, «.j. escu blanc a une vermelle crois ».57 Dans le ms. olim Amsterdam, 1, f. 30 (fig. 11), le blanc chevalier est nimbé et, exceptionnellement, il est identifié dans la rubrique comme étant saint Georges. Le blanc chevalier sous les traits du saint a été à deux reprises représenté ailleurs sur ce manuscrit, dans la Queste : pour délivrer des prises du roi Baudemagus l’écu destiné à Galaad (Ryl, f. 188) et pour expliquer à ce dernier la signification de l’écu (Ryl, f. 188v, fig. 12). La présence de saint Georges dans la Queste est encore plus inattendue car le texte ne spécifie pas la présence de la croix rouge sur l’écu.58 Ces peintres – deux artistes ont participé à l’illustration du ms. Ryl. / Douce / Amsterdam, ce qui explique les différences stylistiques entre les images dans les figures 11 et 12 – ont connu l’iconographie de saint Georges dans le contexte des livres de dévotion, car une miniature, dans un psautier fait dans le même cercle artistique et conservé à Copenhague, représente le saint chevalier portant une ailette et un écu d’argent à une croix recroisettée de gueules, ainsi qu’une lance à bannière aux mêmes armes.59 De plus, l’intervention miraculeuse de saint Georges dans l’Estoire et la Queste rappelle une autre intervention miraculeuse de sa part, BnF, fr 795. Les manuscrits arrageois, Arras, BM 309 et 307, ainsi que le psautier de New York, M.730, n’appartiennent pas à ce groupe (Braem, voir supra note 50). Quoique la ville de Douai se situe dans le diocèse d’Arras, rien ne justifie l’attribution à Arras même proposée par Braem, « Ein Buchmalereiatelier in Arras »; les fortes différences stylistiques des manuscrits faits pour Arras suggèrent le contraire. 56 Sommer, Vulgate Version, t. I, pp. 284–5. Voir aussi Paris, BnF, fr 95, f. 108, Londres, BL Roy. 14 E.III, f. 85v (2e image) et Add 10292, f. 73; aucune illustration dans Rennes, BM 255, Paris, BnF, fr 770, fr 110, fr 19162, fr 24394, olim Amsterdam BPH 1; aucun écu dans Paris, BnF, fr 344, f. 78v. 57 Sommer, Vulgate Version, t.1, p. 62, v. 22–23. Illustré dans Paris, BnF, fr 95, f. 25v, où le chevalier est sans nimbe et porte un surcot rouge-orangé et un écu blanc à une croix vermeille; sans nimbe également dans Le Mans, MM 354, f. 49v. L’étude comparative reste à faire. 58 Pauphilet, La Queste del saint Graal, p. 29. Il est intéressant de constater que les deux manuscrits apparentés de près, Londres, BL Roy. 14 E.III, f. 94v et Add 10294, f. 6v, traitent cet épisode autrement: dans Roy., le blanc chevalier porte un surcot et une ailette blancs sans croix; dans Add., il n’y a pas de miniature à cet endroit. 59 Copenhague, GKS, 3384, f. 192v, psautier à l’usage de Gand, illustré par les deux mêmes peintres que Ryl./Douce/Amsterdam. Voir Nordenfalk, Gyldene Boger, n° 54; Stones, Illustrations of Lancelot, pp. 225–226, 322–323, 472–473; Carlvant, « Collaboration in a Fourteenth-Century Psalter »; Stones, « Another Short Note », pp. 189–190.

286

dont le souvenir, dans la région des Flandres où fut fait ce manuscrit, jouissait d’une importance particulière. Selon les récits de la première croisade, saint Georges aida les chrétiens luttant contre l’infidèle à Antioche en 1098, et de nouveau à Jérusalem.60 Le succès de la première croisade fut marqué dans le Nord de la France et les Flandres par le don de la relique du Saint Sang à Boulogne-sur-Mer en 1101 par Godefroi de Bouillon,61 et à Bruges en 1148 par Thierry d’Alsace.62 Le culte du Saint Sang dans ces régions semble avoir joué un rôle très particulier dans l’iconographie du Graal dans le ms. Ryl. / Douce / Amsterdam, ainsi que dans Roy. 14 E.III et Yale, ms. 229, comme nous l’avons montré ailleurs.63 Il est possible qu’un culte particulier de saint Georges dans ces régions y ait exercé un rôle aussi. La participation du saint à la première croisade avait été commémorée, exceptionnellement, à Saint-Omer. En effet, ouvrant la série des miniatures en pleine page, dans la Bible en images faite pour l’abbaye bénédictine de Saint-Bertin dans les

60 Cité dans Acta sanctorum (AASS), Aprilis, III, p. 153, d’après l’Historia Hierosolymitana de Robert le Moine et l’Itineris Hierosolymitani de Pierre de Tudebode; cf. aussi le récit dans les Gesta francorum, p. 66, ainsi que dans les chroniques de Foucher de Chartres, Historia hierosolimitana 1095–1127, pp. 216–233; d’Albert d’Aix, dans Recueil des historiens des croisades (RHC), Histoires occidentales (H O), IV, pp. 358–432 et de Raymond d’Aguilers, RHC, HO, III, pp. 241–259. Pour un résumé des légendes de saint Georges, voir Matzke, « The Legend of St. George ». Saint Georges a dû être souvent représenté aux XIIe et XIIIe siècles, comme en témoigne le cycle d’images, sur le mur nord de l’église Saint-Botolph à Hardham (Sussex), décrit dans Tristram, English Medieval Wall Painting, p. 31. La popularité du saint en Angleterre, dont le culte fut lancé par Richard Cœur de Lion, culminait dans la proclamation de sa fête, le 23 avril, comme fête nationale au synode d’Oxford en 1222. Une nouvelle impulsion fut donnée à l’iconographie de saint Georges par la fondation de L’ordre de la Jarretière sous son patronage, par Édouard III en 1348. Plusieurs portraits du saint existent dans les manuscrits anglais du XIVe siècle, voir Sandler, Gothic Manuscripts, 1285–1385, figs. 172, 217, 218, 351, 388, 343 (cr Stones). Pour l’iconographie de saint Georges en France, voir Deschamps, « La légende de saint Georges ». Nous tenons à remercier Rosemary Argent pour ses conseils sur les croisades, saint Georges, et Saint-Omer. 61 Conservée actuellement dans l’église Saint-François-de-Sales à Boulogne-sur-Mer dans le splendide reliquaire en émail de plique attribué à l’orfèvre parisien Guillaume Jullien, offert à l’église Notre-Dame de Boulogne par Philippe le Bel en commémoration du mariage de sa fille Isabelle avec Édouard II d’Angleterre en 1308, qui eut lieu dans cette église. Voir Trésors des églises de France, n° 42, pl. 111, et Gauthier, Les émaux du Moyen Age occidental, n° 158. 62 Offerte par le Patriarche de Jérusalem en commémoration de la participation de Thierry à la deuxième croisade. Pour le reliquaire, voir Vloberg, L’iconographie de l’eucharistie, p.149, se référant à Malou, Du culte du Saint Sang de Jésus-Christ. 63 Stones, « Illustrations of BnF, fr 95 and Yale 229 » (voir supra, note 45).

LES DÉBUTS DE L’HÉRALDIQUE

287

premières années du XIIIe siècle, La Haye, KB, 76 F 5,64 l’image du plan de Jérusalem surmonte celle de saint Georges, aux armes d’argent à la croix de gueules, mettant un sarrasin en déroute.65 Nous ne savons pas pour qui fut fait le ms. Ryl. / Douce / Amsterdam, mais il existe un lien – qu’il faut admettre très indirect – avec Saint-Omer grâce à un manuscrit qui lui est apparenté de près, qui fut offert par Guilbert de Sainte-Aldegonde à la chartreuse de Longuenesse près de Saint-Omer en 1323.66 Son calendrier et sa litanie, d’usage incertain, mettent en valeur saint Georges. La représentation de ce saint dans la Queste, et sa rubrique, montrent une image, et une interprétation de celle-ci, qui sont devenues beaucoup plus spécifiques et particulières que la simple mention du chevalier blanc dans le texte, et qui évoquent une autre dimension d’associations. Saint Georges est devenu le type non seulement du chevalier chrétien et croisé, mais aussi du bon chevalier, et son image est transférée comme telle, de la même façon que celle de saint Martin partageant son manteau avec le pauvre. Ce dernier, qui se présente d’ailleurs sur la même page que l’image de saint Georges, dans le psautier de Copenhague illustré dans le même cercle que Ryl. / Douce / Amsterdam, devient le type par excellence de la charité67. Comme nous l’avons montré ailleurs pour le cas de l’ermite posé dans l’arbre, dans le ms. Douce, les peintres du manuscrit Ryl. / Douce / Amsterdam ont su exploiter sur plusieurs niveaux l’idée du transfert de signification par le moyen d’un jeu de modèles picturaux qui vont au-delà des mots littéraux du texte68. Pour parvenir à leurs interprétations visuelles, ils ont fréquemment utilisé des intermédiaires sous forme d’esquisses et de deux sortes de Byvanck, Les principaux manuscrits à peintures de la Bibliothèque Royale des Pays-Bas et du Musée Meermanno-Westreenianum à La Haye, pp. 13–15, n° 3, pl. III–IV; Schatten van de Koninklijke Bibliotheek, n° 19. 65 Reproduits dans Ornamenta ecclesiae, t. III, n° H 4, aux pp. 76 et 77. On a daté le plan des environs de 1170, ce qui nous paraît un peu tôt pour le reste de l’enluminure. 66 Saint-Omer, BM ms. 270, voir l’analyse du calendrier dans Leroquais, Les psautiers manuscrits, II, p. 203, pl., reprise dans Stones, « Another Short Note », p. 188. Il n’est pas certain que le calendrier soit à l’usage de Tournai, comme le pensait Leroquais, puisque Éleuthère est absent; une localisation à Saint-Omer n’est pas justifiée non plus. Saint Georges est dans le calendrier et il suit saint Piat dans la litanie. 67 Ainsi l’image de saint Martin partageant son manteau est souvent choisie, vers 1300, pour illustrer le chapitre sur la largitas dans l’Éthique d’Aristote; voir, par exemple, Avranches, BM 22 et Boulogne-sur-Mer, BM, 110, etc. 68 Voir Stones, « Indications écrites » (supra note 53), repris dans « Arthurian Art Since Loomis », aux pp. 39–41, pl. 18–23. 64

288

notes marginales, à l’intention d’une part du rubricateur et de l’autre de l’enlumineur69. Aussi peut-on vraiment croire que, dans ce manuscrit, l’iconographie parfois inattendue – puisque le texte ne l’explique pas – a été le résultat d’un choix très conscient, et que ce qui pourrait apparaître comme une erreur ou une déviation curieuse témoigne en fait d’un esprit libre et créateur70. Un deuxième cas de la concrétisation de l’idée représentée par un écu à la croix se trouve dans le ms. BnF, fr. 95, f. 44v (vers 1290–1300, fig. 4). Nascien, à genoux devant le bateau merveilleux, fait (dans le texte) un signe de croix71; le peintre l’a représenté les mains jointes en prière, le signe de croix devenu concret sous la forme insolite d’un écu d’azur, relevé de traits losangés noirs, à une croix aussi d’azur et coticée de rouge-orangé. C’est l’unique apparition dans le manuscrit de cet écu, qu’on ne relève pas non plus dans le manuscrit apparenté, Yale, 229. Même Nascien, dans d’autres miniatures, porte un écu rose à figure en forme de rose et de couleur plus foncée que le fond de l’écu72; et, comme nous l’avons vu dans ce dernier manuscrit (fig. 7), les couleurs portées par les chevaliers peuvent changer même entre les deux parties d’une seule miniature. C’est dans ces exemples insolites, voire déviants du point de vue de l’évolution éventuelle de l’héraldique arthurienne et de son iconographie, qu’apparaît un emploi de l’héraldique et de l’écu souvent plus original et complexe que dans l’ensemble de la tradition. Ces variantes nous permettent d’examiner de plus près quelques-unes des considérations qui gouvernèrent la création des images, et de saisir quelques niveaux de perception et de réception héraldiques qui dépassent les sources textuelles. Elles ajoutent une importante dimension à l’étude de l’héraldique arthurienne dans son ensemble.

Stones, « Indications écrites », aux pp. 325–327, pl. 5–8,19–20; voir aussi la mention de Ryl./Douce/Amsterdam dans Alexander, Medieval Illuminators, pp. 112–115. 70 Nous ne savons pas exactement qui fut responsable : le peintre, son patron ou directeur de travail, ou le commanditaire. 71 Sommer, Vulgate Version, t.1, p. 120. Un sondage préliminaire indique que cet épisode n’est pas normalement illustré : il manque dans Rennes, BM 255, Bonn, LUB 526, Paris, BnF, fr 344, Londres, BL Add 10292, Roy. 14 E.III; il y a une lacune dans le ms. d’Amsterdam à cet endroit. 72 Voir Yale, 229, f. 52v (supra, note 48). 69

XV Lancelot and Identity

T

he story of King Lancelot, grandfather of Lancelot du Lac, occurs at the end of the Estoire del saint Graal, as the last part of a sequence about the descendants of Nascien who will reappear in the later branches of the cycle. The pious king is suspected (unjustly) of illicit relations with the wife of his cousin the Duke of Belle Garde, who plans revenge. As King Lancelot emerges from confession at a hermitage, he pauses to drink at a fountain, whereupon the Duke cuts off his head which falls into the fountain. When the Duke attempts to retrieve it the water boils so he cannot touch the head. He is punished for the murder by divine retribution when he is killed as his castle collapses on top of him.1 Meanwhile the King’s body rests in a tomb that weeps blood and is guarded by two lions who had fought each other for the possession of a stag and were wounded, but cured by the blood from the tomb, which they then guard, and would continue to guard until such time as Lancelot du Lac appears.2 This sequence forms a transition to the Lancelot proper, the third branch of the five-part cycle, which begins with the story of King Lancelot’s son King Ban, whose lands and castle are usurped by King Claudas and whose baby son Lancelot is captured by the Lady of the Lake. Ban dies of grief at the loss of his lands and his widow Queen Elaine takes the veil upon seeing her son disappear into the Lake, which she is powerless to prevent. The young Lancelot remains without knowledge of his parentage First published in Marqueurs d’identité dans la littérature médiévale: mettre en signe l’individu et la famille (XIIe–XVe siècles), ed. C. Girbea, L. Hablot, R Radulescu (Histoires de famille. La parenté au Moyen Age, coll. dirigée par M. Aurell, 17) (Turnhout, 2014), pp. 217–42. 1 For text editions of the Estoire del saint Graal and the placing of illustrations of this episode see Appendix A. 2 For text references, see Appendix A.

290

and ancestry until after he has embarked upon his adventures. It is only after having received from the Damsel of La Dolereuse Garde the shields argent with first one, then two, and finally three bends gules and defeated the knights of La Doleureuse Garde, that he finds the tomb with its slab of metal, decorated with precious stones, which is destined for him; he lifts the lid to reveal his name and parentage: Chi gerra lancelos del lac li fiex au roi ban de benoyc.4 He will learn more only when, much later in the story, he comes upon a severed head in a fountain of boiling water and a tomb guarded by lions: he kills the lions, retrieves the head, and reunites it with the headless body inside the tomb. A hermit hands him a letter explaining that the head and body are those of his grandfather King Lancelot, for whom he is named. These latter events, related in Agravain,5 the last part of the Lancelot proper, are a sequel to the murder of King Lancelot which occurs at the end of the Estoire del saint Graal. How do the illustrations treat these episodes ? This paper examines which copies devote pictorial emphasis to the story of the two Lancelots and what their similarities and differences are. The choice as to whether or not to illustrate the murder and its sequel and how to treat the subjects have much to tell about the interests of patrons and makers of the popular Lancelot-Grail romance and about the strategies they employed to convey this important theme: the presence or absence of illustrations is a measure of whether this was considered a key episode or not. The story of King Lancelot in the ESTOIRE DEL SAINT GRAAL The Estoire del saint Graal survives in some 44 manuscripts, most of which are illustrated with between one and 63 scenes.6 Surprisingly, most copies of the Estoire leave the episode of the murder of King Lancelot unillustrated. The fullest treatment of the subject is found in olim Amsterdam Bibliotheca Philosophica Hermetica 1, made most likely in Therouanne, Saint-Omer, or Ghent c. 1315, one of three copies made by the same team of anonymous

Brault, Early Blazon, pp. 46–47; Pastoureau, Armorial des chevaliers de la Table Ronde, p. 82; Stones, ‘Les débuts de l’héraldique’, reprinted in these essays. 4 Sommer, Vulgate Version, t. III, 152. 38; Kennedy, Lancelot do Lac, p. 194. 30; Micha, Lancelot, VII, p. 332. 5 For text editions of the Agravain and illustrations of this episode, see Appendix B. 6 For lists of the extant manuscripts see Woledge, Bibliographie, p. 71–79, and Supplément, p. 50–55; and Ponceau, L’Estoire, p. xxv–xlviii. 3

LANCELOT AND IDENTITY

291

scribes, decorators and illuminators for a patron who is also unknown.7 Four illustrations are devoted to the murder of King Lancelot in the Amsterdam copy, three of them grouped on the same page — a layout unique in the manuscript, and rare in the Lancelot-Grail manuscript tradition in general, giving particular prominence to this episode. On f. 117v is the first scene, depicting the Duke of Belle Garde trying to lift the head out of the boiling fountain (fig. 1). This is followed on the following recto (f. 118) by the other three miniatures: the immediate punishment of the Duke, showing his castle falling upon him and his courtiers; the tomb weeping blood; and the two lions guarding the tomb and licking the blood (fig. 2). The first two of these four scenes are also found in London, BL Royal 14 E.III, but with differences of detail as the castle crushes knights on the ground below (fig. 3). In Royal the text of the Estoire was left incomplete, lacking a final leaf and no doubt the last two miniatures.8 And the third member of this stylistic group, London, BL Additional 10292, omits the story of King Lancelot from its illustrations, one of a series of differences of selection between Add. and Royal/Amst.; the latter two copies are close to each other in conception although Amst has far fewer actual illustrations, often preferring a large champie initial. Amst. and Royal are not the only, nor the earliest, manuscripts to contain illustrations of the fate of King Lancelot, but none of the other 7 The three copies are olim Amsterdam, BPH 1/Manchester, The John Rylands University Library French 1/ Oxford, Bodleian Library Douce 215; London, BL Royal 14 E.iii, and BL Additional 10292–4. The Amsterdam copy was sold at Sotheby’s on 7 December 2010 as lot 33. Its present owner has chosen to conceal his or her identity, so it is convenient to continue to refer to it as olim Amst. These three copies form the core of the Lancelot-Grail Project where comprehensive bibliography by project participants is listed. The provenance of the group has been established on the basis of stylistic analogies, first with the donor portrait in Saint-Omer BM 270, offered in 1323 to the Chartreuse of Longuenesse by Gilbert de Sainte-Aldegonde; second by the participation in the Rylands/Douce section of the same manuscript of an artist who also worked on the Psalter-Hours of Ghent use, Oxford, Bodleian Library Douce 5–6; third by the participation in London, BL Add. 10292–4, of a painter whose work is also found in the Psalter-Hours of Saint-Omer Use, London, BL Add. 36684 and New York, The Pierpont Morgan Library M.754. Many more related manuscripts cluster around these. 8 The last leaf of the Estoire in Royal 14 E.iii, f. 88, is the last leaf of its quire, and Roger Middleton (personal communication) has read the largely cut off catchword in the bottom margin of its verso as the opening words of the Queste del saint Graal, suggesting that the Estoire was never finished and the second and third branches of the story — Merlin and Lancelot — were never included. This may indicate that Martine Meuwese is right in supposing that Amst. preceded Royal in terms of the sequence of their making — unless

292

illustrated copies devotes as many as four illustrations to the story. The earliest manuscript to illustrate King Lancelot, with a single miniature, is Bonn LUB 526 (f. 58v), written in 1286, where the murder is the chosen subject, both protagonists on horseback and King Lancelot’s head already in water below (fig. 4). This is another manuscript made most likely in Thérouanne or Saint-Omer (despite the scribe saying he wrote it in Amiens — as though Amiens was not normally where he worked), and a second copy, closely related, is Paris, BNF fr 110, made most likely a decade or so later,9 where the story, though present in the text, is left unillustrated.10 Flanders is not the only region to produce copies of the Estoire with this subject: Paris, BNF fr 344 is a product of Lorraine, most likely Metz, made at an uncertain date, most likely the last quarter of the 13th century.11 Here the Duke’s attempt to retrieve the head from the boiling fountain is the subject of the initial O on f. 80–2 (reproduced on Gallica so not shown here), while two miniatures follow the explicit, marking the end of the Estoire, on ff. 81 and 81v: the two lions affronted, fighting upright on their hind legs and then the two lions crouching and licking the tomb (f. 81, see Gallica12), and on f. 81v a miniature showing a seated hooded figure holding a book (the hermit who heard King Lancelot’s confession), looking at head in boiling fountain, a lion licking the tomb, a second lion licking a man lying on the ground (the body of King Lancelot?), and a crowd of men watching (see Gallica). The presence of the crowd is a puzzling feature: is it

both were in production at the same time (Meuwese, « Three Illustrated Prose Lancelots; see also Stones, « The Grail in Rylands French 1 and its Sister Manuscripts». Amst. is a more modest production, in that champie initials correspond to places in the text where Royal has a miniature. There are 40 such initials in Amst.’s Estoire as a whole, of which 36 correspond in placement to miniatures in Royal. Two of the four exceptions correspond to the omitted leaf at the end of Estoire in Royal and the other two (Amst. ff 20v and 21 ) are without a parallel scene in Royal. In Amst. they mark places in the description of Josephé’s Grail mass (illustrated on ff. 18 and 21), which is treated with particularly distinctive and detailed miniatures in Amst. 9 These are my attributions, based on stylistic similarity to Boulogne-sur-Mer BM 192, written in 1295, and Boulogne-sur-Mer BM 131, written in 1297 for Eustache Gomer de Lille, Abbot of Saint-Bertin at Saint-Omer. See Stones, ‘ ... entre Cambrai et Saint-Omer’. 10 BNF fr 110, ff. 45 and 45v. 11 Stones, ‘Tournoi de Chauvency’, p. 155 n. 7, 168, fig. 24 (f. 138). The entire manuscript is excellently well reproduced on Gallica. 12 I am not prepared to pay for BNF reproduction rights as all the images referred to here are reproduced on Gallica and/or Mandragore.

LANCELOT AND IDENTITY

293

a misplaced depiction of the companions of the Duke who are killed with him as his castle falls ? The two Paris manuscripts of the second quarter of the 14th century illustrated by the Sub-Fauvel Master both acknowledge the story of King Lancelot, but in different ways.13 BNF fr 105 gives only a rubric, without a miniature, in which the murder and its punishment are mentioned (f. 124–2, see Gallica),11 while in BNF fr 9123, the subject chosen is the lions guarding the tomb, shown as a slab with a nonsense inscription on the lid (f. 94–2, see Mandragore). Finally there is Paris, BNF fr 113, part of the complete Lancelot-Grail cycle made for Jacques d’Armagnac, Duke of Nemours († 1477), c. 1475 and illustrted by Evrard d’Espingues.15 Three pictures, the first two on the same page, depict elements of the story: the murder, about to be accomplished with a second blow of the Duke’s sword as King Lancelot leans over the fountain; and his tomb, elegantly carved with weepers on the tomb chest and a recumbant effigy of the king on the lid. There is however no sign of blood (see Mandragore). The third picture, on f. 116v, shows the two lions guarding the tomb, and one of them licking it (see Mandragore), an arrangement strikingly similar to what is found in Amst. Here it is worth noting that the episode is unillustrated in BNF fr 117, part of the set inherited by Jacques d’Armagnac from his great-grandfather Jean de Berry, and from the companion volume, Arsenal 3479 (in colour on Gallica), suggesting that Jacques d’Armagnac took a personal interest in the episode. What can be made of these comparisons ? First, it is stiking that, despite similarities of placing, no two manuscripts treat the subjects in quite the same way. The images are not simply copies one of another. Even the closely similar selection in Amst. and BNF fr 113 reveals differences in the treatment of the subjects chosen. The fullest pictorial treatment is that found 13 I prefer to distinguish the artistic level found in these and many other stylistically related books from the very high quality painting in the Roman de Fauvel, Paris, BNF fr 146 and a small number of other books, attributing the former group to a lesser artist and calling him the ‘Sub-Fauvel Master’. See Stones, ‘Artistic Context of le Roman de Fauvel ’. 14 A rubric without a miniature is also found in the copy of the Estoire made between 1480 and 1482 for Jean-Louis de Savoie, Bishop of Geneva, Brussels, BR 9246, and the wording of the rubric is the same as in BNF fr 105, in this instance and with few exceptions throughout the Estoire. See Stones, ‘L’Estoire del saint Graal dans la version adaptée...’ reprinted in these essays. 15 See Blackman, Manuscripts and Patronage of Jacques d’Armagnac, pp. 182–245 and ead., ‘A pictorial Synopsis’. The illustrations are reproduced on Mandragore.

294

1. olim Amsterdam, BPH 1, i, L’Estoire del saint Graal, f. 117v, The Duke of Belle Garde trying to lift the head out of the boiling fountain (© Lancelot-Grail Project)

3. London, BL Royal 14 E.III, L’Estoire del saint Graal, f. 88v, The Duke’s castle falls on him and his knights (© Trustees of the British Library)

LANCELOT AND IDENTITY

2. olim Amsterdam, BPH 1, i, L’Estoire del saint Graal, f. 118, His castle falls on the Duke; the tomb weeping blood; the two lions guarding the tomb and licking the blood (© Lancelot-Grail Project)

295

296

4. Bonn, LUB 526, L’Estoire del saint Graal, f. 58v, The Murder of King Lancelot (© Bonn, Landes-und Universitätsbibliothek)

LANCELOT AND IDENTITY

297

5. New Haven, Yale University, Beinecke Library 229, Agravain, Lancelot rides away from an abbey; he meets his companions; he meets a dwarf; he kills the lions at the tomb, he retrieves the head from the fountain, he presents it to the hermit who points to the tomb, he lifts the lid to reveal the shrouded body (© Yale University)

298

6. London, BL Add. 10293, Agravain, f. 334v, Lancelot meets a dwarf (© Trustees of the British Library)

7. London, BL Add. 10293, Agravain, f. 335, Lancelot having retrieved the head from the fountain hands it to the hermit for reuniting with the headless body in the tomb (© Trustees of the British Library)

LANCELOT AND IDENTITY

299

8. London, BL Add. 10293, Agravain, f. 335v, Lancelot and the hermit talking (© Trustees of the British Library)

9. London, BL Add. 10293, Agravain, f. 336, Lancelot and his squire follow the stag and four lions (© Trustees of the British Library)

300

10. Bonn, LUB 526, Agravain, f. 380v, Lancelot killing the lions who guard the tomb (© Bonn, LUB)

11. Manchester, The John Rylands University Library, Fr. 1, Agravain, f. 76v, rubric (© Lancelot-Grail Project)

LANCELOT AND IDENTITY

301

12. Manchester, The John Rylands University Library, Fr. 1, Agravain, f. 77, Lancelot shows the head he has retrieved from the fountain to the hermit (© Lancelot-Grail Project)

13. Manchester, The John Rylands University Library, Fr. 1, Agravain, f. 76v, champie initial (© Lancelot-Grail Project)

302

in Amst., which in other respects is spare in its illustrations — so this episode must have been considered critical to whoever had charge of constructing the illustrative program: whether producers or patron. Both are anonymous and the subsequent history of the manuscript casts no further light on its particular circumstances of production beyond what can be deduced from its style and iconongraphy. That the Estoire in Royal was allowed to end incomplete shows that, unlike the preference clearly expressed in Amst., the presence of the final page — which would undoubtedly have had the same two images as in Amst., was not considered critical at all — not worth embarking on a fresh quire for one page. This is surprising since otherwise Royal’s Estoire is the most fully illustrated of all the extant copies. Here the need to finish the book quickly may have been the determining factor. Bonn is limited in its illustrations compared with Royal so it is not surprising to find only a single image for this episode. BNF fr 344 is unusual in positioning the final miniature after the end of the text, a practice however that is also found in other branches of fr 344, and may be considered an idiosyncracy of the producers. As to the two Parisian copies, illustrated by the same painter, the unexpected thing is the difference in choice of subject between the two, while Jacques d’Armagnac’s copy with its emphasis on the beautifully carved tomb made for King Lancelot — for which there is no textual justification and no parallel in any other copy, and this tomb is not weeping blood as the text specifies — may reflect an interest in sepulchral monuments on the part of its patron. This is also the only copy to show the action of the head-cutting so the choice of moment is also distinctive — and ironic in that shortly after this Jacques d’Armagnac himself would be executed for treason — while the lions licking the tomb find parallels in BNF fr 344, BNF fr 9123 and in Amst. BNF fr 113 is the only copy whose commission is known and where a personal interested in the subject may be suggested as postulated above. Lancelot discovering his name This episode, related above,16 is an important step in Lancelot’s selfdiscovery since it provides him with both a name and with noble parentage, and anticipates the more elaborate explanation of his relationship to King Lancelot that will come later, in the Agravain. One might expect the episode of Lancelot’s discovery of his name to have been a popular choice in the 16

See n. 4 for text references.

LANCELOT AND IDENTITY

303

illustration of the Lancelot proper, but in fact only two illustrations of the tomb survive, and both show it made of stone, not of metal with the jewelled lid and inscription inside as in the text. They are BNF fr 118, f. 190v (see Mandragore), the copy purchased by Jean de Berry from Jacques Raponde in 1406, and inherited by his great-grandson Jacques d’Armagnac c. 1465, and its twin, Paris, Bibliothèque de l’Arsenal 3479, p. 420 (see Gallica)17 Although repainted in part under Jacques d’Armagnac (notably the figure of Lancelot), this subject was not chosen to illustrate the same passage in Jacques d’Armagnac’s copy of the Lancelot-Graal illustrated by Evrard d’Espingues, Paris BNF fr 113–116 (see Mandragore). It is possible, even likely, that Lancelot discovering his name once formed part of the distinctive illustrative cycle in the Morgan Lancelot, New York, The Pierpont Morgan Library M.805, made c. 1310 and illustrated in large part by the Master of the Vie de sainte Benoîte d’Origny. There is a lacuna between ff. 36v and 37 in M.805 which corresponds to this section of the text. The presence of a distinctive illustration — many of those in M.805–6 spread across all three columns of the text — could have tempted someone to remove a page containing such a picture (illustrations available on Corsair and on the Index of Christian Art). Lancelot discovering the remains of his ancestor Agravain, the last part of the Lancelot proper, was extremely popular, surviving in some 70 manuscripts, of which most contain one or more illustrations.18 Like the Estoire del saint Graal, the Agravain was usually copied together with other parts of the Lancelot-Grail cycle, whether the other parts of the Lancelot proper, or the Queste del saint Graal and the Mort Artu, or as part of the complete five-branch cycle. What is striking, as with the King Lancelot episode in the Estoire, is how few of the Agravain manuscripts include illustrations of Lancelot’s discovery of his identity: only eight copies illustrate the episode. The events take the pictorial form of six or seven scenes: Lancelot meets a dwarf, who tells him that the forest is full of adventures, Lancelot fights and kills the two lions who guard the tomb, he retrieves the head from the fountain, he presents the head to the hermit, he 17 Blackman, ‘A pictorial sysnopsis...’, p. 15; Hermant, ‘Le Lancelot-Graal de Jean de Berry et Jacques d’Armagnac’. 18 See Woledge, Bibliographie; Brandsma, ‘Lancelot Part 3’.

304

raises the lid of the tomb to reveal the headless body inside, he listens to the hermit who tells him how his grandfather died, he (sometimes with a vallet) sees the White Stag accompanied by four lions, which links this sequence to the Grail Quest.19 As was the case with the illustrations of the murder of King Lancelot in the Estoire, not all the manuscripts that include this episode in pictorial form give all seven scenes, and not all the manuscripts which included the Estoire episode also include the Agravain sequel and vice-versa. Three manuscripts, Yale 229, BNF fr. 344 and Add. 10293, have the encounter between Lancelot and the dwarf as a lead-in image, in Yale shown as part of a composite miniature, combined on the top level with the killing of the lions, and on the bottom with the retrieving of the head, the handing it to the hermit, and the revealing of the headless body in the tomb, to which the hermit directs Lancelot with a pointing finger (f. 110, fig. 5). It is an elaborate and complex illustration complete with a border scene depicting an amorous crowning. Here it is significant that no murder scene was included in Yale’s likely pendant, Paris, BNF fr 95, made by the same craftsmen and containing the Estoire and Merlin, so the Agravain scene was not anticipated pictorially but was none-the-less considered important in the Agravain programme, and placed on a page with the preceding episode about Lionel, creating a page-layout unique in Yale and memorable because of it. BNF fr 344 gives the encounter on its own, taking place on horseback — and this is the only illustration for the entire sequence, a remarkably meagre choice given the attention paid in the Estoire to the murder of King Lancelot (f. 427–2, see Gallica). Add., from the same workshop as Royal and Amst., did not include the murder in its Estoire illustrations but now devotes no fewer than four scenes to the Agravain episode: the encounter with the dwarf (f. 334v, fig. 6), the finding of the head and body (f. 335, fig. 7), the explanation (f. 335v, fig. 8), and the vision of the White Stag and lions (f. 336v, fig. 9). Yale and Add. both devote a separate illustration to the dialogue between Lancelot and the hermit (Yale f. 112), underlining the importance of the information being transmitted, but in Yale the subject is given a small historiated initial, whereas in Add. a miniature gives greater prominence to the conversation, another instance of a depiction of debate

See Appendix B for text editions, manuscripts, and table of subjects. For more on the White Stag and lions see Stones, ‘Le merveilleux’ reprinted in these essays. 20 See Stones, ‘Illustration et stratégie illustrative’, reprinted in these essays. 19

LANCELOT AND IDENTITY

305

which is characteristic of Add.’s pictorial treatment in general.20 The action in Add. is all crammed into the middle scene between the two miniatures of talk, whereas in Yale the miniatures give continuous action and the conversation is confined to the small, pictorially less significant, initial. In Bonn much is made of the killing of the lions (f. 380v/391v, fig. 10) and that is the only subject chosen in Arsenal 3480 (p. 324, see Gallica) and BNF fr 119 (f. 440v, see Mandragore). In fr 119 and Ars. 3480 the deed takes place at a cross, to which, incongruously, the reins of Lancelot’s horse are tied. The tomb is visible in fr 119 but not in Ars. and nothing is made of the retrieving of the head, nor of the explanation. A different approach is taken in Rylands, BNF fr 115 and BNF fr 111. In these three manuscripts a miniature is placed at the beginning of the episode, earlier in the text than in the other copies. In Rylands, the presentation of the head is the only subject depicted, and a champie initial Q marks the place in the text where Bonn, Yale, and Add. place their first or only illustration (ff. 76v–77, figs. 11, 12, 13). The miniature shows the presentation of the head to the hermit, clearly considered more important than the killing of the lions which plays such a prominent part in the depictions chosen in other copies and is also found as the first miniature in BNF fr 115 (f. 507v, see Mandragore) and 111 (f. 204–1, see Mandragore). What is surprising in Rylands is the choice of a single scene, whereas in Amst. (the first part of the same set of volumes) the murder and retribution in Estoire was treated with more pictures than in other copies. So the sequel in Agravain is present but not given the same pictorial weight as the beginning of the story in the Estoire. In BNF fr 115 Lancelot kills the lions by the tomb and fountain with head and in BNF fr 111 the tomb and fountain with head are also shown, and in addition, Lancelot killing the first lion is watched by his squire on horseback, then in a second miniature (f. 204–2, see Mandragore) the lions lie dead by the tomb, while Lancelot stands by the fountain and hands the head to the hermit. The vision of the White Stag and Lions serves to conclude the episode in Add. (f. 336v, fig. 9) by presenting the lions as guardians of the sacred stag rather than antagonists as they were at the beginning of the sequence where they fought with each other over a stag. Add. is the only copy to include the stag and lions at the conclusion of finding of King Lancelot’s head, reinforcing again the importance of the head-episode in the career of Lancelot du Lac. Other copies illustrate the stag and lions motif only when it reappears somewhat later in the Agravain; it is seen again by Lancelot, this

306

time in the company of Mordret, and an interpretation is given by a hermit who explains that it is a miracle of God’s work which can be understood only by the achiever of the Holy Grail. Bonn, Yale, BNF fr 344, Rylands, BNF fr 115 and 111 all choose this episode in preference to the appearance of the stag at the conclusion of the head-episode; and Add. marks the same place in the text by the unhorsing of Lancelot and Mordret which precedes their seeing the Stag. The White Stag and Lions motif is depicted in in other contexts as well, once in the Estoire and once in the Queste del saint Graal.22 Here it serves as a thread linking the Agravain to these branches and to the Grail theme.23 Among the three Flemish copies of the Lancelot-Grail romance, Add. is distinguished for the attention its illustrations pay to the legal issues raised in the text.24 Here the pictorial emphasis on the explanation of Lancelot’s ancestry shows a similar attitude towards the attendant issues of selfrecognition and social status. The murder in Estoire was seen to be less significant than Lancelot’s discovery, while in Amst./Rylands the emphasis was the opposite, and an explanation is not readily forthcoming. Although nothing specific is known about either patron, it is very likely that an important land-owner or political figure was the first owner of Add. The closely similar images in BNF fr 111 were made in Poitiers c. 1480–85 for Yvon du Fou, sénéchal de Poitiers and grand veneur de France († 1488), clearly an important political figure. Might Yvon du Fou have had access to Add. itself ? The comparison warrants further investigation. The remaining manuscripts are associated with Jacques d’Armagnac, BNF fr 119 being part of the set inherited from Jean de Berry — whose copy also included Lancelot’s discovery of his name on the metal tomb-lid as shown above; so here it is surprising that the single subject chosen is the killing of the lions rather than the discovery of the head and its explanation. It is also notable that the murder in Estoire was not depicted in the first volume of this set, BNF fr 117. The companion volumes, Arsenal 3479–80 also give the single scene of the lion-killing and had no illustration in the Estoire. BNF fr 115, made for Jacques d’Armagnac and illustrated by Evrard d’Espingues also gives the killing of the lions, but pairs that with the final scene of the vision Sommer, Vulgate Version, I, p. 256. 24; Ponceau 501.791.13. Sommer, Vulgate Version, VI p. 166; Pauphilet, La Queste del saint Graal, p. 234. 23 Sommer, Vulgate Version, V p. 280. 24 See n. 19. For Add.’s interest in the legal aspects of the False Guinevere episode see Stones, ‘Illustration et stratégie illustrative’, reprinted in these essays. 21 22

LANCELOT AND IDENTITY

307

of the stag and lions by Lancelot and Mordret, as in Bonn, Yale, BNF fr 344, Add. and BNF fr 111. The murder was given three scenes in the first volume of Jacques d’Armagnac’s set, BNF fr 113, as discussed above, and similar to what is in Amst., so it would seem that the retrieving of the head and body and the explanation was considered less important than the murder itself, perhaps a reflection of Armagnac’s own penchant for action rather than words, which eventually resulted in his downfall. No further volumes from Yvon du Fou’s copy have survived so a comparison with the murder is not possible. The story of King Lancelot and its discovery was treated in a variety of ways, rarely consistently, in that most of the illustrated manuscripts which include pictures of the episodes make different selections and give varying weight to the Estoire and the Agravain components of the narrative, while the intermediary moment in the first part of the Lancelot is rarely illustrated at all. These differences point towards the illustrative tradition as something dynamic and changing, calling in question the very notion of a stable ‘illustrative tradition’ at all. Rather, individual choices and preferences seem to be what governed the presence or absence of illustrations, reflecting differing perceptions of which components were considered worth depicting and what should be left as text alone. We are still very much in the dark as to who exactly made these decisions but the examples considered here point in directions that suggest a strong participation on the part of the patrons. Future studies should further clarify how the process worked and who was responsible for what.

308

APPENDIX A The murder of King Lancelot Manuscripts Bonn, Universitäts-und Landesbibliothek 526, written in Amiens in 1286 by Arnulfus de Kayo (Bonn) Paris, Bibliothèque nationale de France, français 344, made in Metz (c. 1290 ?) (BNF fr 344) olim Amsterdam, Bibliotheca Philosophica Hermetica 1 (sold at Sotheby’s 7.xii.2010, lot 33), made in Thérouanne or Saint-Omer c. 1315 (Amst.) Paris, Bibliothèque nationale de France, français 105, illustrated in Paris by the Sub-Fauvel Master (c. 1320–40) (BNF fr 105) Paris, Bibliothèque nationale de France, français 9123, illustrated in Paris by the Sub-Fauvel Master (c. 1320–40) (BNF fr 9123) Paris, Bibliothèque nationale de France, français 113, illustrated by Evrard d’Espingues for Jacques d’Armagnac († 1477), c. 1475 (BNF fr 113) Text Editions Sommer, H.O. The Vulgate Version of the Arthurian Romances, 7 vols., Washington DC, 1909–1913 (S V) Ponceau, J.-P. L’Estoire del saint Graal, 2 vols., Paris, 1997 (Pon) Poirion, D. et al., Lancelot, 3 vols., Paris, 2001, 2003, 2009 (Poi III) in the transcriptions below, proper names are capitalized

S I 291. 21 Pon 568.894.1

S I 291.7, Pon 567. 893.1, Poi I 557. 605

Bonn 526 f. 58v Rubric: Mais or se taist li contes a parler d’Alain et retourne a parler d’un duc qui cope le chief Lanselot le roi et uola en en une fontainne. The duke of Bele Garde has beheaded King Lancelot with his sword, the head now lying in the fountain. Text: Or dist li contes que quant Josephés fu trespasses de cest siecle... BNF fr. 344 f. 80-2 No rubric O, The duke who cut off King Lancelot’s head tries to lift it out of the boiling fountain. Text: Or dist li contes que quant Celidoine ce fu partis de son peire...

BNF fr. 105 f. 124-2 Rubric: Comment li sires du chastel de Bele Garde copa le chief dessus vne fontaine. au roy Lancelot et Comment les murs de son chastel cheirent sus lui quant il y fu reuenuz. No miniature. Text: Or dist li contes qiue quant Celydoines se fu partiz de son pere...

BNF fr. 9123 f. 94-2 Rubric: Comment .ij. lyons gardoient la tombe au roy Lancelot si que nulz ni osoit touchier Two lions stand behind tomb with nonsense inscription on lid. Text: Or dist li contes qiue quant Celydoines se fu partiz de son pere...

LANCELOT AND IDENTITY

309

1

Royal f. 88v-2 Space for rubric left blank His castle falls on the duke who killed King Lancelot. E, penflourished initial Ensi uenga nostre sires le roy Lancelot del duc...

S I 295. 25 Pon 575. 904. 16

BNF fr. 113 f. 116-1 Rubric: Comment le duc de Bellegarde coppa la teste au roy Lancelot. The duke raises his sword to finish cutting off King Lancelot’s head as he leans over the fountain. Text: Quant il vit la teste qui gisoite en la fontainne...

BNF fr. 113 f. 116-2 Rubric: Comment le duc fut occis par son chastel qui sur lui tumba ou il y avoit tenebres. The tomb of King Lancelot with his recumbant effigy (and no blood). Text: Ensi vengea nostre seigneur le roy Lancelot du duc...

Amst. i f.117v Rubric: Chi caupa li dus la teste au roy Lancelot le pere Lancelot du Lac 1 The duke who cut off King Lancelot’s head tries to lift it out of the boiling fountain. Q, one-line ink capital Text: Quant il vit la teste qui gisoit en la fontaine... Amst. i f.118r Rubric: Chi fondi li chastiaus au duch sour sa teste et la crauenta et ses gens. His castle falls on the duke who killed King Lancelot. E, one-line ink capital Text: Ensi uenga nostre sires le roi Lancelot del duc...

It is Lancelot’s grandfather King Lancelot, not his father, King Ban, who is killed by the Duke; and the rubric refers to the previous text passage while the picture shows what immediately follows in the text.

Text ends incomplete at S I 295.3 (P 575. 905. 5)

Royal f. 88v Space for rubric left blank The duke who cut off King Lancelot’s head tries to lift it out of the boiling fountain, watched by his knights. Q, penflourished initial Text: Qvant il vit la teste qui gisoit en la fontaine...

S I 295. 3 Pon 574. 903. 13

310

last leaf omitted

S I 296. 4 Pon 576. 906. 11

BNF fr. 344 f. 81 Estoire ends complete: Si se test ore li contes a tant de totes les ligniees qui de celidoinee issirent et se torne ai vne autre brainche que len apelle lestoire de merlin quil couient a aioster ensemble a fine force auec lestoire del saint Graal por ce que branche en est et a partient. Et comence mes sire Robers de borron cele branche en tel maniere. Miniature follows: two affronted lions on hind legs, fighting; two crouching lions licking tomb. f. 81v miniature: seated hooded figure holding book, looking at head in boiling fountain, lion licking tomb, a second lion licking a man lying on the ground, and a crowd of men watching.

last leaf omitted

S I 295. 34 Pon 576. 906. 1+var

Amst. i f.118r-3 Rubric: Chi lequerent li doy lyon le tombe le roy Lancelot et puis le garderent tant que Lancelot du Lac les tua. Two lions guard King Lancelot’s tomb and lick it. S, one-line ink capital. Text: Si auint si biele auenture...

Amst. i f.118r-2 Rubric: Chi sainia par miracle le tombe le roy Lancelot. The tomb of King Lancelot weeps blood. T, one-line ink capital Text: Tant ka. J. iour aujnt ke par deuant la tombe passoit vns lyons...

BNF fr. 9123 f. 95v Estoire ends complete Moult fu iriez li anemis Explicit le liure du saint graal...

BNF fr. 113 f. 116v Rubric: Comment les .ii. lions qui estoient blecez guerirent a la tombe du roy Lancelot et la garderent tant que Lancelot du lac y vient. Two lions guard the tomb of King Lancelot and one of them licks it. Text: Et advint si bell aven/ture qui fit...

LANCELOT AND IDENTITY

311

312

APPENDIX B Lancelot du Lac finds the head and body of his grandfather King Lancelot Manuscripts Bonn, Universitäts- und Landesbibliothek 526, written in Amiens in 1286 by Arnulfus de Kayo (Bonn) Paris, Bibliothèque nationale de France, français 344, made in Metz (c. 1290 ?) (BNF fr 344) New Haven, Yale University, Beinecke Library 229, made in Thérouanne or Saint-Omer c. 1290–1300 (Yale) London, British Library Additional 10293, written in 1317 ns in Thérouanne or Saint-Omer (Add.) Manchester, The John Rylands University Library French 1, made in Thérouanne or Saint-Omer c. 1315 (Ryl.) Paris, Bibliothèque nationale de France, français 119, made in Paris before 1407 (BNF fr. 119), illustrations attributed to the Master of Berry’s Cleres femmes Paris, Bbliothèque de l’Arsenal 3480, made in Paris before 1407 (Arsenal 3480) Paris, Bibliothèque nationale de France, français 115, illustrated by Evrard d’Espingues for Jacques d’Armagnac († 1477), c. 1475 (BNF fr. 115) Paris, Bibliothèque nationale de France, français 111, made c. 1480–85 in Poitiers for Yvon du Fou, sénéchal de Poitiers, grand veneur de France († 1488), illustrations attributed to the artist of BNF fr. 22500 (see Banque d’images) Text Editions Sommer, H.O. The Vulgate Version of the Arthurian Romances, 7 vols., Washington DC, 1909–1913 (S V) Micha, André. Lancelot, 9 vols., Paris and Geneva, 1978–1993 (LM V) Poirion, D. et al., Lancelot, 3 vols., Paris, 2001, 2003, 2009 (Poi III) in the transcriptions below, proper names have been capitalized and apostrophes added

BNF fr. 344 f. 427v-2 No rubric O, Lancelot meets a dwarf in the Perilous Forest. Text: Or dist li contes car quant lancelot fu entres en la forest perilleuse u il ot cheuachie tant que li soleuz fu leuez...

Add. f. 334v-2 Rubric: Ensi ke Lancelot arme chevauche en .i. forest si encontre .i. nain sor .i. noir palefroi si parla a li. Lancelot meets a dwarf in the Perilous Forest. Text: Che dist li contes que quant Lancelot se fu partis de ses compaignons et sez chevaus ot un poi mangie de l’erbe si comme li contes la devise. qu’il monta et entra en la Forest Perilleuse et puis qu’il y fu entres si chevaucha tant que li solaus fu leves...

Ryl. f. 77r No rubric Q, champie initial Text (+ var): Quant Lancelot fu entres en la Forest Perilleuse si erra tant que li solaus fu leves...

Yale f. 110v-2 No rubric. Lancelot (surcoat and housing argent [white]), riding in the Perilous Forest, meets a dwarf on horseback and having dismounted, fights the two lions who guard the tomb, its lid covered with drops of blood, that stands between the pine trees by the little house and fountain; Lancelot lifts the head out of the boiling water of the fountain and shows it to the hermit; Lancelot raises the lid of the tomb to reveal the headless shrouded body inside; bottom border: a standing woman places a wreath on the head of a kneeling man (Hand 2). Text: Or dist li contes que quant Lancelot fu entres en la forest perilleuse et il eut tant cheuaucie que la nus fu passee et li solaus leues.

SV 244. 13 LM V 117 Poi III 514. 467

Bonn f. 391v Rubric: Conment lanselos se combat a .ij. lyons deuant une tombe et les ocist. Lancelot on horseback kills with his sword the two lions who guard the tomb of his grandfather. Text: Or dist li contes ue quant Lanselos fu entres en la Forest Perillouse qui’il chevaucha tant que li solaus fu leves...

Ryl. f. 76v Rubric: Chi achieva Lancelot l’aventure de la teste son daijon [sic] qui estoit en la fontaine qui bouloit et .y. tua .ii. lyons. f. 77 Lancelot holds the head of his grandfather, which he has taken from the boiling fountain, and presents it to the hermit. Text: Or dist li contes ke quant li compaignon du tertre se furent au matin esvillie...

SV 243. 20 LM V 116 Poi III 514. 466

BNF fr. 115 f. 508 Rubric: Comment messire Lancelot partit du tertre desuce et sen alla tuer deux lions qui gardoient le corps son aieus et getta la test hors de la fontaine Lancelot kills the two lions who guard the tomb of his grandfather. Text: Or dit le compte que quant les compainons du tertre se furent au matin esueilles... BNF fr. 111 f 204-2 No rubric Having killed the lions and opened the tomb with the headless body, Lancelot hands the head of his ancestor to the hermit. Text: Or dit le conte que quant Lancelot fut entre en la fourest perilleuse il cheuaucha tant qu ele souleil fu leue...

BNF fr. 111 f. 204 Rubric: Cy parle des compaignons du tertre. Lancelot, watched by his squire, kills the lions who guard the tomb of his grandfather. Text: Or dit le comte que quant les compaignons du tertre furent esveillez... LANCELOT AND IDENTITY

313

Add. f. 335r Rubric: Ensi ke Lancelot arme a piet a ochis .ii. hommes d’enkoste .i. tombe ki degoute sanc et deles le tombe a .i. fontaine et .i. teste dedens. (lions, not hommes) Having killed the lions guarding the tomb of his grandfather and retrieved the head from the fountain, lancelot presents the head to the hermit. Text: Quant Lancelot ot ocis lez .ii. lyons... Add. f. 335v Rubric: Ensi com Lancelot est a pie devant .i. hermitage et parole a l’ermite The hermit tells Lancelot how his grandfather died. Text: Ensi monstroit li prodoms a Lancelot que par un mellour chevalier...

S V 245. 6 LM V 120

S V 247. 12 LM V 127

S V 244. 27 LM V 118

BNF fr. 119 f. 440v At a cross, Lancelot riases his sword to kill the two lions who guard the tomb of his grandfather. Rubric: Comment Lancelot tua .ii. lyons qui gardoient le corps son ayeul. Text: Quant Lancelot voit les lyons ...

Arsenal 3480 p. 324 No rubric At a cross, Lancelot raises his sword to kill the two lions (no tomb). Text: Quant Lancelot voyt les lyons...

314

S V 249. 15 LM V 132

SS V 247.19, LM V 127

Yale f. 112 No rubric. O, Lancelot (surcoat azure), his horse beside him, talks to the hermit who stands in the doorway of his hermitage; border, top: bird, bottom border: dragon terminal biting champie initial (Hand 2). Text: Or me dites fait Lancelot que la preude feme deuient... Add. f. 336v Rubric: Ensi que Lancelot armes chevauche apres .i. blanc cherf qui estoit environnes de .iiii. lyons. Lancelot sees the White Stag led by four lions pass by. Text: Ensi delivra Lancelot le vallet del ours qui mangier le voloit...

LANCELOT AND IDENTITY

315

XVI A Note on the Heraldry of a Very Special Gauvain*

T

he arms borne by Arthur’s nephew Gauvain, oldest son of King Lot and brother of Agravain, Guerrehet and Gaheriet, had by the fifteenth century stabilized in the rolls of arms to take the form purpure a double-headed eagle or, beaked and clawed azur.1 By then it had become customary to include King Arthur’s knights as a heraldic category preceding historical personages, a practice found in relation to the warriors of Troy of the late thirteenth century, added, probably in the seventeenth century, to the beginning of the Armorial Le Breton.2 As Pastoureau has noted, by the fifteenth century King Lot was considered long dead, so Gauvain was the head of the lineage and his younger brothers differenced their arms in relation to his. But in earlier periods another choice altogether was made for Gauvain, explained primarily as a difference in relation to the arms of Lot, who bore plain arms of argent.3 So, as his son, Gauvain differenced his * It is a pleasure to offer this note as a tribute to Keith, a long-term friend and collaborator from whom I have learned so much. First published in “Li premerains vers”: Essays in Honor of Keith Busby, ed. C.M. Jones and L.E. Whalen (Amsterdam and New York: Rodopi, 2011), pp. 433–47. 1 Pastoureau, Armorial de la Table Ronde, no. 83. 2 Paris, Archives nationales, AE I 25, no. 6 (MM 648), see L’armorial Le Breton, p.14, and Vielliard, “L’armorial des fils de Priam,’’ ibid. pp. 37–39. The Troy arms are on pp. 2–3 of the manuscript, which is paginated, and form shields nos. IB to 43B. The armorial is named after its seventeenth-century owner, Hector le Breton (d. 1652), who was responsible for the present binding. The armorial is composed of several parts of which the earliest (pp.18–46, nos. 136–715) dates to c. 1292–1294 (de Boos p. 16) and the Troy component is approximately contemporary; the rest of the armorial consists of two fifteenth-century components. 3 Pastoureau, “Étude d’héraldique arthurienne: les armoiries de Gauvain,’’ p. 6.

A VERY SPECIAL GAUVAIN

317

arms in relation to his father, and was commonly depicted bearing argent a canton gules, usually shown with white rather than silver or tin used for argent. The arms of the Knights of the Round Table, found mostly in the context of the Lancelot-Grail romance, were rarely stable in manuscript depictions before c. 1300, and even thereafter, surprising variants occur.4 Whereas certain shields, like those of Galaad (argent a cross gules) and Lancelot (argent 3 bends gules), are specified in the text of the Lancelot-Grail romance, the shield of Gauvain was not described and its origins are obscure.5 At the same time the arms argent a canton gules were not always reserved for Gauvain. Sometimes they were borne by other knights, and on occasion they were even given to knights or soldiers in biblical illustration. Here I outline some of the variants found in relation to the arms of Gauvain and identify a set of manuscripts in which the arms borne by Gauvain are particularly telling, pointing perhaps to a specific individual who chose to have himself depicted as Gauvain. Where does argent a canton gules first occur? A surprising instance, perhaps the earliest, is found in the Epistolary written in 1266 by Johannes Phylomena for the Bishop of Cambrai, Nicolas de Fontaines, Cambrai Méd. mun. 190, f. 71 (fig. 1).6 The arms that will come to be Gauvain’s are borne there by one of the three knights shown sleeping at the sepulcher while Christ rises from the dead, in the historiated initial R for High Mass on Easter Sunday.7 What determined this choice? The connotation is not entirely a positive one since the soldiers are sleeping rather than seeing the Resurrection, which of course the three Maries, first visitors to the sepulcher, learn about not from the soldiers but from an angel. One might

4 I outline some of these variants in Stones, “Les débuts de l’héraldique”, reprinted in these essays. For the use of heraldry in the illustrations of the manuscripts of Chrétien de Troyes, see Pastoureau, “Les armoiries arthuriennes”. 5 Even the noted Arthurian heraldrist Brault found no explanation: “A satisfactory explanation for Gawain’s arms cannot be given at the present time” (Early Blazon, p. 42). 6 These manuscripts were first brought to prominence by Beer, “Das Scriptorium des Johannes Philomena”. All the illustrated pages are now reproduced on Enluminures. 7 The same subject is also found in Toledo, Archivo de la Catedral 56. 19, the Pontifical of Cambrai made for Nicolas’ successor Enguerrand de Créquy c. 1274, on f. 24. Here, however, the (two) soldiers’ shields are positioned so as to show the brown insides not the outsides of the shields; the comparison underlines how deliberate was the choice made in Cambrai 190’s version.

318

surmise that a local family may be alluded to, but likely candidates are not forthcoming.8 In the context of Arthurian heraldry, the earliest occurrence of argent a canton gules that I know is in BnF, fr 770, borne by Gauvain on f. 169, in the Suite Vulgate,9 Gauvain is depicted in numerous miniatures in BnF, fr 770 but this is the only one where he is shown bearing these arms (fig. 2). Illustrations by the artist of BnF, fr 770 reappear in two other Arthurian manuscripts, Le Mans MM 354 (Estoire del saint Graal, written by Walterus de Cayo) and Oxford, Bodl. Digby 223 (Agravain incomplete, Queste, Mort Artu). There is no firm date for the group, but a reasonable range is between 1274, when the somewhat related BnF, fr 342 was written (by a female scribe), and 1282, when one Wautiers dou Kai (the same person as the scribe in Le Mans 354?) wrote a copy of the Image du monde by Gossouin de Metz, BnF, fr 14962. Gauvain does not appear in the Estoire, so there is no occasion for depicting him with arms in Le Mans 354. Digby 223 is sparsely illustrated, most of its narrative switches marked not by an illustration but by a decorative initial, but on f. 94v Gauvain is depicted holding a shield argent a canton gules, standing over Meliant who lies ill in bed in the Queste del saint Graal.10 This subject also occurs in Manchester, The John Rylands University Library, MS French 1, f. 193v as an added miniature, but there Gauvain has no shield; the two related copies made in the early fourteenth century, London, BL Add. 10294 and Royal 14 E.III give different subjects: Gauvain, Gaheriet and Yvain riding to the Castle of Maidens (Add. f. 10v) or killing the seven knights at the Castle of Maidens (Roy. f. 9v)—and in both those miniatures Gauvain bears a shield argent a canton gules.11 It is in the 8 Pastoureau (“Armoiries de Gauvain” p. 9) notes that according to the fourteenthcentury Armorial du héraut Navarre, the Croquoison family bore argent a canton gules; see Douët d’Arcq, Armorial de France, p. 228, no. 1047 (Paris, BnF, fr 14356, f. 82). But the Croquoison were an Artesian family, with land holdings in the regions of Montreuil-sur-Mer and Hesdin, not Cambrai in the County of Hainaut. The Croquoison family may be more relevant to BnF, fr 770, an Artesian manuscript which I discuss below. 9 This occurs at II 134.28 in Sommer’s edition (hereafter S). It is based on London, BL Additional 10292–4. The corresponding miniature in Add. 10292 is on f. 115, illustrated by the second painter; the crowd of knights have ailettes painted plain orange, or white with a bend, cross, or saltire gules. In general this artist is not interested in consistent heraldry, see further below. 10 This occurs at S VI 37.25 and at 51.17 in La Queste, ed. Pauphilet. First published in 1923, the earlier editions are differently paginated. 11 Add. = London, BL Additional 10292–4, containing the Lancelot-Grail cycle complete in 3 volumes, one of which has the date 17 February 1316 (1317 new style) carved on a tomb;

A VERY SPECIAL GAUVAIN

319

work of the major painter of the three copies Add. 10292–4, Royal 14.E.III and Amsterdam, Bibl. Philosophica Hermetica 1/Oxford, Bodl. Douce 215/ Rylands French 1, that consistent heraldry for Arthur’s knights makes its next appearance, following on the use of argent a canton gules for Gauvain in BnF, fr 770 and Digby 223.12 The second artist in Add. 10292 however shows no interest in consistent heraldry and Gauvain’s arms are forgotten altogether.13 Even in the work of the major painter of the group there are heraldic surprises. In the scene in the Queste of the three Grail knights and Perceval’s sister discovering Solomon’s enchanted ship and its contents, Douce f. 35 shows Galaad bearing the arms argent a canton gules—arms otherwise reserved for Gauvain who of course is not one of the chosen Grail Winners (fig. 3). Does the choice of heraldry suggest the wish that Gauvain had been present, or was it simply an error? The absence of Perceval’s sister in the Douce miniature suggests the latter, since it was she who explained the meaning of the colored spindles, and the crown and sword, even before the knights discover the written explanation in the purse; and she is present in the comparable miniature in Add. and Royal (figs. 4 and 5). Against this background, Gauvain’s heraldry in another cluster of Lancelot-Grail volumes stands out as very different. Three incomplete sets of Lancelot-Grail volumes from the same workshop have recently been convincingly re-aligned thanks to new discoveries in Bologna and Turin. Two of them have long been known: Paris, BnF, fr 749 (Estoire, Merlin, Suite Rylands = Manchester, The John Rylands University Library, MS French 1. Together with olim Amsterdam, Bibliotheca Philosophica Hermetica MS 1 and Oxford, Bodleian Library MS Douce 215, Rylands Fr. 1 forms a second set, with many lacunae. Royal 14 E.III in the British Library contains the Estoire, Queste, and Mort Artu. All three copies were made by a team of scribes, decorators and illustrators working in Saint-Omer, Thérouanne, or Tournai in the early fourteenth century. They are the pilot project manuscripts of the collaborative Lancelot-Grail Project, chosen because Add. 10292–4 contains more illustrations (748) than any other surviving copy, and because it was Sommer’s base manuscript. For the LancelotGrail Project, see http://www.lancelot-project.html. 12 Other illustrated manuscripts of the first half of the thirteenth century are generally unsystematic in their use of heraldry for Arthur’s knights. Even in the second half of the century, important manuscripts like Bonn, LUB 526, written by Arnulfus de Kayo (a relative of Walterus?) in 1286, include a great deal of fictitious heraldry, but nowhere a shield argent a canton gules. The same observation holds good for the pair of volumes, BnF, fr 95 and New Haven, Yale University, Beinecke Library, 229, made in the region of Thérouanne in the decade of the 1290s. 13 See also n. 9. Add.’s second artist is known elsewhere for his work on the Vincent of Beauvais Speculum historiale, Boulogne-sur-Mer BM 130, dating after 1297 because it is a

320

1. Cambrai, Médiathèque municipale 190, f. 71 (photo: author)

2. Paris, Bibliothèque nationale de France, fr 770, f. 169 (photo: Bibliothèque nationale de France)

A VERY SPECIAL GAUVAIN

3. Oxford, Bodleian Library, Douce 215, f. 35 (photo: reproduced by permission of The Bodleian Library, University of Oxford) (photo: author)

4. London, British Library, Additional 10294, f. 44 (photo: Trustees of the British Library Board)

321

322

5. London, British Library, Royal 14 E.III, f. 130v (photo: Trustees of the British Library Board)

7. Oxford, Bodleian Library, Ashmole 828, f. 63v (photo: reproduced by permission of The Bodleian Library, University of Oxford

A VERY SPECIAL GAUVAIN

6. Paris, Bibliothèque nationale de France, fr 749, f. 311 (photo: Bibliothèque nationale de France)

323

324

Vulgate) and Oxford, Bodl. Ashmole 828 (Lancelot, incomplete). They are stylistically related to several other clusters of secular manuscripts — the four prose Alexander manuscripts, Berlin, Staatliche Museen Preussischer Kulturbesitz, Kupferstichkabinett 78.C.1; Brussels, BR 11040, London, BL Harley 4947 and a manuscript in private hands;14 four copies of Brunetto Latini’s Trésor, London, BL Yates Thompson 19; St Petersburg, National Library of Russia Fr. F.v. III. 4; Paris, BnF, fr 567; and Florence, Bibl. Mediceo-Laurenziana Ashburhman 125;15 and several liturgical and devotional books: the Missal of Saint-Nicaise de Reims, St Petersburg NLR Lat. Q. v. 1.78, made before 1297 since the feast of St Louis (Louis IX), canonized 1297, is absent;16 the Bible, Reims, BM 39–42;17 the Hours of uncertain Use, Baltimore, Walters Art Museum, W. 98, whose calendar (also lacking St Louis), is Stockholm, Nationalmuseum B.1648,18 and the Psalter perhaps for the Use of the Augustinian house at Ypres before 1297, BnF, Smith-Lesouëf 20.19 These devotional and liturgical books help to place the secular manuscripts because of the indications of provenance and liturgical direct copy of MS Boulogne 131 which was written in that year; and the Psalter-Hours of Saint-Omer Use, BL Add. 36684 and New York, Pierpont Morgan Library M.754, dating after 1318 because it includes reference to an indulgence granted by Pope John XXII in that year. In none of these manuscripts is the artist concerned to depict accurate heraldry. 14 See Stones, Illustrations of Lancelot, pp. 182–207, 447–48. For the Berlin, Brussels, and Harley Alexander manuscripts see ead., “Notes on three illustrated Alexander manuscripts,“ pl. 9, f. 6v. For the manuscript in private hands, see Stones and Ross, “The Roman d’Alexandre in French Prose”. For the Berlin manuscript see also Rieger, ed. L’Ystoire du bon roi Alexandre, Appendix by A. Stones, pp. 247–59; for the Brussels manuscript see Rieger, Historia de Alejandro Magno. For a comprehensive treatment of the prose Alexander see now PérezSimon, Mise en Roman, mise en image. 15 Stones, Illustrations of Lancelot, pp. 184, 187, 190–96, 443–44; Roux, L’iconographie du Livre du trésor and ead., Mondes en Miniatures; Stones, “A Note on the North French Manuscripts of Brunetto Latini’s Trésor”; Hunt, Illuminating the Borders, pp. 40, 114–19. For the St Petersburg copy see the facsimile: Li livres dou Tresor. 16 See particularly Mokretsova and Romanova, Les Manuscrits enluminés français du XIIIe siècle, pp. 194–231; Avril in Rois maudits, p. 320, no. 218. 17 Ibid., no. 219. It contains the arms of Toucy-Châtillon for Jeanne de Toucy Châtillon, wife of Thibaud de Bar, who in ber widowhood donated a window to Saint-Nicaise depicting herself and her children. 18 Randall, Medieval and Renaissance Manuscripts, France, pp.119–23, no. 49. The calendar contains the added obit of Henri de Bayon, a nobleman of Champagne who died in 1334. 19 Leroquais, Psautiers, II, p. 326, no. 496; but I owe the Ypres suggestion to the late Christopher Hohler.

A VERY SPECIAL GAUVAIN

325

use that they contain, so that the origins of the group somewhere in the region of Flanders or Champagne may be posited, and a date before 1297 is most likely. Differences in size and script between BnF, fr 749 and Ash. 828, noted by Martine Meuwese, suggest they were actually parts of two different sets of volumes.20 She relates the two bifolia from an Estoire, Bologna, Archivio di Stato, b.l.bis,21 to Ash. 828, while the other newly discovered fragment, Turin, BN L.III.12 (Merlin and Suite Vulgate),22 and BnF, fr 749 (Estoire, Merlin, Suite Vulgate) remain two closely related but distinct copies of parts of the cycle. As indicated above, the idea of producing multiple illustrated copies of the same text is a feature characteristic of the “workshop” (however defined) that made these books, and a strong interest in heraldry characterizes the illustrations, particularly of the Alexander manuscripts (Stones, “Four Illustrated” Table, n. 14), and to a lesser degree the Trésor manuscripts, notably in relation to the arms of Charlemagne (Stones, “Trésor” Tables, n. 15). There is a striking dissimilarity, however: the four Alexander manuscripts each differ in their choice of shield given to the eponymous hero, whereas the Lancelot-Grail manuscripts as a group are consistent in their use of heraldry for Arthur and his knights. King Arthur’s arms in BnF, fr 749 are or 3 crowns gules (Brault, Early Blazon pp. 44, 176)—a variant on azure 3 crowns or which emerges for Arthur in the Add. / Royal / Amsterdam / Douce / Rylands group in the early fourteenth century — and Meuwese has drawn attention to the presence of or 3 crowns gules for Arthur in the Turin fragment as well (“Crossing Borders” 175).23 What, then, of Gauvain in these manuscripts? Far from depicting him with argent a canton gules, he is shown with a considerable degree of consistency bearing something quite different: azure a lion or crowned argent [white], a fess gules overall (figs. 6 [best seen on Bodley’s Luna web site], 7). Where do these arms come from? The answer is far from obvious. The fess gules suggests that Gauvain’s arms are a difference of azure a lion or, made to distinguish him, or the person whose arms they represent, from a father figure, just as the canton gules distinguished Gauvain from Lot’s plain argent—although in these manuscripts, Lot bears not plain argent, but Meuwese, “De omzwervingen”; and ead., “Crossing Borders“. Longobardi, “Ancore nove frammenti della Vulgata”. 22 Castronovo, La biblioteca dei Conti dì Savoia, p. 189. 23 Meuwese also notes the use of or three crowns gules for Arthur in the Middle Dutch Spiegel Historiael of Jacob van Maerlant, The Hague, KB KA XX (“Crossing Borders” p. 176). 20 21

326

argent a chevron gules, so there is no effort to link the arms of father and son, and another explanation is called for. Azure a lion or is less commonly found in the armorials and other manuscript sources than azure a lion argent, which can be traced in a number of contemporary examples. For instance they are borne in the Armorial Le Breton by Enguerran, seigneur de Fiefes (no. 504), and again, with a difference of a label of three points gules, by an unidentified knight (no. 528). Azure a lion or and azure a lion argent both with a bordure of the same, a bend gules overall, the one possibly a difference of the other, occur in the Hannover copy of the Somme le Roi, Gottfried Wilhelm Leibniz Bibliothek-Niedersächische Landesbibl. MS I 82, written in French and Occitan in the second half of the thirteenth century, perhaps in central France.24 Promising families in the Auvergne whose arms are azure a lion or are those of Saint-Floret and Montaigu;25 while azure semé of lions argent each surmounted by a fleur de lis gules is found in the Antiphoner of Beaupré (O. Cist., Dioc. Cambrai), where it has been associated with a knight of the Viane family.26 Neither quite corresponds to what is borne by Gauvain. At present the solution has not come to light. The recurrence of the same arms in more than a single manuscript suggests that several members of a commissioning family might have requested them, or that they were in some way representative of the interests of the “workshop.” If the answer is not immediately forthcoming, the presence of this special Gauvain invites further investigation of what other features of iconography and heraldry these triplet Lancelot-Grail manuscripts offer and how they differ from the rest of the Lancelot-Grail tradition. Perhaps such investigation will eventually reveal whom it was who chose to single 24 They are not mentioned in Handschriften der Niedersächsischen Landesbibliothek Hannover, I, pp. 96–103, color pl. of ff. 9v, 67v, 86v; the shields are noted but not identified by Kosmer, Style and Iconography of a Thirteenth-Century Somme le roi, I, p. 302, II, pp. 20–22, nor by Rouse and Rouse, Manuscripts and their Makers, I, pp. 146, 149, 151, 159–61, 165, 364, n. 13 and fig. 18; II, App. 6B, ills. 82, 89, 90 (ff. 12v, 86v, 92). For the text and manuscripts of the Somme le roi, see Brayer and Leurquin-Labie, Somme le Roi. 25 de Remacle, Dictonnaire généalogique, II, pp. 571–73, III, pp. 282–84, kindly drawn to my attention by Hélène Loyau. 26 Randall, Medieval and Renaissance Manuscripts, Belgium, pp. 55–56, Cat. 219 A–D; ead., “Three Newly Acquired Illuminated initials”. But the shield is best matched by that of Gerard IV of Zottegem, married to an unnamed lady of the Viane family at an uncertain date (Warlop, Flemish Nobility, part II, nos. 236/39). His shield on his seal of 1226 (Warlop, pl. 246) is billety a lion with a fleur de lis on its shoulder. This date is too early for books written in 1289 and 1290 and further investigations are called for.

A VERY SPECIAL GAUVAIN

327

out Gauvain for such distinctive visual attention. For the moment at least I hope to have drawn attention to his startling presence in these books and to invite others to join the search for his alter ego.

XVII Le merveilleux dans le Lancelot-Graal: l’exemple du cerf accompagné de quatre lions

L

e motif merveilleux par excellence dans le roman du Lancelot-Graal est le saint Graal, objet central dans l’Estoire comme dans la Queste. Mais la tradition iconographique de ces branches admet une certaine réticence à ce sujet; nombreux sont les manuscrits qui n’incluent pas d’image du saint vessel, objet trop vénérable pour être représenté, alors que d’autres manuscrits le représentent sous diverses formes, soit une écuelle suivant le texte de l’Estoire, ou bien comme un objet tiré de la liturgique chrétienne — un ciboire, une calice, parfois une calice contenant une croix ou le corps du Christ enfant, là encore suivant le texte.1 Ici nous mettons en avant un autre sujet merveilleux dont la tradition iconographique est tout aussi variable, présent seulement dans certains manuscrits, absent dans la plupart: le cerf accompagné de quatre lions qui, cette fois, se retrouve dans trois des cinq branches du Lancelot-Graal: l’Estoire del saint Graal, le Lancelot (Agravain), la Queste del saint Graal. Des inversions du motif du cerf se présentent dans le Merlin: Merlin déguisé en cerf devant Jules César;2 dans le Lancelot, Cet article parut dans Le merveilleux au Moyen Âge, ed. A. Latimier-Ionoff, J. PavlevskiMalingre, A. Servier (RILMA Études 8), Turnhout, 2017, pp. 137–46. 1 Voir A. Stones, « Seeing the Grail », dans The Grail, A Casebook, éd. D. Mahoney, New York, 2000, p. 301–66; M.L. Meuwese, « The Shape of the Grail in Medieval Art », dans N.J. Lacy, éd., The Grail, the Quest, and the World of King Arthur, Cambridge, 2008, p. 13–27. 2 Voir I. Fabry-Tehranchy, « Comment Merlin se mua en guise de cerf: écrire et représenter la métamorphose animale dans les manuscrits enluminés de la Suite Vulgate du Merlin » Texte-Image, Varia, 2, 2010, p. 1–32 (13.7.2015); ead., Texte et images des manuscrits du Merlin et de la Suite Vulgate (XIIIe– XVe siècle): L’Estoire de Merlin ou les Premiers faits du roi Arthur (Texte, Codex et Context 18), Turnhout, 2014, p. 351–378.

LE MERVEILLEUX DANS LE LANCELOT-GRAAL

329

notamment Lancelot tuant un cerf et l’offrant à la Dame du Lac; et la Mort Artu: Lancelot blessé dans la cuisse par un chasseur de cerf; et de nombreux manuscrits incluent une inversion sous la forme de la chasse au cerf dans les marges, sans rapport évident avec le contexte textuel.3 Nous laissons les inversions de côté ici. Comme est le cas du saint Graal dont les origines chrétiennes sont déclinées dans l’Estoire, le cerf jouit depuis l’époque paléochrétienne de symbolique christique.4 Le passage du cerf, orné d’une chaine autour du cou et accompagné de quatre lions, est lié dans l’Estoire à l’apparence du Graal. Josephé, ayant joui d’un repas à la table du Graal en compagnie de ses fidèles, fait détruire par sa prière les idoles du païen Matagran et fait ressusciter le païen Argan qui se convertit. Puis Josephé retrouve ses compagnons au bord de l’eau de Celice et ils se demandent comment ils vont la traverser. A ce moment-là passe le cerf accompagné des lions et entre dans l’eau; Josephé et ses compagnons les suivent en marchant miraculeusement sur l’eau.5 Plus tard, trois des compagnons de Josephé, Alain, Bron et Perron, demandent à leur maître ce que signifient les animaux. L’explication est manifestement christique. Josephé leur répond, « ...li cers, ce savez vos bien, quant il est el point de venir en viellece, se rajovenist en lessant son cuir et son poil en partie et revient de viellece en joence, ce est autresi comme de mort a vie. Tot ausint revint Jesucrist, cil beneoiz Prophetes, cil beneoiz Sires, de mort a vie qant il laissa el travail de la croz le cuir, ce fu la char mortel que il avoit prise el ventre de la beneoite Virge. Et por ce que en cel beneoit Segnor n’ot onques teche de pechié nos apparut il ensemblance de cerf sanz tache; par la blanchor dont il estoit coverz devez vos entendre virginitez, qui dedenz lui fu si hautement herbergiee qe onques en lui ne pot l’en veoir signe ne tache de luxure. Par la chaiene que il avoit entor le col devez vos entendre humilité que l’en vit dedenz lui si naturelment que il fu humilitez meesmes. Par les qatre bestes qui Voir surtout A. Fisch Hartley, « La chasse » dans J. Wirth et al., éds., Les Marges à drôleries des manuscrits gothiques (1250–1350), Genève, 2008, p. 181–206. 4 Dès le Ve siècle le cerf apparaît dans un contexte christique, par exemple dans le mausolée de Galla Placidia à Ravenne où deux cerfs se dirigent vers l’eau pour s’abreuver, rappelant le vers 2 du Psaume 41, « Quemadmodum desiderat cervus ad fontes aquarum, ita desiderat anima mea ad te, Deus ». 5 Sommer I 254–256, Ponceau 498–503, Poirion 494–500. Pour la liste des éditions de texte voir l’Appendice. 3

330

compaignie li fesoient devez vos entendre les qatre evangelistres, les qatre beneürees persones qui en escrit mistrent partie des oeuvres Jesucrist qu’il fist tant come il fu entre nos come terriens hom.»6 L’emplacement précis de l’illustration de cet épisode varie dans les quatre manuscrits qui l’incluent, même dans Londres, BL Add. 10292 (fig. 1) et Royal 14.E. III (fig. 2), attribués au même atelier flamand aux environs de 1310–20, alors que dans le troisième membre de ce group, olim Amsterdam, BPH 1, on trouve un autre sujet — Josephé et l’épée brisé (fig. 3) — placé au même endroit dans le texte que la miniature du cerf et des lions dans Royal.7 Les manuscrits plus tardifs des XIVe et XVe siècles favorisent rarement le cerf et les lions. Dans BnF fr 105 (Paris, deuxième quart du XIVe siècle) l’illustration montre la guérison d’Argan et la traversée de l’eau mais n’inclut pas les animaux (fig. 4). Aucune illustration n’est placée à cet endroit dans l’autre exemplaire de l’Estoire confectionné dans le même atelier, BnF fr 9123, alors que dans la version réalisé pour Jean-Louis de Savoie, évêque de Genève, Bruxelles, BR 9246, datant de 1480–1482, dont les rubriques et le choix des sujets (à deux exceptions près) sont les mêmes que ceux de BnF fr. 105, c’est la vision du cerf et des lions qui est représentée (fig. 5). Ce choix fut sans doute dicté par le commanditaire, un haut dignitaire ecclésiastique, mais qui néanmoins n’a pas mis en avant l’illustration du Graal, laissé sans représentation aucune dans son manuscrit.8 Enfin la version confectionnée

Cité d’après Ponceau p. 507. cf. Sommer I 259, Poirion 502. Voir l’Appendice. Le cerf et lions ne sont pas illustrés dans Rennes BM 255; Le Mans MM 354 et BNF fr. 770 (même atelier); Berkeley UB 106; Bonn UB 526 (copié en 1286), BnF fr. 110 (même atelier); BnF fr. 19162 et 24394 (même atelier); Cologny, Bodmer 147; BnF fr. 95; BnF fr. 749, BnF fr. 344; St Petersbourg NLR Fr.F.v.XV.5; Tours, BM 951; Yale University, Beinecke 227 (copié en 1357), BnF fr. 96; New York, Morgan Lib. M.207; Oxford, Bodl. Douce 178 (Italien), ni dans les manuscrits ne contenant qu’une seule illustration. Dans BnF fr. 105 c’est le passage sur l’eau de Josephé et ses compagnons qui est représenté et non la vision du cerf et des lions, alors que le sujet est absent dans BnF fr. 9123, deuxième exemplaire réalisé dans le même atelier parisien au deuxième quart du XIVe siècle. 8 Pour le texte, édité par Guillaume de la Pierre pour Jean-Louis de Savoie, voir surtout la thèse inédite d’A. Delamarre, soutenue à l’École des chartes en 2003, intitulée Copier au XVe siècle du français déjà ancien : l’exemple de l’Estoire del Saint Graal. Commentaire linguistique et édition partielle du ms. B.R. Brux. 9246. Nous remercions vivement Mlle. Delamarre de nous l’avoir communiquée. Pour les illustrations voir A. Stones, « L’Estoire del saint Graal dans la version adaptée par Guillaume de la Pierre pour Jean-Louis de Savoie, évêque de Genève : sources et traitement pictural », Bulletin de la Société nationale des antiquaires de France (2013), 2015, 109–25. Une liste des 34 manuscrits illustrés de l’Estoire est inclue aux p. 123–125. 6 7

LE MERVEILLEUX DANS LE LANCELOT-GRAAL

331

1. Londres, BL, ms Add. 10292, L’Estoire del saint Graal, f. 66v, Joseph, Josephé et leurs compagnons voient un cerf blanc et quatre lions (photo: British Library).

2. Londres, BL, ms Royal 14 E.III, L’Estoire del saint Graal, f. 78v, Joseph, Josephé et leurs compagnons voient un cerf blanc et quatre lions (photo: British Library).

332

3. olim Amsterdam, BPH ms. 1, i, L’Estoire del saint Graal, f. 104, Josephé tenant l’épée brisée (photo: DIAMM for the Lancelot-Grail Project).

4. Paris, BnF, ms fr 105, L’Estoire del saint Graal, f. 111v, Argan ressuscité par Josephé ; Josephé et ses compagnons marchant sur l’eau (photo: BnF).

LE MERVEILLEUX DANS LE LANCELOT-GRAAL

333

5. Bruxelles, BR, ms. 9246, L’Estoire del saint Graal, f. 164, Un homme sur la rive observe Josephé et ses compagnons marchant sur l’eau, suivant le cerf et les quatre lions (photo: DIAMM, reproduit avec la permission de la BR).

6. Paris, BnF, ms. fr 113, L’Estoire del saint Graal, f. 103, Josephé se prosterne en prière alors que ses compagnons prient debout; à côté d’eux se trouvent le cerf et les quatre lions (photo: BnF).

334

7. Londres, BL, ms. Add. 10293, Agravain, f. 336v, Lancelot voit un cerf et quatre lions passant devant lui (photo: British Library).

8. Manchester, The John Rylands University Library, ms. Fr. 1, Agravain, f. 90, Lancelot et Mordred voient le cerf blanc accompagné de quatre lions (photo: DIAMM for the Lancelot-Grail Project).

LE MERVEILLEUX DANS LE LANCELOT-GRAAL

335

9. Londres, BL, Add. ms. 10293, Agravain, f. 347, Lancelot et Mordred désarçonnés par deux chevaliers (photo: British Library).

10. Bonn, Landes-undUniversitäts bibl., ms. 526, Agravain, f. 386v/397v, Lancelot et Mordred, chevauchant, suivent le cerf blanc et trois lions (photo: Bonn, LUB).

336

11. Oxford, Bodleian Library, ms. Rawlinson Q.b.6, Agravain, f. 274v, Lancelot et Mordred chevauchent, suivant le cerf et les quatre lions (photo: Bodleian Library).

12. Londres, BL, ms. Add. 10294, La Queste del saint Graal, f. 45v, Perceval et Galaad voient le cerf blanc accompagné de quatre lions (photo: British Library).

LE MERVEILLEUX DANS LE LANCELOT-GRAAL

337

pour Jacques d’Armagnac vers 1475, BnF fr. 113, inclut la vision du cerf et des lions (fig. 6); comme l’a démontré Jean-Paul Ponceau, cet exemplaire est très proche de Londres, BL Royal 14.E.III et du manuscrit naguère dans la BPH d’Amsterdam,9 donc il est d’autant plus remarquable que ce dernier n’inclut pas l’image du cerf. Le cerf et ses lions font apparence aussi dans l’Agravain, la dernière partie de la branche du Lancelot. Là encore les exemples illustrés sont rares — seulement, encore, quatre exemples — et les protagonistes, Lancelot et Mordred, surprenants.10 Cette fois, à la différence de Josephé, évêque, et ses fidèles, chrétiens, les protagonistes se montrent indignes d’en savoir plus. L’épisode est présenté comme précurseur de l’arrivée de Galaad et l’achèvement de la Queste del saint Graal qui suivra dans la quatrième branche du cycle. Lancelot voit le cerf et les lions à deux reprises: la première fois il est accompagné par un vallet; il fait nuit et ils cherchent à se loger; toutefois ils reconnaissent l’événement comme un phénomène extraordinaire, « ...Certes fait il [Lancelot] ore ai iou ueu la greignor meruelle que iou onques mais ueisse...».11 Lancelot se demande s’il s’agit du commandemant de dieu ou d’enchantement, « ...quar sans faille ie quide bien que sans vertu ou dencantement nest ce mie que li lyons ait plus sens en lui que nature ne li aporte. Et pour ce sai iou bien quil font chou par le commandement de dieu ou par encantement ... cest vne cose dont iou ne serai iamais a aise deuant que iou le sache...».12 Mais il faudra attendre une deuxième apparence des animaux symboliques avant qu’il puisse trouver quelqu’un à qui il peut en demander l’explication. Il fait encore nuit et Lancelot est accompagné cette fois par Mordred, fils illégitime du roi Artur, qui dans la dernière branche, la Mort Artu, deviendra l’ennemi de Lancelot et sera lui-même tué par son père; l’apparence du cerf et des lions sert donc à mettre en avant le défaillance et

J.-P. Ponceau, Etude de la tradution manuscrite de l’Estoire del Saint Graal, roman du XIIIe siècle (thèse de doctorat, Universitee de Paris IV–Sorbonne, 1983, révisée 1986), p. 142– 147; pour les illustrations, voir surtout S.A. Blackman, « A Pictorial Synopsis of Arthurian Episodes for Jacques d’Armagnac, Duke of Nemours », in K. Busby, éd., Word and Image in Arthurian Literature, New York and London, 1996, p. 3–49 (esp. 6 et n. 10). 10 On ne peut que faire écho au sentiment exprimé dans Poirion et al., p. 1543 paragraphe 523 concernant la participation de Mordred, « promis au mal » à cette vision « qui ne lui servira aucunement ». Toutefois Mordred n’était pas présent lors de la première rencontre de Lancelot et le cerf. 11 Sommer, V 249. 29–30. 12 Sommer, V 249. 32–38. 9

338

l’éventuel échec de Lancelot et le mauvais caractère de Mordred. Lorsque Lancelot cherche l’explication du mystère auprès d’un hermite, sa demande est refusée, « ...Or sachies fait li hermites que cest vne des gregnors meruelles que vous onques ueissies. Ne ce nest mie cose que vous puissiez sauoir ne achieuer ne vous ne homme nuls. fors seulement li boins cheualiers qui de bonte et de cheualerie passera tous lez terriens cheualiers et cils achieuera lauenture des lyons et du cerf et fera a sauoir el monde en quel maniere li lyon pristrent en garde le cerf. Car bien sachies vraiement que ce nest nus encantemens ne oeure de dyable ains est vns merueilleus miracles qui auint iadis par la uolente de nostre seignor. Sire fait lancelot puis que nous nel poons sauoir par uous ne par autrui fors que par cel boin cheualier a qui diex en donra lonour Ie ne vous en esforcerai ia plus...»13 L’emplacement de la scène varie ici aussi. La première apparence des animaux symboliques n’est illustrée que dans Add. 10293 (fig. 7), alors que dans Ryl. (pendant d’Amst.) la deuxième apparence a été choisie (fig. 8);14 ce même point dans le texte est marqué dans Add. par un autre sujet, le désarçonnement de Lancelot et de Mordret effectué par deux chevaliers, encore un événement négatif pour les protagonistes (fig. 9). Les deux autres manuscrits qui contiennent une image de la deuxième apparence du cerf et des lions sont Bonn UB 526, copié à Amiens en 1286 (et illustré par un peintre audoumarois ou morinien) (fig. 10), et Oxford, Bodl. Rawlinson Q.b.6, réalisé à Paris peu après 1300 (fig. 11). Ce sont les seules images consacrées au cerf et aux lions dans ces manuscrits, malgré la présence dans Bonn de l’Estoire et de la Queste, et dans Rawl. Q.b.6 de la Queste; là encore, c’est l’échec de Lancelot qui est mis en avant. On s’attendrait peut-être à ce que les manuscrits qui incluent le cerf et les lions dans l’Estoire ou l’Agravain favoriseraient le motif aussi dans leurs choix d’images pour illustrer l’apparence de ces animaux dans la Queste del saint Graal. Seuls les cinq manuscrits, Bonn 526, Rawl. Q.b.6, Add.10292–4, Royal 14.E.III, BnF fr. 113–116, contiennent aussi la Queste. Mais l’apparence du cerf et des lions dans la Queste est illustrée uniquement dans Add. 10294: Perceval et Galaad, deux des chevaliers du Graal, sont représentés voyant le cerf accompagné des quatre lions. Cette fois selon le texte les animaux entrent dans un hermitage où un preudome se prépare Sommer, V 280. 6–16. Royal, troisième membre de ce groupe de manuscrits Flamands, ne contient pas le Lancelot. 13 14

LE MERVEILLEUX DANS LE LANCELOT-GRAAL

339

pour célébrer la messe; les chevaliers (y compris Boort et la soeur de Perceval, qui ne sont pas représentés dans l’image d’Add.) voient la transformation des animaux: le cerf devient le Christ en majesté entouré des symboles des quatre évangélistes: Matthieu l’homme, Marc le lion, Luc le boeuf, Jean l’aigle; puis ceux-ci disparaissent par une fenêtre sans la briser. Le preudome donne aux chevaliers la même explication qu’avait donnée Josephé à ces disciples dans l’Estoire. Un seul autre manuscrit fait allusion à cet épisode: la version postVulgate réalisée en Lombardie vers la fin du XIVe siècle inclut une image montrant les trois chevaliers (sans la soeur de Perceval) agenouillés devant une chapelle meublée d’un autel sur lequel repose le Christ en majesté entouré des symboles des évangélistes; un prêtre s’adresse aux chevaliers, sans doute expliquant la vision.15 Mais le cerf et les lions n’y sont pas représentés; et nous ne connaissons pas d’autre exemple parmi les nombreux autres manuscripts qui contiennent la Queste. On peut conclure que les attitudes envers cet épisode, comme devant le Graal, étaient ambigües, voir ambivalentes; tantôt ce phénomène merveilleux était associé à la réussite, tantôt à l’échec. Les commanditaires et dirigeants de la production manuscrite ont dû avoir été sensibles aux dichotomies que présentaient ce sujet, si pittoresque pourtant et rempli de symbolisme par ailleurs, pour avoir évité d’en faire un choix incontournable dans l’illustration du Lancelot-Graal.

15 BnF fr. 343, f. 58, reproduit sur Mandragore (12 juillet 2015) et dans M. Pastoureau et M.-T. Gousset, Lancelot du Lac et la quête du Graal, Arcueil, 2002, p. 85.

C Estoire and Queste Illustration

XVIII The Earliest Illustrated Prose Lancelot Manuscript?

L

ong recognized as the most widely read and most copiously illustrated of all the Arthurian romances, the Vulgate Cycle in Old French prose has held for many decades a high place in the published work of historians of medieval literature and of medieval secular art. Yet uncertainty still surrounds the origins of the prose cycle; its precise date and place of origin remain unclear and the early development of its accompanying picture cycles has never been thoroughly investigated. An examination of the illustrations in MS. Rennes, BM 255,1 as yet unknown to Art Historians and underestimated by Old French scholars, may clarify some aspects of the beginnings of the illustration of vernacular texts in France and at the same time help to assure this important manuscript its rightful place in the history of the manuscript tradition of Grail romances.

My work in Rennes was funded by the American Council of Learned Societies and the Graduate School of the University of Minnesota. I am grateful for their support. Monsieur L. Rousseau, Conservateur of the Bibliothèque Municipale, Rennes, was kind enough to allow me to study and photograph MS Rennes 255 and answered queries. My thanks are also due to Dr. F. Bogdanow, Prof. Dr. R. Haussherr, and Dr. P. Noble, for their assistance. In 2005 I selected Rennes 255 for the Lancelot-Grail Project and Mme Sarah Toulouse, Conservateur at the Bibliothèque des Champs libres (formerly Bibliothèque Municipale) kindly gave permission for us to have the manuscript photographed by DIAMM and to post the images on the web. That phase of the project was funded by a grant from the National Endowment for the Humanities. This article was first published in Reading Medieval Studies 3, 1977, pp. 3–44. 1 Catalogue général, pp. 119–20, MS 255 (148). See Appendix A for a description of the manuscript and Appendix B for a list of the illustrations.

344

I Rennes 255 contains the first three branches of the prose Lancelot cycle: Estoire, Merlin (without continuations), and Lancelot (ending incomplete shortly before the end of le Conte de la charette). The Rennes manuscript has not been used in any of the editions of these three texts,2 nor, with one exception, does the manuscript figure in the most recent discussions of their composition;3 indeed it has been wrongly dated by Micha in his study of the Merlin and Lancelot manuscript traditions: 1302–3 is the date given, no doubt in error for Rennes, BM 593, a compendium that includes the Prophecies de Merlin and was finished (at least in part) in January 1303. 4 No such date appears in Rennes 255 and I hope to show that the manuscript must in fact date from at least as early as the second quarter of the thirteenth century if not earlier. The place of the prose Estoire and prose Merlin in the compositional sequence of the Vulgate Cycle remains paradoxical. While there is general agreement that both depend on the verse Estoire and Merlin of Robert de Boron,5 there has been some difference of opinion as regards their chronology in relation to that of the Lancelot proper — Queste — Mort Artu branches, often referred to as ‘the pseudo-Map cycle’. While Lot considered Estoire earlier than Queste,6 they have been thought since Frappier7 and Micha8 to The text editions are listed in the bibliography at the end of this book. See now Ponceau’s edition of the Estoire. 3 The most significant studies are listed in the bibliography at the end of this book. My early dating of Rennes is cited in Arthur of the French by Roger Middleton, ‘The Manuscripts,’ p. 8 and by Kennedy et al., ‘Lancelot with and without the Grail,’ p. 277, and by Ponceau, I, pp. xxvii, xlvii, xlviii. 4 Micha ‘Les Manuscrits du Merlin en prose. ’ For Rennes 593(147) see Catalogue général, p. 238 ff. This MS. was the ‘MS de base’ used by Paton in Les Prophécies de Merlin. 5 Ed. Nitze, Roman de l’Estoire dou Graal, based on the only MS, Paris, BnF fr 20047: the Estoire text is preceded by Image du Monde and followed by a verse Merlin (ff. 55v–62v) which ends incomplete and is transcribed by Nitze at the end of his edition of Estoire (pp. 126–30). An earlier edition is Michel, Roman du Saint-Graal. Michel’s text was reprinted by Furnivall as an appendix to his edition of Estoire. See now O’Gorman, Joseph d’Arimathie, and Gowans, ‘What did Robert de Boron really Write ?’ and, for Rennes 255, Ponceau, passim, and Delcourt, Légende du roi Arthur, pp. 15, 22, 23, 46, 48. 6 Lot, Étude, stresses the unity of the Lancelot cycle as a whole and considers it all the work of one author. 7 Frappier, Mort Artu, 1936, pp. 27–146; 1959; 1961, appendix. 8 Micha, ‘Vulgate Merlin’. For Merlin illustrations see now Fabry-Tehranchi, Texte et image des manuscrits du Merlin, and pp. 26, 38, 37, 40, 44, 45, 46, 78, 150, 151. 2

THE EARLIEST ILLUSTRATED PROSE LANCELOT

345

be later reworkings of the Robert de Boron material composed as a preface to the pseudo-Map cycle, with the continuations of Merlin being the very latest compositions in the Vulgate Cycle. In fact, there is very little evidence for the date of composition of Estoire and Merlin or indeed for the rest of the Vulgate Cycle. Broadly speaking, the cycle falls between the date of Robert de Boron’s Estoire and mentions of the Vulgate Cycle, or parts of it, in later writings for which a date can be arrived at, although one cannot discount the possibility of lost sources for all. The significant dating points can be summarized as follows: Robert de Boron’s verse Estoire was written ‘o mon seigneur Gautier . . . qui de Mont Belyal estoit’.9 According to Villehardouin, Gautier de Montbelial joined the fourth crusade in 1201 and died in the Holy Land in 1212.10 Thus, to accommodate both the ‘o’ and the ‘estoit’ of the citation, the terminus ante for Robert de Boron’s Estoire would be 1201 or soon after, and so c. 1202 would be the terminus post for the prose Estoire. The Lancelot proper and Queste were used by Manessier, who wrote for ‘Jeanne la comtesse qui est de Flandre dame et maistresse’.11 Lot first interpreted this as meaning that Manessier wrote between 1214 and 1227 during Jeanne’s first regency caused by the imprisonment of her husband Ferrand of Portugal following his captivity at the battle of Bouvines,12 but revised his opinion to include also the second regency of Jeanne, 1233–7, between the death of Ferrand and her remarriage to Thomas of Maurienne.13 A second mention of the prose Lancelot was noticed by Paul Meyer in an anonymous prologue which prefaces the Philippide of Guillaume le Breton in MS London, BL Add. 21212.14 It refers to ‘le livres de Lancelot ou il n’a rime un seul mot’.15 Nitze, Estoire, lines 3490–1. Conqueste de Constantinople, ed. de Wailly, paragraph 533. Lot, Étude, p. 132, n. 6, gives 1202 for Gautier’s departure and 1212 or 1214 for his death, following Mas-Latrie, L’Histoire de Chypre, I, pp. 167, 171, 178–81; II, 12, 13, 14 and note. Gautier married Bourgogne, eldest daughter of Amaury de Lusignan, King of Jerusalem. 11 Perceval le Gallois, ed. Potvin, vol. VI, p. 155, 157 note. There is no reference either to Manessier or to Jeanne in the Mons MS; they occur in an epilogue in MSS Paris, BnF fr 12576, BnF fr 12577, and Montpellier, BIU Sect. Méd. H 249: see Nixon in Les Manuscrits de Chrétien de Troyes, Catalogue of Manuscripts, nos. 23, 39, and 25 respectively. 12 Lot, Étude, p. 135. 13 id., review of Pauphilet, Queste. 14 id., Étude, p. l35, n. 2; Meyer, ‘Prologue en vers’. This text occupies f. 4 of the MS and is followed on f. 5 by the Philippide in Latin. There is no reference to Lancelot in the Philippide itself. 15 ibid, lines 101–2. 9

10

346

This prologue was composed for a ‘seigneur de Flagi’ identified as Giles, châtelain de Sens, documented between 1203 and 1236; a reference to the death of Louis VIII in the same prologue narrows down its date to the period soon after the death of the king in 1226 to 1234 when the regency of Blanche of Castille came to an end. 16 Thus, the range of dates for the Vulgate Cycle would be from c. 1202 for Estoire at the earliest to c. 1234 for Lancelot and c. 1237 for Queste at the latest, and no indications for Merlin or Mort Artu. In so far as recent authors have concerned themselves with dating questions, the opinions are as follows: Frappier suggests a date sequence for the pseudo-Map cycle of c. 1215–20 for Lancelot, with the continuations of Lancelot and Agravain c. 1220–5, Queste 1225–30, and Mort Artu 1230–35;17 Hutchings dates Lancelot 1221–25;18 Pauphilet dates Queste c. 1220.19 Carman has suggested that the pseudo-Map cycle was conceived before the death of Eleanor of Aquitaine in 1204 and was still being composed up to the death of Isabelle of Angoulême, wife of John Lackland, in 1246. He assigns Estoire and Queste to the main period of 1217 or 1222 to the mid1230s.20 II Little effort has been made to relate these dates to those of the surviving manuscripts containing Vulgate Cycle texts. Studies on the manuscript tradition have established the existence of two versions, short and long, for the three texts contained in the Rennes MS; the place assigned to the MS in the manuscript tradition has been of variable significance. A clear picture of development is difficult to arrive at for the lengthy Lancelot proper and critics disagree as to whether the long or the short version is the earlier: Lot 21 and Kennedy22 consider the short version (otherwise known as the London version) the earlier, while Micha opts, at least in part, for the Paris or long ibid, lines 115–8: ‘. . . Loois/ le roi qui tant fu posteis/ Dom nos sommes tuit irascu/ De ce que si poi a vescu. ’ 17 Frappier, Mort Artu, p. 138. 18 Hutchings, Lancelot, p. LI. 19 Pauphilet, Queste, p. iii. 20 Carman, Pseudo-Map Cycle, pp. 128–31. 21 Lot, Étude, pp. 359–77. 22 Kennedy, ‘The two versions of the false Guinevere. ’ 16

THE EARLIEST ILLUSTRATED PROSE LANCELOT

347

version as the earlier. 23 For the equivalent of Sommer volume III Micha notes that the Rennes MS oscillates between groups ‘i’ and ‘e’,24 while it is omitted from his study of Sommer volume IV.25 For Merlin Micha considers the short version, ‘o’, closer to Robert de Boron’s poem and the long version, ‘B’ a later version composed as a preface to the Lancelot/Queste/Mort Artu cycle. He includes Rennes 255, with the siglum ‘i’, in group ‘a’,26 the early group. It is only in Dr. F. Bogdanow’s study of the Estoire that the significance of the Rennes text fully emerges.27 Dr. Bogdanow has shown that only the Rennes MS and the Portuguese Josep Abarimatia oscillate between the short and long versions; she concluded her 1960 article with two alternatives: both must contain a late version of the text that derives from the long as well as the short version, or both must stand at the beginning of the textual tradition.28 An examination of the extant prose Lancelot manuscripts, considering palaeographic and codicological aspects, so far as is possible, in addition to illumination, supports the view that the place of the Rennes MS is in fact among the earliest rather than the latest manuscripts. Few scholars have risked opinions on the precise dates of the surviving Vulgate Cycle manuscripts. Indeed, the question is fraught with difficulty due to the lack of dated copies, the absence of firm dating criteria for palaeography, codicology, or illumination, and the poor quality and bad condition of so many of the manuscripts. Lot had consulted Prinet and Martin about MS Paris, BnF fr 768, which he considered the earliest Vulgate Cycle MS; they were unable to be more precise than ‘milieu du règne de saint Louis’.29 ‘fin du treizième 23 Micha, ‘Études sur le Lancelot en prose,’ 1955, pp. 334–41; 1966, p. 214; 1973, p. 423, n. 9. 24 Micha, ‘Tradition manuscrite du Lancelot,’ p. 317. 25 ibid., pp. 478–517. 26 id., ‘Les manuscrits du Merlin,’’ p. 174. 27 Bogdanow, ‘Portuguese Josep. ’ 28 ibid., p. 375, ‘The fact that PR [Portuguese Josep and Rennes] alternate between the long and short versions can only mean one of two things: either they are the result of contamination or they represent the archetype from which the other two redactions derive. Any attempt to decide between these two alternatives calls for an examination of all the variant readings, and I propose to deal with the problem and its implications in a separate study. ’ Dr. Bogdanow was kind enough to inform me in a letter of October, 1975, that she favours the first of these two alternatives and considers that the Rennes MS. contains an earlier version of Estoire than the other French manuscripts. 29 Lot, Étude sur le Lancelot, p. 135 and n. 3.

348

siècle’ is the date assigned to MS Paris, BnF fr 20047, the only surviving manuscript containing Robert de Boron’s verse texts.30 The earliest dated Vulgate Cycle manuscript is Paris, BnF fr 342,31 written in 1274 by a female scribe,32 and containing Lancelot, Queste and Mort Artu; the earliest dated complete cycle is in MS Bonn, Landes-undUniversitästsbibliothek 526, written by Arnulphus de Kayo at Amiens in 1286.33 Both are copiously illustrated and provide unshakable termini for the existence of illuminated Vulgate Cycle manuscripts. However, these copies are both late in the textual and in the illustrative development of the Vulgate Cycle, and it is to the pre-1274 period that Rennes 255 and the earliest manuscripts belong. Of the 145 or so prose Lancelot manuscripts listed by Woledge34 there are some 76 that date from the thirteenth century, of which 32 may be ascribed to the last quarter of the century by stylistic analogy between their illuminations and those of Paris, BnF fr 342, Bonn LUB 526 Nitze, Estoire, p. v. This MS is Pauphilet’s and Frappier’s ‘D’. For its illuminations, see Vitzthum, Pariser Miniaturmalerei, p. 123; Loomis, Arthurian Legends, p. 93, pl. 213–6; Porcher, Manuscrits à peintures, no. 60; Stones, Illustrations of Lancelot, ch. 3, pp. 148–164, 410. See now Delcourt, Légende du roi Arthur, pp. 23, 167, 188. 32 The colophon on f. 234v reads ‘cis romans fu par escris en lan/del incarnation nostre segnor mil/deus cens & sixante et quatorse/le semedi apries les octaves de le trinite/pries pour ce li ki lescrist’. ‘ce li’ is a feminine pronoun, cf. Pope, From Latin to Modern French, p. 325. 33 The texts in Bonn are Hutchings’ ‘Z’ and Frappier’s ‘B’. Pauphilet gives no siglum. For its illuminations, see Olschki, Die romanischen Literaturen des Mittelalters, frontispiece; Loomis, Arthurian Legends, pp. 94, 96, pl. 217–23; Stones, Illustrations of Lancelot, ch. 5, p. 208–24, 451–2; now ead. Gothic Manuscripts, Part I, Cat. no. 121. Walterus de Kayo was the scribe of MS Le Mans 354 which contains Estoire and Merlin and was edited in part by Hucher, Estoire. The writing of the two scribes is quite distinct and so are the illuminations; Le Mans 354 is closer as far as its illustrations are concerned, to Paris, BnF fr 342 and the group of books associated with it, than to Bonn 526. It is closest to Paris, BnF fr 770, with which its text is also closely related, see Stones, ch. 3 and Hucher, Estoire. Hucher thought ‘Kayo’ referred to Caix (Somme) but Cayeux-sur-Mer (Somme) is an alternative worth considering. In either case the use of the place-name as a surname should mean that the individual concerned resided elsewhere and the ‘Ambianis’ in the Bonn colophon may equally indicate that Arnulphus did not normally work there. For fr 770 see now Delcourt, Légende du roi Arthur, pp. 23, 58, 115 no. 34; Fabry-Tehranchy, Merlin, pp. 26, 29, 45, 88, 91, 112, 123, 125, 141, 144–50, 225. 34 Woledge, Bibliographie and Supplément. The figure of 145 excludes fragments and takes no account of manuscripts that are now separate but might originally have been different volumes in a complete cycle. See the revised list at the end of this volume. 30 31

THE EARLIEST ILLUSTRATED PROSE LANCELOT

349

or other later dated manuscripts.35 This leaves a group of 44 which may be considered as ‘early’ examples.36 They may be divided into four groups: those with no illustrations (20); those with limited illuminations, usually only a single historiated initial or miniature at the opening of each branch (17); two whose miniatures have been cut out; leaving five that contain a sequence of illuminations. Rennes 255 belongs to the last of these groups. Possible contenders for the place of ‘earliest manuscript’ might be sought among any one of the four groups of pre-1274 manuscripts. The primary consideration which governed the quality of manuscript production in this as in any period must have been expense, and it would be misleading to assume a simple developmental progression from unillustrated manuscripts to minimally illustrated copies to fully illustrated versions over the period from the composition of the Vulgate Cycle sometime between c. 1202 and 1237 or even 1245, to the existence of a lavish illustrative sequence for at least part of the Cycle in 1274 and certainly all of it by 1286. In fact, there are indications that sparsely illustrated manuscripts were produced late in the thirteenth century: Paris, BnF fr 12580, for instance, contains one small miniature at the beginning of each branch (Agravain, Queste, Mort Artu) and is the product of a workshop operating in the last quarter of the century,37 while Paris, BnF fr 12581 and Paris, Bibl. de l’Arsenal 2997, also with one illustration for each branch, are both included in compendia written in 1284 and 1301 respectively,38 and the unillustrated Paris, BnF fr 24430 includes a text written after the fall of Acre in 1291.39 Any remarks about the unillustrated manuscripts which do not contain dating evidence must be tentative since the palaeographical and codicological evidence is, as yet, insufficiently precise for dating on these grounds to be firm. The one criterion that does appear to be significant is the relation between the top line of script and the top line of ruling. The 35 For the detailed justification of these stylistic attributions see Stones, Illustrations of Lancelot chs. 3–9. 36 See Appendix C for a list. 37 Folda, in Crusader Manuscript Illumination, p. 120, n. 20, relates this MS. and Città del Vaticano, BAV Vat. Reg. lat. 1490 to the work of the ‘Hospitaller Master’ which also includes a Rent Book of Saint-Germain des Prés dated 1276 and continues in Acre with Chantilly 590 in 1282. 38 For BnF fr 12581 see Omont, Ancien supplément français, II, pp. 566–7; for Ars. 2997 see Martin, Catalogue de l’Arsenal, III, p. 187. 39 Omont, Anciens petits fonds français, II, pp. 357–8.

350

change from ‘above top line’ to ‘below top line’ has been traced in English manuscripts by N. R. Ker40 and provides a useful key to the chronology of English productions of the thirteenth century. No such analysis of French manuscripts has yet been produced and the full impact of this change, in particular the precise dates when the change occurred, remain to be determined for French copying. The type of book is one factor that must be taken into account: bibles and long prose texts are written ‘above top line’ while contemporary psalters and verse texts are written ‘below top line’, even in the first decades of the thirteenth century. Thus, the verse Roman de Troie by Benoît de Saint-More, MS Paris, Bibl. de l’Arsenal 3340, written in 1237,41 has script ‘below top line’ while the prose Lancelots, London, BL MSS Landsdowne 757 and Roy. 20 B. viii (both unillustrated), Roy. 19 C. xiii and Paris, BnF fr 768A (both minimally illustrated), and Rennes 255 (copiously illustrated), all have script ‘above top line’. The prose texts need not be before 1237 but they are most likely to have been produced before the middle of the thirteenth century, by which time prose texts were normally written with script ‘below top line’.42 III While it would be premature to establish a date before 1237, or even before 1250, for Rennes 255 on the basis of the rulings alone, the illuminations and the school to which they belong confirm an early date for the manuscript and set this book apart from the other early illustrated prose Lancelot MSS. There are only five Vulgate Cycle MSS with sequences of illuminations that are earlier in date than 1274. They are Brussels, BR 9627–8; Modena,

Ker, ‘From “above top line” to “below top line” ’, p. 13. Martin, Catalogue de l’Arsenal, Ill, pp. 337–8. The text was edited from the available MSS by Constans. MS Paris, Bibl. de l’Arsenal 3340 is Constans’ ‘A’. It contains 26 foliate initials and two historiated initials: f. l, initial ‘S’, King Peleus with Jason and Hercules (?) (Constans, line 1); f. 5, initial ‘P’, Jason and Hercules set sail; Jason fights the serpent (Constans, line 715). These illustrations are earlier than the earliest illustrated MS in Buchthal, Historia Troiana, which is Paris, BnF fr 1610, written in 1264 and containing an extensive sequence of full-page miniatures, see below, n. 48. 42 This is the case in the mid-thirteenth-century Lancelots discussed below, section III: Brussels, BR 9627–8, Paris, BnF fr 339, and Ex-Phillipps Coll. 3643 (now University of California Berkeley 106); and also in the William of Tyre MSS produced in the same period, see below. 40 41

THE EARLIEST ILLUSTRATED PROSE LANCELOT

351

Bibl. Estense E 39; Paris, BnF fr 339; Ex-Phillipps Collection 3643 (now University of California Berkeley MS 106); 43 Rennes 255. In the case of Modena E 39 it is hard to assign a more specific date than the middle of the thirteenth century since its historiated initials are small, rubbed, and poor in quality (fig. 9a) so that little can be said about their style.44 The format of the historiated initials resembles the type that is current generally throughout the second and third quarters of the thirteenth century in French manuscripts, with initials enclosing a gold ground and bars decorated with acanthus, circles, wavy line or greek key motifs, the whole set on a blue or pink field with white wavy line, circle, or three dot motifs and sometimes bounded with an outer band of gold. This is also the format of the historiated initials in the other four early Lancelot MSS. Ex-Phillipps MS 3643 presents the same problem as the Modena MS: historiated initials of poor condition and quality and one can only postulate the same midthirteenth-century date for it as well. In both cases a large margin of error either side of 1250 should be allowed for. With Brussels, BR 9627–8 45 (fig. 9c) and Paris, BnF fr 33946 (fig. 9b) one is on firmer ground. Their illuminations are not the work of the same hand but both are related to one of the main styles current in the mid-thirteenth century in Paris and its immediate vicinity and whose characteristics are best exemplified in the Vie de Saint Denis of 1250, MS 43 Sotheby‘s 28. 11. 67, lot 92, pl. 10, now University of California, Berkeley, MS 106. Textual contents: f. 1 Vie des pères, f. 92v Gautier de Coincy, Vie de Théophile, f. 101 Jehan de Blois, Le Conte dou Baril, f. 105 Les neuf Joies de Notre-Dame, f. 105v La Passion de Jesus-Christ, f. 111 Vie de sainte Catherine, f. 117 L’Estoire del saint Graal; vol. Il f. l Merlin. There are 20 historiated initials in Estoire. See also Woledge, Bibliographie, p. 72; Micha, Manuscrits du Merlin, p. 89 (MS ‘m’); Stones, Illustrations of Lancelot, ch. 2, pp. 128–47, 299; Krause and Stones, Gautier de Coinci, pp. 305, 348, 350, 354, 375, 438. 44 See Camus, Codici francese della Regia Biblioteca Estense, no. 39, MS XI, B. 9 (E 39/ & L 930), where it is wrongly assigned to the 14th century. Textual contents: f. 1 Josèphe, f. 13v Merlin, f. 44v Perceval, f. 75 Lapidaire, see above, n. 2. The MS contains 15 historiated initials. See also Woledge, Bibliographie, p. 75; Micha, ‘Manuscrits du Merlin,’ p. 91 (MS ‘T’); Stones, Illustrations of Lancelot, ch. 2, pp. 128–47, 396–97. 45 Gaspar, and Lyna, Les principaux Manuscrits à peintures, vol. l p. 163, no. 69, pl. xxxiv. Textual contents: f. 1 Queste, f. 69 Mort Artu. This is Frappier’s MS ‘E’; Pauphilet gives no siglum. There are 37 historiated initials and two small miniatures. See also Woledge, Bibliographie, p. 72; Stones, Illustrations of Lancelot, ch. 2, p. 128–47, 398–9. 46 Textual contents: f. 1 Lancelot, f. 231 Queste, f. 264 Mort Artu. There are 120 historiated initials. This is Hutchings’ MS ‘O’, Pauphilet’s ‘A’, Frappier’s ‘I’. See also Woledge, Bibliographie, p. 74; Stones, Illustrations of Lancelot, ch. 2, p. 128–47, 399.

352

Paris, BnF naf 1098,47 made for Saint-Denis. The stylistic features of this and the works associated with it do not reveal the mid-thirteenth century as a high-point in the development of thirteenth-century French painting. The colours are monotonous: maroon, pink, deep blue and grey for the most part; the figure drawing makes use of square-shaped repetitious faces with blank expressions, multiple parallel lines for wig-like hair with a straight line at the brow; simplified drapery with pointed edges and broad straightline folds with little modelling. In the Vie de Saint Denis itself there is an extensive cycle of full-page miniatures unparallelled in the secular books of this period with the sole exception of the Roman de Troie MS Paris, BnF fr 1610, written in 1264.48 Other secular works illustrated in the style of the Vie de Saint Denis include several manuscripts of William of Tyre’s Histoire de la guerre sainte: Bern, Burgerbibl. 112 and 163; Paris, Bibl. de l’Arsenal 5220; Paris, BnF fr 779, fr 2630, fr 2827, fr 24208.49 In these and the Lancelots associated with this style, additional characteristics are the 47 Facsimile, ed. Omont; Vitzthum, Pariser Miniaturmalerei, pp. 2, 7; Porcher, Manuscrits à peintures, no. 5. 48 See above, note 41. Vitzthum, Pariser Miniaturmalerei, p. 12; Porcher, Manuscrits à peintures, no. 9; Saxl, ‘Troy Romance’; Buchthal, Historia Troiana, pp. 9–13, pl. 1–3. This manuscript contains 4 full-page miniatures with scenes in three registers, fols. 17v–18, 154v– 5. Four further miniatures from the manuscript are in the collection of Mr. J. H. van Heek at ‘s-Heerenberg, Holland, cf. Saxl, p. 129, n. 5 and Buchthal, p. 9, n. 2 and pl. 4–5. Vitzthum thought the manuscript was produced in the region of Cambrai and Buchthal follows Saxl in opting for an eastern French provenance, while Porcher ascribed it to Paris on the grounds of stylistic similarity with the Oxford/Paris/London moralized bible. In 1977 I subscribed to the latter view, on which see also Branner (Manuscript Painting, pp. 81, 198, cat. 220), Morrison (Illuminations of the Roman de Troie, pp. 83–105 and ead. ‘From Sacred to Secular,’ pp. 19–20, 25 n. 26), and Hedeman (‘Presenting the Past’ pp. 70–71 and fig. 36). I now (2013) agree with Jung that a better provenance is Burgundy, for Jung (La Légende de Troie, pp. 215–20) on linguistic grounds, and for me by analogy with Paris, BnF fr 24428, written in part by Omons in 1265 and attributable in my view to Metz or Burgundy (Stones, Gothic Manuscripts, Part II, cat. no. V–4). Other secular manuscripts with full-page miniatures of similar date are Paris, BnF fr 12558 and 12559, both copies of the Chevalier au cygne, and Paris, BnF fr 2186, Roman de la Poire (probably c. 1270) . In general, the full-page format is most infrequent in French secular illuminations. See Stones, ‘Sacred and Profane Art’, reprinted in these essays, p. 6, n. 17; ead., Gothic Manuscripts, Part I, cat. nos. III–3, III–8, I–10 respectively. 49 See Folda, Crusader Manuscript Illumination: Bern 112, pp. 33 n. 38, 125 n. 32; Ars. 5220, p. 168 n. 29; BnF fr 779, p. 114 n. 196, 120 n. 21; BnF fr 2630, pp. 33 n. 36, 85 n. 48, 157; BnF fr 2827, pp. 33 n. 36, 126; BnF fr 24208, pp. 33 n. 36, 85 n. 48, 125 n. 32, 157. These manuscripts all include continuations up to 1231.

THE EARLIEST ILLUSTRATED PROSE LANCELOT

353

treatment of armour, in particular helms and mail. Helms are often pointed on top and the sight is a single slit serving both eyes; mail is usually shown painted blue with white parallel lines alternating with rows of white dots (figs. 9b and 9c). The dating limits of this group will no doubt be clarified in R. Branner’s forthcoming book òn Parisian illumination in the time of Saint Louis. 50 By 1250, the date of the Vie de Saint Denis, the style had reached a developed form, and a related version of it is found as late as 1264 in the Roman de Troie. Another version of the Vie de Saint Denis style also appears, together with the Muldenfaltenstil of earlier in the century in the Oxford/ Paris/ London copy of the moralised bible which has most recently been dated c. 1240.51 Thus, an approximate date of c. 1240–1265 would seem appropriate for Brussels, BR 9627–8 and Paris, BnF fr 339. The illustrations of Rennes 255 belong to an earlier stylistic phase. Its 57 historiated initials are by an artist from a well-known and fairly clearly defined workshop operating in Paris in the second and third decades of the thirteenth century and whose major products include on the one hand a group of psalters with martyrological calendars and on the other hand the earliest of the four early thirteenth-century moralized bibles. What characterizes the best products of this workshop is above all the quantity and the quality of their illumination: the lengthy moralized bibles with their six historiated roundels per page have received only recently the detailed investigation they deserve,52 while the psalters, with their full cycles of prefatory miniatures and See, in the meantime, ‘The “Soissons Bible” Paintshop’ for a workshop that may be seen as marking the transition between the style of the Vie de Saint Denis and the Muldenfaltenstil of the earlier part of the century. 51 Haussherr, ‘Eine Warnung’, p. 390; id., ‘Petrus Cantor, Stephen Langton und Hugo von St. Cher,’ p. 348. This copy comprises MSS Oxford, Bodleian Library, Bodley 270b, Paris, BnF lat 11560, and London BL Harley 1526–7, and is reproduced in its entirety in Laborde, La Bible moralisée. The other early thirteenth-century moralized bibles are the onevolume MSS Vienna, ÖNB 1179 and 2554 and the three-volume MS in Toledo Cathedral Library, of which the last gathering is in New York, Morgan Library, M. 240. On the bibles moralisées see now the facsimile of the Toledo manuscript (Biblia de San Luis), the updated bibliography on Corsair for M. 240, and the study by Lowden, The Making. 52 See above, note 51 and in addition Haussherr, ‘Christus-Johannes-Gruppen’; id. ‘Beobachtungen an den Illustrationen zum Buche Genesis’; id., ‘Templum Salomonis und Ecclesia Christi’; id., ‘Bible moralisée,’ in Lexikon der christlichen Ikonographie; id., ‘Sensus litteralis und sensus spiritualis’; id., Commentary to facsimile of Vienna, ÖNB cod. 2554. Haussherr’s essays are now reprinted in Gesammelte Schriften, ed. König et al. 50

354

historiated initials exhibit a profusion and variety that clearly rank them as luxury products. The nucleus of the psalter group is formed by three psalters which have been the object of detailed study by R. Haussherr.53 They are MSS Manchester, John Rylands University Library lat. 22,54 St Petersburg, National Library of Russia Lat. Q. v. I, 67,55 and Paris BnF lat 1392.56 Within the illuminations of these three books there are slight stylistic variations which justify the notion that they were all products of a large workshop employing several illuminators. Haussherr includes the one-volume moralized bible Vienna ÖNB cod. 1179 and the three-volume Toledo version as products of the same shop,57 together with the Kristina Psalter, Copenhagen Kgl. Bibl. Gl. Kgl. Saml. 1606 4°,58 the Philadelphia Psalter, Philadelphia Free Library, Lewis Coll. E 185,59 the psalter-hours Paris, BnF lat 1073A,60 and (less

Haussherr, ‘Ein Pariser martyrologischer Kalendar’. ibid., and James, Descriptive Catalogue of the Latin Manuscripts, vol. I, pp. 64–71; Laborde, Bible moralisée, vol. 5, pp. 16–17; Haseloff, Psalter-illustration, pp. 34–36; Nordenfalk, ‘Insulare und kontinentale,’ p. 119; Peterson, Iconongraphy of the Historiated Psalm Initials; ead., ‘Visual Errors’; ead.,‘Scholastic Hermeneutics.’ 55 Haussherr, ‘Ein Pariser martyrologischer Kalender,‘ and Laborde, Bible moralisée, vol. 5, p. 72; id., Manuscrits a peintures de St. Petersbourg, I, pp. 3–4; Nordenfalk, ‘Insulare und kontinentale Psalterillustrationen,’ p. 118; Peterson, Iconongraphy of the Historiated Psalm Initials; ead., ‘Visual Errors’; ead.,‘Scholastic Hermeneutics.’ 56 Haussherr, ‘Ein Pariser martyrologischer Kalender,’ and Leroquais, Psautiers manuscrits, Il, pp. 137–9; Porcher, Manuscrits à peintures, no. 3; Haseloff, Psalterillustration, pp. 21–23, 78–80, 82; Peterson, Iconongraphy of the Historiated Psalm Initials; ead., ‘Visual Errors’; ead. ‘Scholastic Hermeneutics.’ 57 Haussherr, ‘Ein Pariser martyrologischer Kalender’, pp. 1100–3; Peterson, Iconongraphy of the Historiated Psalm Initials; ead., ‘Visual Errors’; ead.,‘Scholastic Hermeneutics.’ 58 Haussherr, ‘Ein Pariser martyrologischer Kalender’; Haseloff, Psalterillustration, pp. 78, 36; Nordenfalk, ‘Insulare und kontinentale Psalterillustrationen’, p. 118; Peterson, Iconography of the Historiated Psalm Initials; ead., ‘Visual Errors’; ead.,‘Scholastic Hermeneutics.’ See also Vidas, Christina Psalter. 59 Nordenfalk, ‘Insulare und kontinentale Psalterillustrationen’; Wolf, Catalogue of the John Frederick Lewis Collection, pp. 200–4; Illuminated Books, no. 52; Nordenfalk, ‘Insulare und kontinentale Psalterillustrationen’, p. 119; Peterson, Iconongraphy of the Historiated Psalm Initials; ead., ‘Visual Errors’; ead.,‘Scholastic Hermeneutics.’ 60 Haussherr, ‘Ein Pariser martyrologischer Kalender’; Leroquais, Livres d’heures manuscrits, l, pp. 52–5; Haseloff, Psalterillustration, pp. 78, 86; Peterson, Iconongraphy of the Historiated Psalm Initials; ead., ‘Visual Errors’; ead.,‘Scholastic Hermeneutics.’ 53 54

THE EARLIEST ILLUSTRATED PROSE LANCELOT

355

closely related), the hours New York, Morgan Library M. 9261 and the Psalter of Jacob Suneson, London BL Egerton 2652. 62 A number of bibles have been attributed by G. Schmidt to the same workshop: Göttweig 55, Vienna ÖNB 1144 and Baltimore, Walters Art Museum MS W 56. 63 To these may be added New York, Morgan Library, MS G. 3164 and Douai Bibl. Mun. 22.65 The illuminations in Rennes 255 and the bibles of the group consist of historiated initials which are more modest in format than that of the psalters and moralized bibles. The basic layout, border motifs and colour scheme of the former group is similar to that of the mid-thirteenth-century Lancelots and their contemporaries, and indeed this repertoire in the latter books is based on models established in the second and third decades of the century. However, the opening initials for each of the three branches of the text in the Rennes MS are similar in scale to the large historiated initials and roundels in the psalters and moralized bibles and allow a monumentality of design and execution not displayed in the mid-thirteenth century Lancelots. What is decisive in establishing the connection between the Rennes MS. and the psalter workshop is, first, the figure style and, secondly, some aspects of the minor decoration. The term Muldenfalten is generally used to describe the style current between c. 1200 and 1240 which is dominated by a drapery convention displaying a high degree of modelling and numerous parallel, hairpin-like, folds. A late phase of this style is found alongside the Vie de Saint Denis style in the Oxford/Paris/London moralized bible of c. 1240.66 It is to the early phase of this style that the Rennes Master’s figure style belongs, together with that of the psalters and early moralized and other bibles. The figures are elegantly proportioned and adopt graceful poses: the drapery folds are loose and lack the mannered rigidity of the later phases of Muldenfaltenstil, which begins even among the psalter group (cf. figs. 2a Haussherr, ‘Ein Pariser martyrologischer Kalender’; Burlington Fine Arts Club, no. 124; James, M. R., Catalogue of the Library of J. Pierpont Morgan, no. 74. 62 Haussherr, ‘Ein Pariser martyrologischer Kalender’ and Gad, ‘Psalter’; Peterson, Iconongraphy of the Historiated Psalm Initials; ead., ‘Visual Errors’; ead.,‘Scholastic Hermeneutics;’ Vidas, Christina Psalter. 63 Schmidt, ‘Buchmalerei,’ p. 112. 64 Plummer, Glazier Collection, no. 29, pl. 29. 65 Catalogue générale, VI, Douai. MS 22 is a glossed bible in 11 volumes from the Abbey of Anchin. 66 See Laborde, Bible moralisée. 61

356

and 2b); the faces are oval rather than square, with large eyes and loosely flowing hair with curls or waves at the brow. The best parallels for the faces and draperies of the Rennes Master seem to lie in the work of the main hand of the Manchester Psalter, Ryl. lat. 22, and there are some close affinities also with the psalter-hours, Paris, BnF lat 1073A, as well as with the bibles Douai 22 and New York, Morgan Library G. 31. Compare the format and motifs of the initial, and especially the treatment of pose, drapery, and faces, of the king in bed (figs. 6a, b, c, d) and the large profile head in the border of Rennes 255 (figs. 5c, d) and the heads in the Massacre of the Innocents in Paris, BnF lat 1073A (fig. 5b). The motif of the border head is hard to parallel exactly but its closest relation is with the initial on f. 42 of Vienna, ÖNB 1179 (fig. 5a), the earliest of the moralized bibles and the one that has most stylistic affinities with the Rylands psalter. Another connection with the early moralized bibles is the ‘sepulchre’ motif in the opening initial of Estoire in Rennes 255 (fig. la) which is commonly used in Vienna 1179 for the Temple of Jerusalem or for the Sepulchre of Christ.67 The allusion here is that of the sepulchre since the scene shows Christ appearing to the author (a hermit) and handing him the book containing the story of the Grail. In the background is an altar with a chalice on it, which may allude to the Grail itself, although elsewhere in the illustrations, as in the text, the Grail is a dish (escuele) rather than a chalice (cf. fig. 3a). Indeed, the use of a chalice here may be an indication that the Rennes Master was simply adapting a stock scene for this context.68 The minor decoration of Rennes 255 consists of small painted initials, of which the most distinctive are the ‘heraldic’ initials69 and the initials See Haussherr, ‘Templum Salomonis und Ecclesia Christi.’ Another scene which is comparable in many ways with this opening miniature is the first initial in the William of Tyre MS Paris, BnF fr 9081, f. 1 which shows Christ appearing to Peter the Hermit at the Holy Sepulchre; Peter is shown in a lying pose beneath a domed, round-arched structure (fig. 1 b). Folda (Crusader Manuscripts, p. 31 and n. 27) considers this is the earliest MS of the History of Outremer and dates it between 1244 and 1248; see Branner, Manuscript Painting, p. 59, 207, no atttempt at dating. Apart from this instance, I do not consider the study of the iconography of Rennes 255 pertinent to the present examination and I reserve it for a future occasion. See, however, Appendix B. 69 These initials cannot be considered truly heraldic. No serious attempt at heraldry in Lancelot manuscripts is met with before c. 1300 although the shields of owners appear in psalters and books of hours from the third quarter of the thirteenth century. An early appearance of heraldry in a secular manuscript (apart from Matthew Paris and the early rolls of arms) is in le Roman de la Poire, Paris, BnF fr 2186; the arms have not been identified. See 67 68

THE EARLIEST ILLUSTRATED PROSE LANCELOT

1a. Rennes 255, Estoire, f. 1. Initial C: Christ appears to the hermit-author with the books containing the story of the Grail (photo: Lancelot-Grail Project)

1b. Paris, BnF fr 9081, Guillaume de Tyre, Histoire de la guerre sainte, f. 1. Initial L: Christ appears to Peter the Hermit at the Holy Sepulchre; Peter the Hermit leads the First Crusade (photo: BnF)

357

358

2a. Rennes 255, Merlin, f. 101. Initial M: Harrowing of Hell (photo: Lancelot-Grail Project)

2b. Philadelphia, Free Library, Lewis Coll. E. 185, Psalter, f. 17. Three Maries at the Sepulchre; Harrowing of Hell (photo: author)

THE EARLIEST ILLUSTRATED PROSE LANCELOT

359

3a. Rennes 255, Estoire, f. 76v. Initial O: Josephé and his companions with the Grail (photo: Lancelot-Grail Project)

3b. Rennes 255, Lancelot, f. 137. Initial E: Aramont Duke of Brittany is handed a crown by King Uterpandragon; they embark with their knights (photo: Lancelot-Grail Project)

3c. Rennes 255, Lancelot, f. 188v. Initial O: standing man holding book in veiled hand (photo: Lancelot-Grail Project)

360

4a. Manchester, The John Rylands University Library, latin 22, Psalter, f. 12, Miracles of Christ (photo: The John Rylands University Library)

THE EARLIEST ILLUSTRATED PROSE LANCELOT

361

5a. Vienna, ÖNB 1179, Bible moralisée, f. 42. Initial V and scenes of Baptism (photo: Warburg Institute)

5c. Rennes 255, Lancelot, f. 197v. Initial Q: Lancelot rides off following his host (photo: Lancelot-Grail Project)

5b. Paris, BnF latin 1073A, Psalter-Hours, f. 7v. Massacre of the Innocents (photo: BnF)

5d. Rennes 255, Lancelot, f. 197v. Male head terminal (photo: Lancelot-Grail Project)

362

6a. Rennes 255, Estoire, f. 72v. Initial C, Duke Ganor in bed (photo: Lancelot-Grail Project)

6b. Rennes 255, Estoire, f. 10v. King Evalach in bed (photo: Lancelot-Grail Project)

6c. Manchester, The John Rylands University Library, latin 22, Psalter. f. 130v, Psalm 4, initial C: King David in bed (photo: The John Rylands University Library)

6d. New York, Pierpont Morgan Library and Museum, G.31, Bible. F. 106, Kings IV, initial P: Elijah addressing King Ahaziah in bed (photo: Pierpont Morgan Library and Museum)

THE EARLIEST ILLUSTRATED PROSE LANCELOT

363

7a. Rennes 255, Estoire, f. 96. Initial O: Josephé blesses his brother Galaad after crowning him (photo: Lancelot-Grail Project)

7b. Manchester, The John Rylands University Library, latin 22, Psalter. F. 130v, Psalm 51, initial Q, Doeg telling King Saul that David has come to the house of Ahimelech (photo: The John Rylands University Library)

7c. Douai, Bibl. municipale 22, Glossed Bible, vol. 5, f. 9. Sapientie, Initial D: Solomon receiving the sword of justice (photo: author)

364

8b. Rennes 255, Lancelot, f. 179. Initial Q, Sciapod (photo: Lancelot-Grail Project)

8a. Rennes 255, Estoire, f. 49, minor initials (photo: Lancelot-Grail Project)

8c. Rennes 255, Estoire, f. 100. Initial Q, Centaur (photo: LancelotGrail Project)

8e. Manchester, The John Rylands University Library, latin 22, Psalter, Psalm f. 71v Initial ‘I’, Large dog-like creature attacks small dogs (photo: The John Rylands University Library)

8d. Manchester, The John Rylands University Library, latin 22, Psalter, Psalm 78, D initial, Dog-like creature attacks man in underwear (photo: The John Rylands University Library)

THE EARLIEST ILLUSTRATED PROSE LANCELOT

365

9a. Modena, Biblioteca Estense, E 39, Merlin, f. 13v. Initial M: Harrowing of Hell (photo: Biblioteca Estense)

9b. Paris, BnF fr 339, Queste, f. 257. Initial O: Galaad, Boort and Perceval in Solomon’s enchanted ship (photo: BNF)

9c. Brussels, BR 9627-8, Mort Artu, f. 124. Initial O: Arthur and his men sail from London (photo: BR)

366

containing animals (figs. 6a, b, c). The ‘heraldic’ initials are outlined in gold on a ground of blue or pink and contain a coloured field of pink or blue against which is set a lion rampant or an eagle in white. These motifs are similar to the bird and animal motifs that occur occasionally in the backgrounds of Vienna 1179 (fig. 5a) and are common in the backgrounds between the historiated roundels in the full-page miniatures of the psalters (figs. 2b, 4).70 The animal initials are painted in full colour with modelling on the animals in white; the animals are set against gold grounds, enclosed within the coloured bar of the initial which, in turn, is on a coloured ground outlined in gold — the same format as the historiated initials. As with the historiated initials, these animal initials find their closest parallel in the minor initials of the Rylands Psalter (cf. figs. 8d, e). The animal initials in Rennes 255 would appear to be the work of a different hand than the artist of the historiated initials; two of them contain human figures which are somewhat different in style from the figures in the historiated initials (fig. 3c). This artist may be the painter of the miniature of the prince in Gilles de Paris’ Miroir des Princes, Paris, BnF lat 6191, f. 1v,71 although it would be rash to base a comparison on only three examples of the style. The lavish pen decoration of the psalters is only occasionally found in Rennes 255, where it embellishes painted initials (fig. 3c, 5c). Coloured ink is also used on a modest scale for paragraph marks in the margins of text columns and as infilling to enliven capitals in the body of the text, an unusual practice in secular books and one which, along with the size, clarity, and relative lack of abbreviations in the writing, serve to set this volume apart from the unillustrated Lancelots and those produced in the middle of the century (figs. 3b, c, 5d, 8a, b, c). IV The date of the Rennes Lancelot can be established only approximately since the volume itself contains no evidence, but it cannot be too different also below, n. 78, for the Navarre and Champagne shields in the Toledo moralized bible. For Arthurian heraldry see Stones, Illustrations of Lancelot, ch. 10, pp. 359–84 and now ead., ‘Les débuts’ and ‘Gauvain,’ both reprinted in these essays. See also my discussion of the Wollaton manuscript WLC/LM6 in relation to Rennes 255, also reprinted in these essays. 70 This ‘heraldic’ motif continues to be used for painted initials until c. 1300 and is not in itself a significant factor in establishing stylistic groups. 71 Reproduced in Branner, ‘Manuscript Painting in Paris around 1200,’ pl. 13.

THE EARLIEST ILLUSTRATED PROSE LANCELOT

367

in date than the products of the Parisian workshop with which its closest comparisons lie, in particular Ryl. lat. 22, Douai 22, Morgan G. 31, Vienna 1179, BnF lat l073A, and possibly also BnF lat 6191. The earliest is BnF lat 6191 which was written for Louis VIII while dauphin and presented to him on 3 September 1200. One page, showing the Virtues, was added to the book between 1216 and 1223; it is just possible that the initial showing the prince dates from the same period,72 although it is more likely, if this manuscript is indeed the presentation copy, that the initial would have been completed by 1200. Of the other books close in style to Rennes 255, only Ryl. lat. 22 and Vienna 1179 contain evidence for dating. Ryl. lat. 22 must be after 1220 since its calendar contains the Translation of Saint Thomas of Canterbury which took place in that year;73 among the psalter group, however, its close relative BnF lat 1392 was made before 1220, since the Translation of Saint Thomas is lacking,74 and the Suneson Psalter, BL Egerton 2652 was produced before 1224 since it contains a calendar of Roskilde that lacks Abbot William, canonized in 1224.75 Thus, it is likely that Ryl. lat. 22 belongs to the early 1220s. Vienna 1179 contains on f. 246 a portrait of its royal patron with either a scribe or an illuminator, and an accompanying poem that would no doubt reveal the identity of the king represented in the miniature were it still readable.76 Whether the king is Philippe Auguste (d. 1223), Louis VIII (d. 1226) or Louis IX (d. 1270) is unclear from the miniature alone. One may however compare this portrait with the one in the other early moralized bible, the three-volume Toledo version, which, in its New York section contains on f . 8 a portrait of a young king and a queen, who are generally interpreted as representing Saint Louis and his mother, Blanche of Castille, and this version is normally dated to the regency of Blanche, between 1226 and 1234.77 An additional reason for dating the Toledo version before 1234 is that it contains in volume 2 (f. 78) the shields of Navarre and Champagne which are shown separately but 72 Branner, ibid., p. 176 and p. 179 considered the initial and the Virtues page the work of the same hand, a view which I do not share. For the adding of the Virtues page, see ibid., p. 179, referring to Delaborde, ‘Note sur le Carolinus de Gilles de Paris. ’ 73 Haussherr, ‘Ein Pariser martyrologischer Kalender’ especially pp. 1095–6. 74 ibid., p. 1100. 75 ibid., and Gad, ‘Psalter. ’ 76 Haussherr, ‘Sensus litteralis,’ p. 365. 77 Laborde, Bible moralisée, V, pp. 41–73; Haussherr, review of Sauerländer, Von Sens bis Strassburg, p. 318. For the portraits compared, see Haussherr, ‘Sensus litteralis,’ pl. 67–8.

368

were united in 1234.78 On stylistic grounds Haussherr dates the Toledo bible c. 1230 and considers it slightly later than Vienna 1179.79 Since the king in Vienna 1179 appears alone it should represent Louis VIII, or, possibly, Philippe Auguste, and should have been at least planned, if not completed, by 1226.80 Such a date accords well with the other activities of the workshop on the psalters during the early 1220s. That Paris was the location of the workshop is indicated not only by the royal portraits in these two moralized bibles but also by the exclusively Parisian feasts in Rylands lat. 22.81 On the basis of the evidence of illumination, then, the production of Rennes 255 most probably took place in Paris in the early 1220s in the workshop that made the luxury psalters and the early moralized bibles. There is no surviving parallel in French secular illumination for a book of this date;82 there are no illustrated epics before the middle of the thirteenth century,83 the earliest Troie manuscript dates from 1237,84 while the earliest illustrated Chrétien manuscript probably dates also from the middle of the century85 and the earliest copy of William of Tyre’s Histoire de la guerre sainte is currently dated c. 1235–45 or 1244–48.86 Thus, the Rennes MS. Branner, ‘Soissons Bible,’ p. 23, n. 34. Haussherr, ‘Sensus litteralis,’ p. 357; id., ‘Warnung,’ p. 390 and n. 2; id., ‚Petrus Cantor,‘ p. 348. 80 Haussherr, ‘Sensus litteralis,’ p. 365 asks ‘Sollte es sich um Ludwig VIII. (1223–1226) handeln?’ and dates Vienna 1179 between 1220 and 1230, see id., ‘Templum Salomonis,’ p. l06; id., ‘Petrus Cantor,’ p. 347; id., ‘Warnung,’ p. 391. For debate on date, see now Lowden, The Making, p. 8 (c. 1234–35 for Oxford/Paris/London}, p. 50 (c. 1220–30 for Vienna 2554, following Haussherr), p. 51 (mid-1220s for Vienna 1179; late 1220s–1230s for Oxford/ Paris/London). 81 id., ‘Ein Pariser martyrologischer Kalender,’ p. 1100. 82 There is, however, an important illustrated MS containing classical texts in Latin by Virgil, Statius, Lucan and Claudian, Paris, BnF lat 7936. See Avril, ‘Un Manuscrit d’auteurs classiques.’ And see now the Wollaton manuscript, discussed elsewhere in these essays, 83 Notable exceptions are the Roland window at Chartres, see Lejeune and Stiennon, La légende de Roland, pt. ll, ch. 3, pIs. VII–XVIll; the Codex Callixtinus of the mid-twelfth century (in Latin), ibid., pt. l, ch. 5, pl. I, II. See also the review by Ross; see now Maines, ‘Charlemagne Window’; Stones, ‘Iconography of Charlemagne’; Schenck, ‘Taking a Second Look. ’ 84 See above, notes 41 and 48. 85 Paris, BnF fr 12576. I am preparing a study of Chrétien illustration. See now Stones, ‘Illustrated Chrétien Manuscripts’; and Delcourt, Légende du roi Arthur, p. 101 cat. 25. 86 Folda, Crusader Manuscript Illumination, p. 31, n. 27 refers to Branner’s forthcoming book where this date is assigned; in fact, Branner is unspecific about the date (Manuscript Painting, p. 59). See also note 68 above. 78 79

THE EARLIEST ILLUSTRATED PROSE LANCELOT

369

occupies a place at the forefront not only of prose Lancelot illustration but of French secular illustration in general. The early 1220s bring one very close to the likely date of the composition of the first parts of the Vulgate Cycle itself and again the Rennes MS stands at the beginning of the manuscript tradition. There is, unfortunately, no evidence to connect the book with a specific patron and one can do no more than speculate as to any possible relationship between its original owner and the author of the prose text. Could this important manuscript, written and illustrated in a Parisian workshop that executed royal commissions, be the presentation copy of a prose version of the Lancelot cycle specially composed by royal command ?

XIX Two French Manuscripts: WLC/LM/6 and WLC/LM/7

T

he Wollaton library collection includes two 13th-century French vernacular manuscripts, MSS WLC/LM/6 (figs. 1–13, 23–25) and WLC/LM/7 (fig. 23). WLC/LM/6 is a lengthy miscellany with 83 illustrations in the form of small framed miniatures in half a text column, of which 68 illustrate the characters and actions in the various narratives and 15 are decorative panels containing hybrid motifs or birds. It has received attention mostly from the various text editors, and has usually been dated late in the 13th century.1 However, its codicological and decorative features allow this opinion to be radically modified, returning to the date in the first quarter of the 13th century proposed by W. H. Stevenson in 1911 and Henri Omont in 1912.2 Here I examine the decorative and artistic context to which the manuscript belongs and concur with Omont’s early 13th-century date and with Terry Nixon’s view that WLC/LM/6 is ‘the earliest fully illustrated romance collection’.3 WLC/LM/7 is an incomplete copy of the Estoire del Saint Graal, the first branch of the popular Lancelot-Grail romance in French prose. Although it contains no illustration, codicological features corroborate Ponceau’s view that WLC/LM/7 falls very early in the manuscript tradition of the Estoire for First published in The Wollaton Medieval Manuscripts. Texts, Owners and Readers, ed. Ralph Hanna and Thorlac Turville-Petre, Woodbridge Suffoks and Rochester NY, 2010, pp. 41–56. I have renumbered the illustrations in this version. 1 Chanson d’Aspremont, ed. Brandin, I, vi (‘troisième quart du XIIIe siècle’), followed by all the other text editors. 2 Stevenson, Report, pp. 221–34; Omont, ‘Manuscrits de lord Middleton’. 3 Nixon, ‘The role of audience’, p. 447.

TWO FRENCH MANUSCRIPTS

371

which a date in the first quarter of the 13th century may also be posited.4 The distribution of its party-bar (two-colour) pen-flourished initials (see fig. 23) shows that a system of markers was developed early in the textual tradition and was expanded upon and accompanied by illustrations in later copies.

ms wlc/lm/6 The beginnings of the illustration of literary manuscripts in French are confined to a handful of manuscripts with a limited number of pictures. This is in sharp contrast to Latin texts such as the Lives of the Saints which were copied across Europe as fully illustrated shrine books from the 11th century onwards.5 Even Pseudo-Turpin’s Latin account of the campaign of Charlemagne in Spain and the Battle of Roncevaux was illustrated from the mid-12th century with narrative pictures, beginning with Santiago de Compostela, Archivo de la Catedral, MS 1, written and illuminated by French craftsmen.6 If the early 13th-century Architrenius of Master Johannes Havillensis (Edinburgh UL, MS D.b.VI.6), has only a single historiated initial of Xerxes sailing over Mount Athos and crossing the Hellespont,7 the late 12th-century Virgil has a series of historiated initials (BnF lat 7936), and Latin texts continue to enjoy a robust illustrative tradition through the 13th century.8 The earliest literary manuscripts in French are only sparsely illustrated. A single historiated initial in the late 12th- or early 13th-century Tristan of Thomas (Oxford, Bodl. MS Fr. d.16, f. 10) depicts a figure, the eponymous hero or Yseut, playing the harp and possibly singing (the paint on the face has flaked, so one cannot be sure),9 and the famous ‘Guiot’ manuscript of Chrétien de Troyes also has a single historiated initial, at the opening of the L’Estoire del saint Graal, ed. Ponceau, I, pp. xxvii, xlvii, xlviii. Important examples made in France in the late 11th or early 12th century are the Lives of Saints Martin of Tours, Aubin of Angers, Omer of Saint-Omer, Quentin of Saint-Quentin, the Miracles of Sainte-Foi de Conques, and many others. 6 Stones, ‘Four Illustrated Jacobus Manuscripts’; Stones and Krochalis, The Pilgrim’s Guide to Santiago de Compostela, Critical Edition, I, pp. 28–30, pl. 19–21. 7 Borland, Catalogue of the Western Mediaeval Manuscripts in Edinburgh University Library, pp. 30–31, assigning it to England. 8 Avril, ‘Un manuscrit d’auteurs classiques’. 9 The initial marks the introduction of Tristan’s or possibly Yseut’s musical accomplishments as he or she sings a lay. The costume and headgear worn by the figure leave the gender of the 4 5

372

Chevalier de la charrette, depicting Marie de Champagne, at whose request the text was written (BnF MS fr 794, f. 270.10 The Guiot manuscript has been related to a large number of manuscripts, many of them in French, illuminated or decorated in the style of the Manerius Bible, Paris, Bibl. Sainte-Geneviève, MSS 8, 9, 10.11 Whereas the Bible is beautifully and extensively illustrated with historiated initials, the vernacular manuscripts are considerably more limited in their decoration. Against this background, WLC/LM/6 marks an exciting new departure since its 83 illustrations are without antecedent among vernacular manuscripts in French.12 They take the form of small framed miniatures set in half a text column (Plates 5a–m). These replace the first letter of the first word that follows, suggesting that a format of historiated initials rather than miniatures was originally envisaged, as was the case with the Tristan and Chevalier de la charrette referred to above.13 The historiated initial format was also chosen for the two earliest illustrated Arthurian manuscripts in prose, Modena, Bibl. Estense, MS E 39 (olim alpha L 9 30) (Robert de Boron’s Joseph and Merlin in prose, followed by Perceval in prose, Mort Artu and a Lapidaire, ending incomplete, fig. 19), and Rennes, Bibl. mun., MS 255 (L’Estoire del Saint Graal, Merlin, and a fragmentary Lancelot in prose; figs. 14–18, 20).14 These are the two literary manuscripts which, despite many differences, offer the closest parallels for WLC/LM/6. Other literary manuscripts mentioned below are much more sparsely illustrated. Both Rennes and Modena are larger in format than WLC/LM/6: Rennes measures 434 X 310 mm, and is written in three columns of 45 lines; Modena measures 310 x 215 mm and is written in two columns person open — deliberately so ? For other instances of early Tristan iconography, see Stones, ‘The Artistic Context of Some North French Illustrated Tristan Manuscripts’, reprinted in these essays. 10 See Les Manuscrits de Chrétien de Troyes, II, cat. no. 8. Copied by Guiot in Provins, it is attributed to the second quarter of the 13th century and was perhaps a commission by Thibaut le Chansonnier. See also the exhibition catalogues, Splendeurs de la Cour de Champagne, pp. 11 (fig.), 15, 38, cat. no. 34, and La Légende du roi Arthur, pp. 21, 42, 58, 98,100, cat. no. 23. 11 Stirnemann, ‘Some Champenois Vernacular Manuscripts’, and ead., ‘Une Bibliothèque princière au XIIe siècle’, and Splendeurs, cat. nos. 32–45. 12 The texts and subjects of the illustrations are listed in the Appendix below. 13 There are many letters written in leadpoint in the margins of WLC/LM/6, corresponding to the letters displaced by the miniatures. They may have been guide letters, put in at the planning stage, or they may have been added later. 14 Stones, ‘The Earliest Illustrated Prose Lancelot Manuscript?’ reprinted in these essays.

TWO FRENCH MANUSCRIPTS

373

of 38 lines. Both Rennes and Modena were written above top line,15 in large elegant scripts (several related scribes wrote Rennes 255) with few abbreviations. While WLC/LM/6 was also written above top line it is a smaller, more oblong book, measuring only 297 x 200 mm, written in two columns of 48 lines, and its scripts are correspondingly smaller and more compressed. All three manuscripts have minor decoration consisting of pen-flourished initials in red and blue, alternating, with flourishing in the other colour. In Rennes these initials are rare but elegant, in gold and blue, in a single quire in the Lancelot section (quire T, ff. 145–152v; fig. 17); in Modena (fig. 19) and WLC/LM/6 they are used throughout, although in WLC/LM/6 there are two flourishers at work, a major decorator who uses a light shade of blue and does extremely neat if exuberant flourishing, and a second decorator who uses a darker shade of blue and whose flourishing might be described as sloppy (in the Ille et Galeron section, quires Q, R, S, ff. 157–187v; and half-way through the Aspremont section to the end of Raguidel, quires AA–GG, ff. 313–335v, with the exception of ff. 306 and 314 (fig. 25) which may be the work of a third pen-flourisher). Of interest in WLC/ LM/6 is the occasional use of flourished descenders in the ink of the scribe, a feature which suggests that the scribe may have also been the primary flourisher (especially noteworthy on ff. 153v, 161v–164 [fig. 1], 224v). Rennes 255 has 57 historiated initials, several initials with animal motifs drawn from the bestiary (figs. 14, 16), similar to the animal initials in WLC/ LM/6 (figs. 2, 7, 9, 10).16 All the illustrations to L’Estoire and the opening historiated initials of Merlin (fig. 15) and Lancelot are narrative in content and depict subjects that depend closely on the text, but there is only one more illustration in the Merlin, and most of the initials in the rest of the Lancelot section are by a lesser artist, working in a smaller format (four-line initials as opposed to seven or eight lines), often depicting a single figure with little or no relation to the text. There are also several changes of decorator in the Lancelot section of Rennes. Modena E 39 has 15 historiated initials and two gaps where initials have been removed. The illustrations, though tiny in scale and poorly preserved, are all closely dependent on the text. Modena That the text is written above the top ruled line suggests that it was written before c. 1240, although there is still no comprehensive study of the question for French manuscripts. For England, see Ker, ‘From “Above Top Line” to “Below Top Line”.’ 16 Illustrated on http://www.lancelot-project.pitt.edu. See Cassagnes-Brouquet with Clouzot, Les Romans de la Table ronde: premières images. 15

374

lacks initials that are painted in full colour but are non-narrative in content. This is a feature that Rennes and WLC/LM/6 have in common. In both manuscripts these initials are the work of the main artist.17 Whereas the Rennes manuscript was most likely made in royal circles in Paris c. 1220 and is stylistically related to the cluster of royal psalters and to the early Moralised Bibles and other books, the Modena manuscript is extremely hard to place stylistically, and parallels are not readily forthcoming. I tentatively suggest the region of Champagne or north France, the former because it provided an important patronage base at the turn of the 12th to 13th century, although none of the manuscripts discussed by Stirnemann look quite like Modena; on the other hand, the characteristic shape of the historiated initials with borders designed in a rectilinear step-pattern arrangement is similar, for instance, to the format of the initials in the noted missal of Arras, Bibl. mun. MS 888 (13th century, second quarter?) and in the Bible, Arras, Bibl. mun., 561/Boulogne, Bibl. mun. MS 4 (mid13th century?).18 But the small-scale figures of Modena remain unmatched. Among the Chrétien manuscripts, the sparsely illustrated (two historiated initials) and poorly preserved manuscript Bern, Burgerbibliothek, MS 354, written above top line, perhaps offers the closest stylistic analogy.19 Terry Nixon compared it with the Roman de Troie written (also above top line) in 1237, Paris, Bibl. de l’Arsenal, MS 3340, with a single historiated initial at the beginning, depicting the walls of Troy.20 As noted above, Rennes is similar to WLC/LM/6 in its choice of decorative initials containing non-narrative motifs unrelated to the text — birds, harping animals, a centaur (fig. 16), lions, eagles (fig. 14), a sciapod (fig. 18) — some of which are motifs borrowed from the bestiary tradition. Occasionally the choice fits the text, so that an eagle initial is placed at a passage where blindness is discussed in the text. In WLC/LM/6 these initials are distributed across all the texts, according to a rationale that is In WLC/LM/6 they seem to me perfectly consistent with the work of the miniaturist in framing, draughtmanship, and colour palette, contrary to the opinion of Thorpe, Roman de Silence, p. 6, and Jung, Légende de Troie en France, p. 131. 18 These may be consulted on the website Mandragore: http://mandragore.bnf.fr/html/ accueil.html. 19 Bern 354 has been assigned variously to northern Champagne, Nièvre, Allier and Haute-Marne (summary in Les Manuscrits de Chrétien de Troyes, cat. 10). 20 Ibid. See also Samaran and Marichal, Manuscrits datés, I, pl. 12; and, for the iconography, Stones, ‘Seeing the Walls of Troy,’ reprinted in these essays. 17

TWO FRENCH MANUSCRIPTS

375

now difficult to reconstruct. Some are hybrid creatures, not derived from the bestiary, but others depict an eagle (f. 227; fig. 10), a pelican apparently ripping its breast, but without its young (f. 218v; fig. 9), a griffon biting its tail (f. 217); a lion with flowering tail (f. 213). There are more of these motifs in the Roman de Silence than in the other texts, suggesting that textual markers were needed but subjects from the text had not been planned for that section. In terms of numbers of illustrations, WLC/LM/6 contains many more than either Rennes or Modena, and they differ in format, as noted above. Overall they are considerably less specific to the text than those in Rennes and Modena. While most of them are broadly appropriate to their textual context, the depiction of particular details from the text is rare — examples are Emenidus lacing up his mail chausses (f. 240v; fig. 11) and Hector’s squire attaching the lacing on Hector’s helmet (f. 84; fig, 4 ), both exceptional. Many illustrations depict a single figure, a portrait of the author, shown writing (f. 126v), laying his hand on his text (f. 188; fig. 8), pointing to the text (f. 137v), or the hero of the story (Alexander, f. 224, Charlemagne, f. 244v; King Manuel, f. 277v). Occasionally single figures seem to be misunderstandings, as Emenidus at the end of the Fuerre de Gadres (f. 270). Standard scenes are two figures talking; two or three riders riding; a pair of lovers (f. 170), a protector and his protegée (ff. 195v, 333v; fig. 12). Overall there is a general avoidance of the complexity of battle and the drama of death, even in the Roman de Troie and the Chanson d’Aspremont, and many colourful events narrated in the various texts are not depicted in the illustrations — Girart having Agolant buried and his head (still in its helmet) embalmed at the end of Aspremont, Namles’ fight with the giant bird and killing the bear and leopard earlier in the story. Nevertheless, the number of miniatures marks an impressive beginning to what would develop in the course of the 13th century into a serious interest in illustrating French texts with captivating details of the stories. The miniatures in WLC/LM/6 were considered (at an uncertain date) precious enough to be protected with silk, of which the sewing holes may still be seen on many folios (e.g. f. 244v). It is notable that there are no traces of preliminary notes or sketches as is often the case with the invention of a new set of illustrations, as in the late 13th-century Roman de la rose, Città del Vaticano, Biblioteca Apostolica Vaticana, MS Urb. lat. 376,21 or 21

König with Bartz, Der Rosenroman des Berthaud d’Achy.

376

Jacobus de Voragine’s Legenda aurea, San Marino CA, Huntington Library, MS HM 3027.22 Rennes 255 and Modena also lack such guides, although they, too, appear to stand at the beginning of their respective illustrative traditions. Questions remain as to when and where WLC/LM/6 was made, and for whom. Whereas later ownership by the Laval family is attested by the 14th-century inscription on f. 249v, ‘cest liure est Madame de la Val’, the notion that an ancestor of the couple, such as Gui de Laval and Béatrix de Gavre (married in 1286), was responsible for the commission, must remain speculative. Certainly 1286 is far too late for the making of this manuscript. Similarly, the dedication to Thibaud de Blois, count of Champagne († 1191) in the epilogue of Ille et Galeron, unique to WLC/LM/6, is most likely too early for this manuscript and was probably copied from its exemplar.23 If Rennes and Modena, dating in all likelihood at the end of the first quarter of the 13th century, offer some analogies for WLC/LM/6, neither is stylistically comparable. The palette of WLC/LM/6 with its bright colours and an emphasis on orange, white, green and blue shading from dark to light tones is not matched in either Rennes or Modena. I suggest that WLC/LM/6 may be related to a cluster of very distinguished psalters made in monastic circles in northern France at the end of the 12th century, not in Champagne but further north, on the border of Artois and Flanders. Two manuscripts in particular offer a similar palette of light bright colours on gold in which an orange/light green palette is prominent. Both are now housed in the Koninklijke Bibliotheek, The Hague. One is the so-called ‘Fécamp Psalter’ (The Hague, Koninklijke Bibliotheek, MS 76 F13). It is very fully illustrated, with 36 calendar miniatures depicting the Labours of the Months, 27 full-page miniatures of the Infancy and Passion of Christ, the portrait of an unidentified lay female donor, presumably the patroness, and 11 historiated initials illustrating the major divisions of the psalter.24 Walter Cahn outlined reasons why this psalter could well have been made for use at the Augustinian house of Ham (county of Artois, diocese of Thérouanne) — the relics of St Waningus, who is prominent in the calendar, 22 Alexander, ‘Preliminary Marginal Drawings,‘ pp. 308–9, fig. 2; Stones, ‘Indications écrites et modèles picturaux,’ pp. 328–9, fig.13. See now Alexander, Medieval Illuminators. 23 Gautier d’Arras, Ille et Galeron, ed. Eley, p. XV. 24 Fully reproduced on the Koninklijke Bibliotheek website (http://www.kb.nl/ manuscripts/).

TWO FRENCH MANUSCRIPTS

377

were transferred there from Fécamp in the 9th century.25 There may also be stylistic connections to a second manuscript, the ‘Saint-Bertin Psalter’, The Hague, Koninklijke Bibliotheek, MS 76 F 5.26 Often called a ‘Picture Bible’, it has a remarkably extensive series of 45 full-page prefatory miniatures in two registers, depicting scenes from the Old and New Testaments and the Lives of the Saints. The manuscript was owned by the Benedictine abbey of Saint-Bertin at Saint-Omer, also in the county of Artois and diocese of Thérouanne, and was most likely made there. Both these manuscripts have been attributed to the last quarter of the 12th century. They are, of course, very much more copiously illustrated than the miscellany WLC/LM/6, but an early literary manuscript with illustrations is likely to have emanated from a cultural context in which distinguished devotional books were being produced for wealthy monastic and lay patrons.27 Saint-Omer is mentioned by name in the Chanson d’Aspremont (line 295 in Brandin’s edition) and Saint-Omer continued a productive centre throughout the 13th century. Many distinguished books were made there for lay, clerical, and monastic patrons alike.28 As a coda I mention another remarkable set of small rectangular miniatures in a French literary manuscript, which may stand in some kind of relationship to WLC/LM/6: the Provençal Chansonnier H, Città del Vaticano, Biblioteca Apostolica Vaticana, MS Vat. lat. 3207. It contains a series of eight portraits of trobaritz.29 They are repetitious, all showing the poetess standing in three-quarter profile and declaiming, gesturing in the direction of the adjacent text, the figures set against a patterned background in red, blue, green or yellow and outlined in black. Attributed to the end of the 13th century and to the region of Padua or Treviso, these simple illustrations, however, display a striking similarity to some of the portraits in WLC/LM/6, suggesting that a re-examination of the date and cultural context of Chansonnier H is called for. Cahn, Romanesque Manuscripts, 2: pp. 160–61, no. 134, pl. 328–32. Fully reproduced on the Koninklijke Bibliotheek website. See also Cahn, Romanesque Manuscripts, 2: pp. 165–67, no. 138, pl. 336–8 and colour pi. XVI. 27 See Stones, ‘L’enluminure au temps de Jeanne de Constantinople et Marguerite de Flandre’, p. 189. 28 Many are discussed in Saint-Omer Gothique, ed. Gil and Nys; see also Stones, ‘La Production de manuscrits littéraires aux environs de 1300: entre Cambrai et Saint-Omer.’ 29 Anglade, ‘Les Miniatures des chansonniers provençaux’; Rieger, “‘Ins e.l cor port, dona vostra faisso”’; Careri, Il canzoniere provenzale H (Vat. lat. 3207); ead. ‘Intavulare’ Tavole 25 26

378

1. Nottingham University, WLC/Lm6, Gautier d’Arras, Ille et Galeron, f. 164, Ille addressing the emperor of Rome; pen-drawn trails in bottom margin (photo: Nottingham University)

TWO FRENCH MANUSCRIPTS

2. Nottingham University, WLC/ Lm6, Benoît de Sainte-Maure, Roman de Troie, f. 14v dragon (photo: Nottingham University)

4. Nottingham University, WLC/ Lm6, Benoît de Sainte-Maure, Roman de Troie, f. 84, Hector and a squire (photo: Nottingham University)

379

3. Nottingham University, WLC/Lm6, Benoît de Sainte-Maure, Roman de Troie, f. 73, Hector and his men ride out (photo: Nottingham University)

5. Nottingham University, WLC/ Lm6, Benoît de Sainte-Maure, Roman de Troie, f. 115v Achilles raises his sword and extends his arm to two men, one also raising his sword (photo: Nottingham University)

380

6. Nottingham University, WLC/Lm6, Benoît de Sainte-Maure, Roman de Troie, f. 133v King Priam ordering his men to move the reliquaries (photo: Nottingham University)

8. Nottingham University, WLC/Lm6, Heldris de Cornuälle, Le Roman de Silence, f. 188, the author laying his hand on his book on a lectern (photo: Nottingham University)

7. Nottingham University, WLC/Lm6, Benoît de Sainte-Maure, Roman de Troie, f. 158 hybrid (photo: Nottingham University)

9. Nottingham University, WLC/Lm6, Heldris de Cornuälle, Le Roman de Silence, f. 218v Pelican ripping its breast (photo: Nottingham University)

TWO FRENCH MANUSCRIPTS

10. Nottingham University, WLC/Lm6, Fuerre de Gadres, f. 227, eagle (photo: Nottingham University)

12. Nottingham University, WLC/Lm6, La Vengeance Raguidel, f. 333v, Gauvain lays his hand on the shoulder of Greviloïne, assenting to her marriage (photo: Nottingham University)

381

11. Nottingham University, WLC/ Lm6, Fuerre de Gadres, f. 240v, Emenidus laces his chausse (photo: Nottingham University)

13. Nottingham University, WLC/ Lm6, Gautier d’Arras, Ille et Galeron, f. 185, Portrait of Ganor (photo: Nottingham University)

382

14. Rennes, Bibl. mun. 255, L’Estoire del saint Graal, f. 33v, eagle (photo: Lancelot-Grail Project)

16. Rennes, Bibl. mun. 255, L’Estoire del saint Graal, f. 100, centaur(photo: Lancelot-Grail Project) 15. Rennes, Bibl. mun. 255, Merlin, f. 101, Harrowing of Hell (photo: Lancelot-Grail Project)

TWO FRENCH MANUSCRIPTS

17. Rennes, Bibl. mun. 255, Lancelot, f. 147v, initials (photo: Lancelot-Grail Project)

383

384

18. Rennes, Bibl. mun. 255, Lancelot, f. 179 Sciapod (photo: Lancelot-Grail Project)

19. Modena, Bibl. Estense, E 39, Merlin, f. 35, Merlin before King Pandragon (photo: Bibl. Estense)

20. Rennes, Bibl. mun. 255, L’Estoire del saint Graal, f. 98v Nascien and King Mordrain at Josephé’s tomb; Celidoine and his son Narpus bid farewell to Nascien (photo: Lancelot-Grail Project)

TWO FRENCH MANUSCRIPTS

21. Le Mans, Médiathèque Louis Aragon, 354, f. 204v, initials (photo: LancelotGrail Project)

22. Paris, BnF fr 1430, Lancelot, f. 231, initials (photo: BnF)

385

386

23. Nottingham University, WLC/Lm7, L’Estoire del saint Graal, f. 73v, initials (photo: Nottingham University)

TWO FRENCH MANUSCRIPTS

24. Nottingham University, WLC/Lm6, Benoit de Sainte-Maure, Roman de Troie, f. 54, initials (photo: Nottingham University)

387

388

25. Nottingham University, WLC/Lm6, La Vengeance Raguidel, f. 314, initials (photo: Nottingham University)

TWO FRENCH MANUSCRIPTS

389

wlc/lm/7 Although it lacks historiated initials or miniatures, WLC/LM/7 is decorated with pen-flourished initials both small (three-line) and large (eight-line, with party-coloured bars in red and blue). The large initials are placed at narrative switches which match the placing of illustrated markers in other copies of the Estoire. Several of the other unillustrated Estoire manuscripts adopt a similar format of large pen-flourished initials.30 What makes WLC/ LM/7 particularly important is that, unlike the other pen-decorated Estoire manuscripts, it was written above the top line of ruling as were WLC/LM/6, Rennes 255 and Modena E 39 discussed above. WLC/LM/7 is comparable therefore to the very earliest manuscripts transmitting the long version of the story of the early history of the Holy Grail. It is one of the control manuscripts collated by Ponceau with Rennes 255, his base manuscript for the second half of the text. WLC/LM/7 is likely to be comparable in date to Rennes 255 for which I posit a date c. 1220, as noted above. The question remains as to which of the two came first, and whether, therefore, illustrations were planned from the start, or whether the set of textual markers (in the form of large decorated initials) were forerunners to full-scale historiated initials or were merely a cheaper way of including textual breaks. Both kinds of initials served to enable the reader/viewer to find his or her place and to break up the narrative into sections for reading in several stages, while the historiated initials and miniatures depicted highlights of the story as well. A comparison among WLC/LM/7 and other Estoire manuscripts dating around 1285 or earlier suggests that WLC/LM/7 was in fact a trend-setter in terms of how the text was broken into readable or viewable sections. Compare its use of large party-bar initials with those of another penflourished copy, BL MS Royal 19 C.xii, and with three illustrated copies, di canzionieri romanzi ... I ... Canzionieri provenzali, 1. Biblioteca Apostolica Vaticana, pp. 293–371; Poe, Compilatio: Lyric Texts and Prose Commentaries; Lemaitre and Vielliard with Arnould, eds, Portraits de troubadours, pp. 97–103, figs. 46–52; Peters, Das Ich im Bild, p. 23 n. 14, p. 25 n. 19, p. 26, p. 27 n. 27, p. 38 n. 85, p. 47. 30 BL MS Add. 32125; Cambridge UL, MS Additional 7071 (both English, late 13th or early 14th century); BNF MS fr 12582 began in the first quire as a pen-flourished manuscript to which miniatures were added in margins; after the first quire small miniatures in half a text column accompany party-bar pen-flourished initials (from Metz? c. 1275?); BL Royal 19 C.xii, BnF fr 747, Paris, Bibl. de l’Arsenal 2997 (all north French, c. 1250).

390

Rennes 255, Le Mans, Médiathèque Louis Aragon, MS 354 and Paris, BNF MS fr 770 (figs. 17, 21, 23). Royal 19 C.xii is of indeterminate north French provenance, perhaps Paris because its small neat script is quite like that of Paris, BNF MS fr 339 (containing Lancelot, Queste, and Mort Artu), which I attribute to Paris, c. 1250.31 Le Mans MM 354 (written by Walterus de Kayo) and BNF fr 770 are both likely to come from Douai or its region since their illustrator also worked on two important manuscripts made for use at Douai: the Martyrology and Obituary of Notre-Dame-des-Prés, Douai (OCist), Valenciennes, Bibl. mun. MS 838, and the psalter-hours of the use of Saint-Amé, Douai (OSB), Brussels, Bibl. Royale, MS BR 9391; many other manuscripts cluster around these.32 There are no places marked with a large initial or a miniature in Royal, Rennes, Le Mans, or BNF fr 770 that are not also marked by a large penflourished initial in WLC/LM/7. On the other hand Rennes, Le Mans and BNF fr 770 all lack a prominent marker at the opening of the section corresponding to WLC/LM/7, f. 1, at Molt durererent longuement cil arbres en cel color ...This is a place that is not commonly marked in later manuscripts; exceptions are Berkeley, University of California Library, MS 106 (c. 1250–75 ?), Paris, BNF MS fr 24394 (c. 1275–85 ?) and BL MS Add. 10292 (1317), but not MS Add. 10292’s sister manuscripts, BL MS Royal 14 E.iii and olim Amsterdam, Bibliotheca Philosophica Hermetica, MS 1.33 The initial of MS Add. 10292 seems unprepossessing, but it is an important marker, as this is the only case in the manuscript where a champ stands on its own without an accompanying miniature, whereas in Rennes the small champ is the type of initial generally used for paragraph breaks, so the place is not specially singled out for attention. There are two additional cases where Rennes lacks historiated initials to correspond with breaks in WLC/LM/7 (ff 43v, 52v), and one case where Royal 19 C.xii lacks a corresponding initial (between ff. 41 and 46v). Thus WLC/LM/7 has more major markers than the other manuscripts, suggesting that the others are reductions of an earlier, more fully marked, format. At the See La Légende du roi Arthur, ed. Delcourt, cat. no. 72. Ibid., nos 34 and 35, with reference to dating arguments made more fully in Les Manuscrits de Chrétien de Troyes. 33 Other manuscripts that mark this place in the text are (in chronological order) BL Add. 32125; BnF fr 19162, fr 95, fr 1426. See Stones and Kennedy, ‘Signs and Symbols in the Estoire del saint Graal and the Queste del saint Graal’, pp. 158–59, reprinted in these essays. See also ‘Un schéma d’emplacement’ reprinted in these essays. 31 32

TWO FRENCH MANUSCRIPTS

391

same time Rennes, Le Mans and BNF fr 770 offer a more extensive level of decoration for the markers they do include, so that one or both of the final episodes at Josephé’s tomb and Celidoine’s departure are marked in WLC/ LM/7 and Royal and illustrated in Rennes and BNF fr 770, and marked with large champs in Le Mans (figs. 20, 21). Le Mans has less historiation than either Rennes or BNF fr 770, using champs in eleven cases where Rennes and BNF fr 770 give a historiated initial or miniature. This suggests that Le Mans was a cheaper product, but one which nonetheless respected established section breaks. It remains to correlate these findings with the other unillustrated and illustrated copies. Several later copies, notably BNF MS fr 344 (late 13th century); olim Amsterdam, BPH MS 1; and BL MS Add. 10292 (dated 1317), BNF MSS fr 105, 9123 (c. 1320–30), include more illustrations than the early manuscripts referred to here. They pictorially emphasize parts of the text which are left without markers in WLC/LM/7 and the other early copies, such as Josephé’s second Grail Mass and the King Crudens/King Agrestes events in England and the final episode about the murder of King Lancelot. This suggests that a simpler set of markers was devised at the start and that the makers or patrons of later copies elaborated upon it in various ways, expressing interest in different parts of the text through their pictorial choices. These are preliminary findings, however, and full confirmation of the sequence of development in Estoire illustration would require consideration of the other copies, illustrated and unillustrated, a task which is beyond the scope of the present essay.34 So far the results suggest that careful thought was given from the beginning to the question of marking, with WLC/LM/7 positioned at that beginning. Where and when WLC/LM/7 was produced is not an easy matter to determine. It contains only one hint among the marginal notes — the tantalizing reference to the Abbess of Fontevrault (Fig. 10). Was the inscription an indication that the book was a loan from the abbess to the person who wrote the note, or had it written? Why was the note written on a middle folio and not on a first or last page, or a flysheet? Questions remain. Monastic manufacture has been suggested for Berkeley 106 as noted above, painted perhaps by an artist of Jumièges (OSB), Normandy, c. 34 Illustrated Estoire manuscripts dating before c. 1275 are Oxford, Bodl. MS Douce 303; Chantilly, Musée Condé, 476(644), which also includes an Image du monde written by Jaquemin d’Acre in 1271; and Berkeley, University of California Library, 106 (see ‘Signs and Symbols’, pl. 12).

392

1275.35 I further suggest that the pen-flourishing in an early (written above top line) copy of the Lancelot, BnF, MS fr 1430 (fig. 22), might have links to production and/or ownership by the Abbey of Pontigny (OCist).36 More might eventually be made of an ecclesiastical context for the making and owning of the Lancelot-Grail manuscripts. So it is worth drawing attention to WLC/LM/7 and its decoration, in the hopes that eventually the dating, placing, and understanding of the manuscript and its position in the tradition of the Estoire and the Lancelot-Grail romance may emerge more clearly.

The attribution is by Avril, reported in ‘Signs and Symbols’. For Pontigny manuscripts, see Peyrafort-Huin, Stirnemannn and Benoit, La Bibliothèque médiévale de l’abbaye de Pontigny. 35 36

TWO FRENCH MANUSCRIPTS

393

Appendix The Illustrations of WLC/LM/6 Illustrations are numbered continuously, with a following folio number to indicate the place of the illustrations. Editions are listed at the beginning of each text. Many of the images are damaged where paint has either flaked or been removed at an uncertain date; this is especially unfortunate on the faces of figures. Benoît de Sainte-Maure, Roman de Troie (fols 1ra–150rb) Line references are to Le Roman de Troie par Benoit de Sainte-Maure, ed. Constans; summaries follow Jung, La Légende de Troie en France au Moyen Age, pp. 124–33. 33 half-column miniatures in coloured frames, of which four contain decorative motifs. Heraldry occasionally is depicted with a correct combination of metal (with white for argent) and colour; but is often shown as colour on colour, disregarding basic rules of heraldry, and is therefore described below in general terms. No title. 1 f. 12 Hesione, wearing a white robe and orange cloak, stands holding a flowering branch (v. 2183: Hercules sets out against Troy). 2 f. 14v (fig. 2) A dragon with tail ending in a foliage motif, (v. 2601: In the battle outside Troy, Castor is assisted by Pollux). 3 f. 20v Hector stands before King Priam, addressing him (v. 3771: Council of the Trojans). 4 f. 40v Hector stands wearing a red tunic and holding a blue cloak over his arm, gesturing towards the text (v. 7641: List of the chieftains of the second battle). 5 f. 44 Hector couching his lance, riding to the right, holding a blue and white shield, his horse’s housing orange with three black lions rampant (v. 8329: Hector sets out to kill Patroclus). 6 f. 55v Blue bear on gold ground (v. 10561: Beginning of the third battle).

394

7 8 9 10

11 12

13 14 15 16 17

f. 66v Knight riding out of city of Troy (v. 12691: Beginning of the sixth battle: a single knight stands for the Trojan army: ‘Mais de la vile sen eissierent/E cil de lost les recoillirent...’). f. 68v Priam’s forces (three men, representing ‘li duc, li prince et li conroi Prianz’) ride out (v. 13065: Agreement between Greeks and Trojans resulting in the exchange of Antenor for Thoas). f. 69 Hector and Achilles, on horseback, address each other, raising hands in speech (v. 13135: Interview between Hector and Achilles). f. 73 (fig. 3) Hector (housing and shield argent [white] semé of lions sable) and two of his men ride out of Troy, all holding their lances aloft, with small white banners attached (v. 13907: Beginning of the eighth battle). f. 78v Hybrid with three heads: serpent, bird, human head, wearing a hood (v. 14977: Truce before the ninth battle). f. 84 (fig. 4) Hector (surcoat, shield and housing argent [white] semé of lions sable) and a squire riding; the latter adjusts the helmet of the former (v. 16007: [H]ector escoute et or et uoit... en son cief a son heaume asis/lacie ia uns damoisiaux.’ This, I think, is a literal depiction of the following lines describing a damoisel lacing up Hector’s helmet and not, as Jung thought (p. 129), an illustration of the death of Hector according to the Histoire ancienne model as depicted in BnF, MS fr 20125 f. 133v, where Hector leans over to remove the helmet of a knight he has killed and is speared in the back by Achilles as he does so. Cf. BnF fr 20125, reproduced on the BnF website, Mandragore). f. 85v King Priam, holding a sceptre, points to the text (v. 16317: Beginning of the lamentations). f. 88v Palamedes and two companions riding; Palamedes points (v. 16881: Council of the Greeks and election of Palamedes). f. 92 Winged mermaid (v. 17489: Anniversary of the death of Hector; Achilles falls in love with Polixena. Jung (130) wonders if the siren might be a reference to Polixena). f. 98v Paris, wearing a petal-shaped helmet, aims his bow; no adversary is shown (v. 18841: In the middle of the 12th battle, Paris has just killed Palamedes with an arrow). f. 100v Two knights ride. The one in front has a housing orange with two black battle-axes and couches his lance; the other holds his lance vertically, with a small white banner attached; his shield is azure a

TWO FRENCH MANUSCRIPTS

395

fleur de lis argent [white] (v. 19207; Beginning of the 13th battle). 18 f.106 Two groups of Greeks addressing each other, one extending his arm in speech (v. 20341: Council of the Greeks before the 16th battle). f. 107 In the bottom margin, a box-like drawing in leadpoint with eight compartments with a standing figure (?) in one of them. 19 f. 109 Achilles, arm extended, pointing with his finger, addressing a man who listens, raising his hand (v. 20865: This corresponds only generally to the text since there is no mention of a particular conversation between Achilles and another person; however, the image could refer to Achilles ordering the dead to be buried, as in lines 20875–8, ‘... seveli furent/Solonc lor lei, si come il durent./N’i remest cors a enterrer/Ne chans ne place a délivrer’). 20 f. 113 Inside Troy, shown as a crenellated structure, two men [sic] grieve, holding heads on hands (v. 21687: Priam grieves; Hecuba laments Troilus). 21 f. 115v (fig. 5) Achilles raises his sword and extends his hand draped in a green cloth to two men (Antilocus, who also raises his sword, and Paris? or vice versa) (v. 22187; Death of Achilles: the picture illustrates lines 22189 90, ‘Son braç molt tost et molt isniel/A ien entors de son mantiel...). 22 f. 118 Paris rides, holding a green shield, raising his lance which has a long white pennant attached (v. 22635: Paris rides out to the 20th battle, in which he will be killed). 23 f. 121v King Priam tells Queen Penthesilea (grieving, head on hand), that her children are dead (v. 23395: Penthesilea’s grief ). 24 f. 124 Two Greek knights ride, holding lances with pennants raised; the one in front carries a shield argent [white] a bend gules (vv. 23843–4: \... Bien ordené e bien devis/Se sont tuit a la veie mis’). 25 f. 126v The Author Dictis seated, writing (v. 24397: Dictis rewrites Dares’ story). 26 f. 128v King Priam’s men ask the king to make peace (v. 24801: Priam and his men discuss terms for peace). 27 f. 133v (fig. 6) King Priam orders three men to move the reliquaries (‘les santualires’) (shown as a cross on a column/altar) (v. 25809: The reliquaries transported out of the city, an exceptional image, as noted by Jung (p. 131)). 28 f. 136 Cassandra, in white, holding a ring (?) stands before

396

29

30

31 32 33

Agamemnon who sits cross-legged holding a sceptre (his crown defaced) (v. 26299: Cassandra given to Agamemnon). f. 137v The Author, pointing to the words of the text (cf. Dictis on f. 126v) (v. 26591: ‘[E]nsi com io uos ai conte/Onques en lestoir ai troue ...’. After the death of Hecuba, the introduction of the disputation of the Palladion). f. 144 Four men riding: Ulysses, Diomedes and Palamedes and a fourth man riding towards the well in which Palamedes will be stoned (v. 27847: The miniature is placed in the middle of the account given to Nauplus of the death of his son Palamedes, illustrating lines 27849–51). f. 146 Orestes rides back to Athens (badly rubbed) (v. 28295: Orestes returns to Athens). f. 147v King Menelaus gives Hermione (wearing white) to Orestes (seated), in marriage (v. 28533: King Menelaus recognises Orestes’ valour). f. 153v A man (a devin), sitting cross-legged, listens to Ulysses telling his dream (v. 29815: Beginning of the second and last episode of Ulysses, illustrating v. 28923).

As Jung has noted, there is no interest in depicting full-scale battles, and the deaths of heroes are similarly absent, represented only by a lead-in scene (in the case of Hector) or a follow-on scene (Paris with his bow, having killed Palamedes, whose death is not shown). On the other hand, the detail of the helmet-attaching on f. 84 and the lead-in to the death of Achilles on f. 115v are instances where the pictures follow exactly what the text says. The ‘io’ of line 26591 (f. 137v) is rendered as an author-portrait. Notable is the Christianisation of the sanctuary motif on f. 133v. Gautier d’Arras, Ille et Galeron (ff. 157ra–187va) Line references are to Gautier d’Arras, Ille et Galeron, ed. Cowper. List of subjects of the illustrations on p. xv, n. 3. No title. 34 f. 157 Agnus with cross staff and banner with small red cross; orangered frame (line 1). 35 f. 158 (fig. 7) Male hybrid holding small black bow (not Ille as a

TWO FRENCH MANUSCRIPTS

397

child, as per Cowper; line 205). 36 f. 160 The vilein Rogelyon as a knight riding, couching his lance and holding a shield orange a C-shaped crescent white; his horse’s housing azure semé of crescents (C-shaped) argent [white] (line 591). 37 f. 164 (fig. 1) Ille addressing the emperor of Rome, who stands, pointing a finger at him, holding a loop of his vair-lined cloak with his right hand and a fleur-de-lis sceptre in his left; Ille asks for a modest place in the Roman army (line 1342). 38 f. 170 Ille and Ganor, daughter of the emperor, holding hands (line 2496). 39 f. 175v Ille, now appointed seneschal, asks permission of the emperor (seated, holding sceptre) to return to Brittany to search for Galeron (line 3354). 40 f. 185 (fig. 13) Portrait of Ganor, wearing a ringlet in her hair and a vair-lined cloak (line 5392). As Cowper has noted, the illustrations privilege the Roman component of the story and Ille’s relationship with Ganor, to the exclusion of Galeron. Heldris de Cornuälle, Le Roman de Silence (ff. 188ra–223rb) Le Roman de Silence, Thirteenth-Century Arthurian Verse Romance by Heldris de Cornuälle, ed. Thorpe, from this unique manuscript; the frontispiece reproduces f. 188, and the subjects of the illustrations are described on pp. 6–8. No title. 41 f. 188 (fig. 8) The author Heldris de Cornuälle seated, one hand on open book on desk, the other holding fold of cloak; wearing a grey hat. Guide letter C [sic] in left margin (line 1) 42 f. 195v The Count of Chester addressing Eufemie, daughter of Count Renalt de Cornuälle and handmaid of Queen Eufeme, wife of King Ebains. He holds a red fleur-de-lis sceptre and raises his hand to Eufemie who is robed in white with a ringlet in her hair, holding the cord of her blue cloak (line 1433). 43 f. 199 This badly rubbed miniature probably shows the nurse wearing white and holding the baby Silence, girl child of Cador and Eufemie who is to be brought up as a boy if Cador is to retain his

398

44

45 46 47 48

49 50

51 52

rights to the County of Cornwall, facing Eufemie who points her finger — or vice versa (line 2127). f. 201 Thorpe (p. 7) read this as two men, possibly the seneschal and Cador, in conversation, presumably about Silence being brought up with the seneschal’s son. One figure (face rubbed) is seated, holding the fold of his cloak and pointing a finger at the other figure who stands. However, the miniature precedes the passage where Nature and Environment debate this decision and disagree with each other, so an alternative reading would be the two personified (although such personifications would normally be shown as female figures). Shortly after this, Cador explains to Silence why he has had her brought up as a boy, so a third possible reading is that the seated figure is Cador and the standing one is Silence dressed as a boy; against this is the depiction of Silence as a child in the following image (fol. 203r), although line 2500 gives her age as 12 years old at this point (line 2497). f. 203 Two (female ?) jongleurs with Silence, one of whom holds her hand. Here she is shown as a child (line 2921). f. 206v Cador, crowned, standing, holding the fold of his cloak, addresses an old man (or is this Silence dressed as a man, on her return from the jongleurs ?) (line 3593). f. 209 Queen Eufeme and Silence sitting together on bench; Eufeme addresses Silence with whom she has fallen in love (badly rubbed) (line 4027). f. 211 Silence, her foot outside the picture frame, kneels and hands a scroll to the King of France; this is the letter originally written by King Ebains asking that Silence be well received at the French court, but for which Queen Eufeme has substituted a letter requesting the immediate execution of Silence; the exchange of letters is revealed upon the king’s request to Ebains of confirmation of its contents (line 4459). f. 213 Lion (not a dog as per Thorpe) with flowering tail (line 4781). f. 214v King Ebain’s chancellor, his foot on the miniature frame, explains to the King how the letter was exchanged. The chancellor holds emblems of his office — a staff with a belt and small pouch attached to it (line 5069). f. 217 Griffon biting its tail (line 5557). f. 218v (fig. 9) Pelican ripping its breast (but not, as is usually the

TWO FRENCH MANUSCRIPTS

399

case, feeding its young); as a symbol of Christ, this image relates to the following lines which refer to God (line 5879). 53 f. 221 Merlin as man (badly rubbed) sitting on red mound; Merlin is playing a madman and is asked to explain his laughter or his head will be chopped off. (He laughs for several reasons, including that Silence is a woman in disguise and that a disguised nun is Queen Eufeme’s lover.) (line 6299). 54 f. 222v Silence stands naked before King Ebains, whom she will marry once Queen Eufeme and her disguised nun-lover have been executed (miniature badly rubbed) (line 6582). C’est d’Alixandre (Le Fuerre de Gadres) (ff. 224–243vb) The Medieval French ‘Roman d’Alexandre’, ed. Armstrong et al, esp. vols II and V, to which reference is given, following Ross, The Illustrations of the Old French Alexander in Verse. It should be noted that this manuscript gives a variant version in which the laisses are sometimes out of sequence in relation to Armstrong’s edition. 55 f. 224 Alexander crowned, enthroned, facing frontally (face rubbed), seated on a faldstool with dragon-head knobs and claw feet (II, 74, laisse 1, line 1). 56 f. 227 (fig. 10) An eagle, wings spread, head turned to the right, possibly a theriomorphic capital (II, 84, laisse 23, line 488). 57 f. 229v Gadifer riding to the right, wearing mail (showing the lacing of the ‘chausses’, a pot-helm, prick spurs, and a shield shown from the inside). The horse’s housing is striped red and yellow with a narrow white line between the stripes (II, 100, laisse 52, line 1189). 58 f. 231 Salatin riding to the right, in full armour, holding a shield gules with an erased charge. The horse’s housing is argent semé of martlets sable (II, 91, laisse 38, line 787). 59 f. 232v A grey-blue canine creature, male and with a bushy tail and lop ears, advancing towards the left, looks back to the right; possibly a theriomorphic capital (V, 20, Version B, laisse 61, line 1). 60 f. 234v The riderless horse of Duke Betis of Gadres (unhorsed in the two previous laisses), with reins broken and trailing, a saddle and cloth, and a white housing with a faint pattern of two leaves hanging from a central stalk (II, p. 112, laisse 73, line 1740).

400

61 f. 237 Gadifer again, riding to the right, holding a shield gules a bordure argent, in dexter canton a cinquefoil voided argent (not the same charge as on fol. 229v). His horse’s housing is the opposite: argent semé of cinquefoils voided gules (V, 46, Version B, laisse 85, line 1). 62 f. 239 A grotesque green two-legged monster with claws, a canine head and foliate prolongations like a forked tail, looking back over its shoulder to the left (II, 106, line 1476). 63 f. 240v (fig. 11) Emenidus, wearing full armour, seated on a green bank, faces left and laces up the mail of his left chausse (V, 60, Version B, laisse 104, line 1). As Ross noticed, this depends directly on line 8 of the text, ‘A terre est descendus pour ses cauces lachier ...’. 64 f. 243 Emenidus (face rubbed), wearing civilian clothes of a long green robe and a red cloak, holds out his left hand, wrapped in his cloak, and makes a gesture of speech with his right hand; at this point Emenidus urges Alexander to leave him and the other wounded behind and ride on and take vengeance on the enemy (V, 82, Version B, laisse 124, line 1). As Ross noted (The Illustrations), miniature 9 is the only one that really depicts what the text says. Furthermore, Ross detected no iconographical parallels with any other illustrated version of Branch II of the Roman d’Alexandre, nor with any other illustrated manuscripts of this part of the Roman. La Chanson d’Aspremont (ff. 244va–303vb) La Chanson d’Aspremont, ed. Brandin. Brandin gives a list of miniatures but does not identify the subjects. Title: C’est d’Iaumont et d’Agoulant 65 f. 244v Charlemagne enthroned holding sword and orb (laisse 1). 66 f. 246v Standing youth holding fold of cloak, hand to face. Cf. BnF MS fr 2495, f. 71, red capital A (laisse 23: Charlemagne entertains Balan, the envoy of the pagan king Agolant, at dinner). 67 f. 253 Two knights ride, one in helmet, one in chain mail coif. The one in front has an orange surcoat, a shield argent [white] a fess gules, a housing orange semé of cinquefoils sable; the other has a green surcoat and housing. Cf. fr 2495, f. 88, with variant (laisse

TWO FRENCH MANUSCRIPTS

68

69

70

71

72 73 74 75 76 77

401

96: Charlemagne and his men [‘Cent milïomes’, not two] ride to Aspremont). f. 255 Between two trees, Duke Namles stands in front of his horse, which dips its head behind frame (drinking?) (laisse 117: Namles dismounts under a tree at night time; he encounters bear cubs and a bear, which he kills; another bear and a leopard appear; he kills the leopard, whereupon the bear runs off). f. 257 The Saracen Gorhans leads Namles before the Saracen king Agolant, who sits with legs crossed, pointing at them; cf. fr 2495, f. 97, D, red capital (laisse 138: Having spared the life of Gorhans, Namles is led by him before Agolant). f. 259 (Page is badly torn and sewn; miniature badly erased.) Namles stands before Charlemagne, recounting his adventures at Aspremont. Cf. fr 2495, f. 102v S, red capital (laisse 153: Namles presents Charlemagne with the claw of the huge bird that had attacked his horse; Jeanroy, ‘Sur un épisode de la Chanson d’Aspremont’, derives this episode from the Descriptio itineris Hierosolymitani). f. 260v Knight on horse, couching lance, housing argent [white] semé of fleurs de lis (rubbed) azure. To judge by the heraldry, this is one of Charlemagne’s knights, representing the army who at this point advances towards King Aumons (laisse 165: Aumons asks Hector, king of Val Penee, whose army is advancing towards them). f. 269 Griffon with flowery tail (laisse 262: Ansquetin the Norman comes to the rescue of Ogier, who is fighting Aumon). f. 270v Knight rides, lance held diagonally in hand, housing argent [white] bendy sinister sable; presumably this is Girars, who delivers an encouraging speech to his men in laisse 272. f. 274 Hybrid man wielding large axe, wearing red cap (laisse 294: Aumon son of Agolant weeps for his dead men; one of them, Sinagon, is wounded by Namles). f. 277v King Manuel, seated frontally holding sceptre; his other hand holds the fold of his cloak as he pleads for peace (laisse 322). f. 282v Charlemagne standing, holding sceptre and fold of cloak (laisse 363: Guitekins suggests a strategy to Charlemagne). f. 288 Seated man (Agolant) sits cross-legged, holding fold of garment and pointing to standing knight wearing chain mail and helmet and holding his lance point up onto the frame of the miniature; Agolant is here addressing his (dead) son (laisse 419).

402

78 f. 294 Roland and Graelans ride, holding a white banner (‘Cascuns d’aus ot sa lance sus levee/La grant ensegne avoit desvelopee...lines 9509–10); the rider in front has a housing orange semé of fleurs-delis sable (laisse 463: Roland and Graelant enter the battle; Graelant attackes Calides). 79 f. 299v The author, standing, hand on hip, pointing to text (‘[S] egnor baron plaist uos a escoter/Por qoi l’en doit le duc Girart amer ...’) (laisse 508: because Girart founded an abbey [Vézelay is not named] and had the body of Agolant buried in the palace and the head, still in its helmet, embalmed so that it could be viewed). La Vengeance Raguidel (ff. 304ra–335vb) Raoul de Houdenc, La Vengeance Raguidel, ed. Roussineau, based on WLC/ LM/6. Illustrations listed on p. 66 and reproduced on pp. 138–41. Title: De roi Artut 80 f. 304 Kings Engenor and Aguisait addressing King Arthur who sits holding a sceptre, other hand resting on his knee, his legs uncrossed (l.1). 81 f. 318 Gauvain and le Chevalier Noir conversing together (l. 2703). 82 f. 328 Hooded male hybrid with long tail and wings (l. 4595). 83 f. 333v (fig. 12) In a crenellated room, Gauvain, holding a sceptre, lays his hand on the shoulder of Greviloïne, thus expressing his assent to her marriage to Yder. (l. 5621; cf. lines 5934–6; as noted by Roussineau [67], Gauvain holds a sceptre as the future inheritor of his uncle, King Arthur). Gautier le Leu, Fabliaux (ff. 336r–345vb) Ending incomplete. Unillustrated.

XX Un schéma d’emplacement pour l’illustration de l’Estoire del saint Graal et les débuts de la tradition manuscrite

Q

uant Lancelot, emprisonné par Morgain la soeur du roi Artur, regarda par la fenêtre de sa prison, il vit un artiste en train de peindre l’histoire de Troie sur les murs d’une piece avoisinante. Cela lui donna l’idée qu’il pourrait lui-même peindre l’histoire de son amour pour la reine Gueniève et raconter comment leur passion s’était déroulée. Les images étaient tellement réalistes que Lancelot baisa la peinture de la reine. La seule représentation de ce moment se trouve dans un manuscrit confectionné probablement à Saint-Omer en 1317.1 Là par contre l’image montre Lancelot embrassant une statue de Guenièvre et non pas une peinture ; l’artiste connaissait sans doute la légende de Pygmalion d’après les Métamorphoses d’Ovide (livre X) mais la tradition manuscrite d’Ovide ne transmet pas d’images de ce sujet avant le XIVe siècle. On voit aux environs de 1300 d’autres reflets de ce sujet en images, par exemple dans le Roman de la Manekine où une statue de l’héroïne et de son enfant est jetée dans le feu à la place de l’héroïne ellemême ou encore au XIVe siècle dans le Voir-dit de Guillaume de Machaut où la statue de Toute Belle se met à pleurer. Quant au topos de Troie, il a été repris par Christine de Pizan qui contemple des images de l’histoire de Troie peintes sur les murs de la Salle des merveilles dans un exemplaire du Livre de la mutacion de fortune. D’autres aspects de l’iconographie du

Publié dans Mémoires arthuriennes, éd. D. Quéruel (Troyes: Médiathèque du Grand Troyes), 2012, pp. 71–87. 1 Londres, BL Add. 10293, f. 325v, reproduit dans Stones, ‘Seeing the Walls of Troy.’

404

Lancelot reflètent une fascination avec la destruction de Troie.2 Dans la dernière branche du Lancelot-Graal, la Mort Artu, Morgain trahit Lancelot en montrant ses peintures au roi Artur qui, pour la première fois, grâce aux images, se rend compte que les rumeurs concernant l’adultère dont il est victime étaient fondées. Là encore, les images illustrant ce moment sont rares dans les manuscrits de la Mort Artu – on les retrouve dans les deux manuscrits que Jacques d’Armagnac a commandité dans les années 1470 et 1475 (Figs. 1 et 2), Mais la leçon du texte, même si elle n’est pas souvent illustrée, met en avant la puissance de l’image au Moyen Âge. Dans la tradition manuscrite du Lancelot-Graal en effet les images jouent un rôle capital, séduisant même, comme en témoigne la remarquable exposition organisée en 2009 par Thierry Delcourt à qui nous rendons hommage ici. C’était un assemblage sans précédent de manuscrits et d’objets arthuriens faisant suite à l’exposition de Rennes de l’année 2008 et suivie en 2011 par l’exposition de Troyes – une trilogie d’expositions qui privilégiaient surtout et à juste titre les manuscrits enluminés. Comment l’emplacement des images dans les textes a-t-il été déterminé ? Les manuscrits du Lancelot-Graal ne sont pas comme la Bible, où une illustration en tête de chaque livre était une formule simple et de rigueur. La structure du Lancelot-Graal par contre est beaucoup plus fluide et n’offre pas une simple solution. En effet, la gamme des illustrations et leur emplacement dans le texte n’est pas standardisé. On retrouve rarement une sélection identique d’images, même dans des exemplaires sortis des mêmes ateliers.3 Ici nous revenons sur les débuts de la tradition manuscrite et en particulier sur les plus anciens manuscrits de l’Estoire del saint Graal afin d’examiner quels sont les premiers schémas d’illustration et quels rapports ils peuvent avoir avec la décoration secondaire, non-figurative. Le manuscrit conservé à Rennes Bibl. mun. 255 est à notre avis le plus ancien manuscrit illustré du LancelotGraal, contenant L’Estoire, le Merlin et le début du Lancelot propre, dont l’Estoire contient 31 initiales historiées.4 Nous avons proposé une datation aux environs de 1220 et une provenance parisienne ou champenoise. D’autres exemplaires à peu près contemporains ne contiennent pas d’initiales historiées 2 Par exemple la destruction des nombreux châteaux : celui du duc qui tua le roi Lancelot, aieul de Lancelot du Lac, celui de Trèbes au début de la branche du Lancelot, celui de la Fausse Gueniève. 3 Parmi les exceptions on peut citer les manuscrits jumeaux Paris, BNF fr 19162 et 24394 (reproduits sur Mandragore, comparaisons sur http://www.lancelot project.pitt.edu). 4 Stones, ‘The Earliest Illustrated Prose Lancelot Manuscript ?’, réimprimé dans ces essais.

UN SCHÉMA D’EMPLACEMENT

405

mais sont ornées de grandes initiales champies ou filigranées à corps mi-parti de bleu et de rouge (souvent appelées “puzzle initials”) qui fournissent des étapes-repaires dans les manuscrits. Quel rôle cette décoration secondaire a-t-elle pu jouer dans la formation de la série d’illustrations présente dans le manuscrit de Rennes et dans d’autres manuscrits illustrés ? Pour répondre à cette question nous analysons d’après une table comparative (à la fin de cet article) deux manuscrits sans initiales historiées contenant de grandes initiales filigranées (Nottingham, Université de Nottingham, WLC/LM7 et Londres, BL Royal 19 C.xii) qui tous les deux se placent très tôt dans la tradition manuscrite5 et deux manuscrits illustrés (Le Mans, Méd. mun. 354 et Paris, BnF fr 770, datant probablement des environs de 1285) à côté du ms. de Rennes.6 Parmi les études portant sur la décoration mineure nous citons les articles par Françoise Gasparri, Geneviève Hasenohr et Christine Ruby et par Roger Middleton dans Les manuscrits de Chrétien de Troyes.7 Toutefois les manuscrits LM7 et Royal ont été confectionnés vraisemblablement une génération avant les manuscrits survivants de Chrétien. Comme le manuscrit de Nottingham ne contient que la deuxième partie de l’Estoire nous limitons notre enquête à cette partie. Parmi les 5 manuscrits comparés, c’est le ms LM7 qui présente le plus ample schéma décoratif comportant 19 grandes initiales filigranées dans cette deuxième partie de l’Estoire del saint Graal. Des différences de choix entre les cinq manuscrits apparaissent dès la première comparaison. LM7 (fig. 3a) et Royal ont chacun une grande initiale pour le passage commençant Mout dura longuement... alors que les trois manuscrits illustrées n’ont qu’une petite initiale de paragraphe. 5 Ponceau, L’Estoire del saint Graal, 1, xxvii, xlvii (siglum Not., version longue, membre isolé de la famille y); xlviii (se rapproche textuellement de Rennes BM 255 et de Cambridge UL Add. 7071). Le ms LM7 a été copié au-dessus de la première ligne de réglure tout comme Rennes 255. Selon Ponceau il appartient comme Rennes au plus ancien groupe de manuscrits. Pour sa provenance et la mention ‘de par madame l‘abbesse de fontderaut’ (ou ‘founderaut’)’ au f. 29v dans la marge inférieure, voir Hanna, ‘The Catalogue,’ dans The Wollaton Medieval Manuscripts, 99–100, fig. 10. Royal 19 C.xii, copié au-dessous de la première ligne de réglure, doit être légèrement plus tardif mais la série d’initiales filigranées qui s’y trouve se rapproche de près de celle de LM7. Dans le ms de Rennes a été inaugurée la série d’initiales historiées, placées à tous les points marqués dans LM7 à l’exception de deux emplacements au début de la deuxième partie du roman, puis reprise en partie dans les autres exemplaires. 6 Un résumé des données se trouve dans Stones, ‘Two French Manuscripts: WLC/LM/6 and WLC/LM/7,’ dans The Wollaton Medieval Manuscripts (cit. n. 5), pp. 47–49. 7 Gasparri, Hasenohr, Ruby, ‘De l’écriture à la lecture’ dans Les manuscrits de Chrétien de Troyes, I, pp. 97–148; Middleton, ‘Coloured Capitals in the Manuscripts of Erec et Enide.’

406

Il s’agit pourtant d’un moment culminant dans l’histoire des couleurs de l’arbre de vie, blanc lorsque Eve l’a emmené hors du Paradis, devenu vert lorsque Dieu commenda à Adam et Eve d’y faire l’amour, transformé enfin en rouge à la suite du meurtre d’Abel, et l’arbre demeura rouge pendant longtemps, dit le texte, avant que Salomon utilise son bois, et le bois des souches vertes et blanches, pour faire construire sa nef et ses fuissiaus.8 La place centrale de ce moment n’a pas été toujours mis en avant dans le reste de la tradition iconographique non plus. Il a été reconnu et marqué par une petite initiale historiée dans le manuscrit conservé à Berkeley, UCB 106, au f. 145v (fig. 3b) et dans les manuscrits ‘jumeaux’ Paris, BnF fr 19162, au f. 65v (fig. 4,9 et Paris, BnF fr 24394, au f. 50v (fig. 5) ainsi que dans Paris, BNF fr 344 au f. 43 (fig. 6) et Paris, BnF fr 95 au f. 49v (fig. 7). Parmi les manuscrits qui incluent une image de l’arbre, on peut constater deux groupes : dans fr 24394, 19162, 95, l’arbre seul prent toute la place et dans fr 24394 et 19162 il change de couleur, de blanc en vert, alors que dans fr 95 trois arbres à couronnes de feuillage de types différents, toutes peintes en vert surmontent deux troncs rouge-orange et un tronc bleu. Dans Berkeley 106 et fr 344 des personnages sont présents, trois dans Berkeley qui se tiennent devant l’arbre; dans fr 344 deux groupes de personnages contemplent un arbre à feuillage vert, peut-être font-ils partie de la cour de Salomon qui va utiliser ce bois pour sa nef symbolique. Derrière les arbres dans Berkeley et fr 344 se trouve un tasseau, évoquant peut-être le bois de la croix du Christ dont de nombreuses légendes circulaient au Moyen Age.10 Par ailleurs dans la tradition manuscrite de l’Estoire, l’arbre de vie ne suscitait pas d’intérêt puisqu’aucune image de l’arbre n’est inclue. C’est le cas du manuscrit conservé à Bonn, LUB 526, copié en 1286, de BnF fr 110 (vers 1295). Il en est de même dans les manuscrits du XIVe siècle : seulement dans Londres, BL Additional 10292, confectionné à Gand ou à Saint-Omer en ou après 1317, une initiale champie tient lieu, exceptionnellement, d’une miniature au f. 33v (fig. 8), alors que dans les deux manuscrits apparentés, Royal 14 E.iii et olim Amsterdam BPH 1, aucune initiale ne marque cette place. L’arbre n’avait apparemment pas d’intérêt pour les commanditaires des 8 Nous résumons cette partie de L’Estoire del saint Graal dans Stones et Kennedy, ‘Signs and Symbols in the Queste’, pp. 157–61. 9 Ibid., Figs. 9, 10. 10 A notre connaissance aucune légende de la croix ne mentionne les trois fuissiaux et le texte de l’Estoire de son côté ne mentionne pas la croix, voir Stones et Kennedy ibid., p. 160.

UN SCHÉMA D’EMPLACEMENT

407

manuscrits parisiens non plus puisqu’il manque aussi dans Paris, BnF fr 105 et 9123, émanant du cercle du Roman de Fauvel aux environs de 1320–1330. Il fait défaut aussi dans les manuscrits du XVe siècle. Dans Rennes, Le Mans et fr 770 tout l’intérêt se porte sur la curiosité de Nascien envers la nef et ses fuissiaux (n° 2 dans la Table à la fin de cet article) et c’est un moment marqué aussi par une grande initiale dans LM7 et Royal (fig. 9–12). Puis la décoration passe à l’aventure de Celidoine fils de Nascien, emporté par des mains (no 3 dans la Table). Cette fois l’endroit est marqué par une grande initiale décorative dans tous les manuscrits à l’exception de Rennes. La suite de l’aventure – Celidoine mis en mer dans son bâteau par les hommes du roi Laban (n° 4 dans la Table) – reste sans décoration dans Rennes, marqué par une petite initiale dans Royal, mais par une grande initiale dans LM7 (fig. 13) et par une image figurée dans Le Mans (fig. 14) et fr. 770 (fig. 15). Donc parmi les trois plus anciens manuscrits LM7 a été le pionnier une deuxième fois en incluant une grande initiale filigranée à cet endroit dans le texte. Puis le récit se concentre sur Flegentine la femme de Nascien et les aventures des messagers qui partent à sa recherche. Tous les manuscrits reconnaissent l’importance de cette partie du récit (initiale historiée dans Rennes, Le Mans, fr 770, de grandes initiales dans LM7 et Royal (nos 5–8 dans la Table), s’arrêtant sur Hippocras qui guérit le neveu de l’empereur Auguste (n° 6) une initiale historiée dans Rennes, une champie dans le Mans, une grande initiale filigranée dans fr 770, LM7 et Royal) ; la découverte du tombeau d’Hippocras (n° 7, des initiales historiées dans Rennes et Le Mans, une miniature dans fr 770, des grandes initiales filigranées dans LM7 et Royal); n° 8, la lamentation de Flegentine (Rennes, fr 770), Flegentine intérrogeant les messagers (Le Mans), des grandes initiales filigranées (LM7, Royal). Dans tout le reste de l’Estoire – encore 11 passages (nos 9–19 dans la Table) sont marqués par une lettre décorée ou historiée – les cinq manuscrits marquent les mêmes passages dans le texte (à l’exception de Royal où la fin du récit, soit 5 passages, manque et la dernière page de fr 770 manque aussi). Par contre les manuscrits plus tardifs ajoutent une ou plusieurs scènes supplémentaires à ce noyau et introduisent d’autres sujets dans le corps du noyau meme – Nascien tuant le roi de Northombrie ; le roi Mordain et ses compagnons en bateau vers l’Angleterre ; la réunion entre le roi Mordrain, Nascien, Flegentine, Célidoine ; la messe de Josephé ; l’histoire du roi Agrestès, plus tard les tombeaux de Chanaam et de ses frères et la conversion du roi;

408

Simeu ou Moys emporté par des mains ; Matagran et Argan, la rencontre avec le cerf et les lions, l’histoire de Perron, le roi Bruillant, la mort du roi Lancelot, sa tête dans la fontaine bouillante et son tombeau... Pour cette dernière partie du texte, les grandes initiales filigranées dans LM7 et Royal correspondent aux emplacements des initiales historiées de Rennes. Dans Le Mans et fr 770 le choix de sujets et d’initiales décoratives varie, dans Le Mans entre une initiale champie (10 cas) et une initiale historiée (n° 14, Pierre découvert dans son bateau par la fille du roi Orcaut au f. 190v, fig. 16) ; dans fr 770, 3 grandes initiales filigranées contre 4 miniatures et 4 initiales historiées. La dernière miniature (n° 19) dans fr 770 accompagnée sur la même page par deux initiales filigranées (fig. 17) dont là encore les deux endroits sont marqués dans LM7 (fig. 18), Rennes (fig. 19) et Le Mans (fig. 20) où LM7 et Le Mans donnent des initiales décoratives alors que deux initiales historiées ont été inclues dans Rennes. Deux choix de sujets qui unissent Rennes et fr 770 concernent le Graal qui est représenté à deux reprises dans fr 770 et une fois dans Rennes : le saint vessel porté en Angleterre par les confrères de Josephé (fig. 21, 22), et Josephé sur son lit de mort transmettant le Graal à Alain dans fr 770 (fig. 23); dans Le Mans 354 la scène de la transmission du Graal, là encore représenté sous forme d’écuelle, se trouve dans le quatrième et dernier compartiment de la miniature en tête du manuscrit. L’image correspondant dans Rennes n’inclut pas le Graal (fig. 24). Le Graal est représenté dans ces trois scènes sous forme d’escuelle alors que dans les manuscrits plus tardifs le Graal prend la forme de vessel eucharistique, calice ou ciboire.11 Parmi les nombreux objets qui ont servi de modèle au Graal – les prétendus calices du seigneur – il n’y a qu’un seul qui prend la forme d’une écuelle. Il s’agit du vaisseau de verre couleur émeraude, probablement de confection fatimide, conservé au trésor de la cathédrale Saint-Laurent de Gènes12, ramené de Césarée par les Croisés génois (fig. 25).13 Mentionné pour la première fois au XIIe siècle par Guillaume de Tyre,14 Stones, ‘Seeing the Grail’; Meuwese, ‘The Shape of the Grail.’ Voir Fra Gaetano da S. Teresa, II Catino. 13 Pour un résumé de son histoire, voir Zahlten, ‘Der “Sacro Catino” in Genua.’ 14 Guillaume de Tyre, Historia rerum in partibus transmarinis gestarum ‘... in hoc eodem oratorio [à Césarée] repertum est vas coloris viridissimi in modum parobsidis formatum, quod praedicti Ianuenses smaragdum reprantes, pro multa summa pecuniae in sortem recipientes, ecclesiae suae pro excellenti obtulerunt ornatu...’ (livre X chapitre XVI, p. 423); voir aussi Aubry des Trois Fontaines, ‘...ubi participes illius victoriae Ianuenses vas 11 12

UN SCHÉMA D’EMPLACEMENT

1. Paris, BnF fr 116, Lancelot, f. 688v: Morgain montre à Arthur les peintures faites par Lancelot. (photo: BnF)

409

410

2. Paris, BnF fr 112, Lancelot, t. III, f. 193v: Le roi Arthur découvre son infortune en regardant les peintures de Lancelot. (photo: BnF)

UN SCHÉMA D’EMPLACEMENT

3a. Nottingham, University, Wollaton WLC/LM7, L’Estoire del saint Graal, f. 1, initiale M. (photo: Nottingham University)

3b. University of California, Berkeley, 106, L’Estoire del saint Graal, f. 145v, Trois personnages devant l’Arbre de Vie et un tasseau derrière. (photo: University of California, Berkeley)

411

412

4. Paris, BNF fr 19162, L’Estoire del saint Graal, f 65v, L’Arbre de Vie; une femme confectionnant du beurre. (photo: BnF)

5. Paris, BNF fr. 24394, L’Estoire del saint Graal, f. 50v, L’Arbre de Vie. (photo: BnF)

UN SCHÉMA D’EMPLACEMENT

6. Paris, BnF fr. 344, L’Estoire del saint Graal, f. 43, Des personnages devant l’Arbre de Vie et un tasseau derrière l’arbre. (photo: BnF)

7. Paris, BNF fr. 95, L’Estoire del saint Graal, f. 49v, L’Arbre de Vie. (photo: BnF)

413

414

8. Londres, BL Additional 10292, L’Estoire del saint Graal, f. 33v, initiale M. (photo: British Library)

9. Rennes, BM 255, L’Estoire del saint Graal, f. 46, Josephé et ses compagnons dans la nef de Salomon découvrent le lit et les fuissiaus. (photo: Lancelot-Grail Project)

UN SCHÉMA D’EMPLACEMENT

10. Le Mans, MM 354, L’Estoire del saint Graal, f. 110, Nascien sur la nef de Salomon découvre le lit. (photo: Lancelot-Grail Project)

11. Paris, BnF fr 770, L’Estoire del saint Graal, f. 66v, Nascien nage pour regagner la berge. (photo: BnF)

415

416

12. Nottingham, University, Wollaton WLC/LM7, L’Estoire del saint Graal, f. 3v initiale O. (photo: Nottingham University)

13. Nottingham, University, Wollaton WLC/LM7, L’Estoire del saint Graal, f. 12v, initial C. (photo: Nottingham University)

UN SCHÉMA D’EMPLACEMENT

15. Paris. BnF fr 770, L’Estoire del saint Graal, f. 73v, Celidoine mis dans une barque. (photo: BnF)

417

14. Le Mans, MM 354, L’Estoire del saint Graal, f. 123, Celidoine mis dans une barque avec un lion. (photo: LancelotGrail Project)

16. Le Mans, MM 354, L’Estoire del saint Graal, f. 190v, Pierre découvert dans son bateau par la fille du roi Orcaut. (photo: Lancelot-Grail Project)

418

17. Paris, BnF fr. 770, L’Estoire del saint Graal, f. 119v Celidoine et ses compagnons chevauchant. (photo: BnF)

UN SCHÉMA D’EMPLACEMENT

419

18. Nottingham, University, Wollaton WLC/LM7, L’Estoire del saint Graal, f. 73v, initiales O. (photo: Nottingham University) 19. Rennes, BM 255, L’Estoire del saint Graal, f. 98v Nascien and King Mordrain at Josephé’s tomb; Celidoine and his son Narpus bid farewell to Nascien (photo:Lancelot-Grail Project).

420

20. Le Mans, MM 354, L’Estoire del saint Graal, f. 204v, champie initial O. (photo: Lancelot-Grail Project)

21. Rennes, BM 255, L’Estoire del saint Graal, f. 76v, Le saint vessel porté en Angleterre par les confrères de Josephé. (photo: Lancelot-Grail Project)

UN SCHÉMA D’EMPLACEMENT

421

22. Paris, BnF fr 770, L’Estoire del saint Graal, f. 98, Josephé sur son lit de mort transmettant le Graal à Alain. (photo: BnF)

23. Paris, BnF fr 770, L’Estoire del saint Graal, f. 117v, Le saint vessel porté en Angleterre par les confrères de Josephé. (photo: BnF)

422

24. Rennes, BM 255, L’Estoire del saint Graal, f. 96v, Alain pleuerant Josephé sur son lit de mort. (photo: Lancelot-Grail Project)

25. Fra Gaetano, Le Sacro Catino (gravure de 1726).

UN SCHÉMA D’EMPLACEMENT

423

La renommée de cet objet a été commentée notamment par Pétrarch dans son Itinéraire en Terre Sainte.15 Il le recommande impérativement à son ami, embarquant à Gènes : Même si vos compagnons de voyage sont pressés et si les marins dénouent déjà les cordes, ne partez pas avant d’avoir contemplé cette précieuse vase taillée entièrement en émeraude que le Christ, par amour de qui vous voyagez si loin de votre pays, a utililsé pour célébrer la Cène. C’est un objet de dévotion si c’est vrai ce qu’on dit, fabriqué avec art...16 Parmi les nombreux récits des pèlerins en Terre Sainte, signalons celui composé par Epiphanius le Moine qui, seul, mentionnne la présence au Saint-Sépulcre d’un vessel en émeraude, disant qu’il s’agissait du vessel dans lequel le Christ but le vinaigre...17 Mais il faut insister sur l’absence de rapport dans ces récits entre le vessel du seigneur et le Graal. Ce sera la contribution de la littérature depuis la première mention du Graal par Chrétien de Troyes, dont les origines chrétiennes ont été élaborées dans l’Estoire del saint Graal, qui verra dans le vessel du seigneur les origines du saint Graal.18 viridissimi coloris repertum et in modum paropsodii formatum pro multa summa pecuniae recipientes in sortem pro excellenti obtulerunt ornatu ecclesiae suae...’ (cité par Cabrol et Leclercq, Dictionnaire d’Archéologie chrétienne et de liturgie, t. VI, cols. 900–901; Iacopo da Voragine e la sua Cronaca di Genova, ‚...erat autem in predicta civitate [Caesarea] vas quoddam smaragdinum et inextimabiliter preciosum... Ianuenses igitur, ceteris alijs partibus omissis, vas illud smaragdinum pro sua parte acceperunt et Ianuam cum multo gaudio ipsum deportaverunt...”, II, p. 309–10. 15 Cachey, ed., Itinerary to the Sepulcher of Our Lord Jesus Christ (Itinerarium ad sepulchrium domini nostri Yesu Christi), facsimile ed. du ms Cremona, Biblioteca statale, Deposito libreria civica, MS BB.1.2.5. Nous remercions Jeanne Krochalis d’avoir attiré cet ouvrage à notre attention. 16 ‘Hinc tu, tametsi socii properent et naute de littore funem solvant, non tamen ante discesseris quam preciosum illud et insigne vas solido e smaragdo, quo Christus, cuius te tam procul a patria amor trahit, pro parapside usus fertur, videas devotum si sic est, alioquin suapte specie clarum opus... (f. 4r, Cachey, para. 2.5 et p. 169, n. 26). 17 Wilkinson, Jerusalem Pilgrims Before the Crusades, p. 208, traduit d’après Donner, ‘Die Palästinabeschreibung de Epiphanius Monachus Hagiopolita,’ p. 83, ‘es ist wie grünlicher Kalkstein...’. La datation du texte est incertain (entre 670 et 800 selon Wilkinson et les éditeurs cités par Donner); les manuscrits sont tous tardifs, le plus ancien datant aux environs de 1300. 18 Les représentations du Graal dans les manuscrits de Chrétien de Troyes le traitent en objet liturgique. Voir Baumgartner, ‘Les scènes du Graal et leur illustration dans les manuscrits du Conte du Graal et des Continuations.’

424

La présence des grandes initiales filigranées et des images dans les cinq manuscrits considérés ici indique l’importance des passages concernant le Graal pour les tout premiers lecteurs de l’Estoire del saint Graal à côté d’autres passages mis en avant par la décoration à travers ce texte si captivant. La comparaison entre le traitement décoratif et illustratif de ces manuscrits indique que, dès les premiers exemplaires, un schéma de points de repère avait été établi dans l’ensemble du texte. Ce fut d’abord dans un manuscrit sans images, LM7, que fut réalisé le schéma, repris plus ou moins dans les autres exemplaires. La popularité du roman est attestée par la quantité remarquable d’exemplaires qui ont survécu – parfois conservés avec soin, en d’autres cas fortement abîmés et usés – des exemplaires qui ont été souvent lus et consultés. Que cette matière continue à fasciner de nos jours, l’exposition de Troyes, comme celles de Paris et de Rennes, en témoigne.

UN SCHÉMA D’EMPLACEMENT

425

Table Comparative : distribution du décor dans les plus anciens manuscrits de l’Estoire del saint Graal Editions du texte : Hucher, Le Saint Graal, t. II et III. Ms de base : Le Mans 354 (H). Sommer, Vulgate Version, t. I (S I), Ms de base : Londres, BL Add. 10292. Ponceau, L’Estoire del Saint Graal, (Pon). Mss de base : olim Amsterdam, BPH 1 au f. 63 (t. 1), puis Rennes BM 255 (t. 2). Micha, Merlin (ici MM). Mss de base : le fragment du Merlin en vers de Robert de Boron, Paris, BnF fr 20047, ff. 55v–62, et la version en prose du Merlin de Robert de Boron sans la Suite Vulgate du Merlin, Paris, BnF fr 747 (Siglum A). Manuscrits : Nottingham, University of Nottingham, WLC/LM/7 Décoration : petites initiales filigranées rouge aux filigranes bleues alternant avec initiales bleues aux filigranes rouges ; grandes initiales filigranées au corps de l’initiale en rouge et bleu aux filigranes rouges et bleus, sur 5 ou 6 lignes. Londres, BL Royal 19 C.xii Decoration : petites initiales filigranées rouge aux filigranes bleus alternant avec initiales bleues aux filigranes rouges ; grandes initiales filigranées au corps de l’initiale en rouge et bleu aux filigranes rouges et bleus, sur 4 à 8 lignes. Rennes Bibl. mun. 255 Décoration : petites initiales champies comportant de nombreux motifs d’oiseaux, d’hybrides ; initiales historiées. Le Mans, Méd. mun. 354 Decoration : petites initiales filigranées, grandes initiales champies ; initiales historiées, miniatures dans une colonne de texte avec antennes et motifs. Paris, BnF fr 770 Decoration : petites initiales filigranées, grandes initiales filigranées, initiales historiées, miniatures.

Royal f. 39 O, grande initiale filigranée

LM7 f. 3v O, grande initiale filigranée (filigranes incomplets)

LM7 f. 6 E, grande initiale filigranée

2. H II 484 S 1136. 35 Pon 291.461 Or dit li contes que grant piece regarda nasciens les .iij. fuissiax...

3. H II 496 S I 141. 33 Pon 299.473 En ceste partie dit li contes que qant les .ix. mains en orent porte celydoine...

Royal f. 40v E, grande initiale filigranée

Royal f. 37 M, grande initiale

LM7 f. 1 M, grande initiale

1. H II 466 S I 130. 6 Pon 279.445 Molt durerent longuement cil arbres en cel color...

Royal 19 C.xii c. 1220-50 ?

WLC/LM/7 c. 1220 ?

Texte d'après LM7

Rennes f. 52v aucune décoration

Rennes f. 46 O, Nascien et ses compagnons regardent la nef au lit orné de fuissiaus blanc, vert et rouge.

Rennes f. 43v M, petite initiale champie

Rennes 255 c. 1220

Le Mans f. 113 E, grande initiale champie

Le Mans f. 110 O, Nascien regarde le lit sur la nef (les fuissiaus sont absents); sur l’antenne en haut un terminal-chevalier tenant une épée; en bas un terminal-clerc tenant un livre.

Le Mans M, petite initiale filigranée

Le Mans 354 c. 1285

BnF fr 770 f. 68 E, grande initiale filigranée

BnF fr 770 f. 66v miniature: Nascien dans l’eau à côté de la nef contenant le lit (les fussiaus sont absents).

BnF fr 770 M, petite initiale filigranée

Paris, BnF fr 770 c. 1285

426

LM7 f. 18 U, grande initiale filigranée

6. H III, 29 S I 171.1 Pon 349.545 Uoirs fu ce dit li contes e lestoire de phisophes le tesmoigne que ypocras fu li plus souereins clerc... Royal f. 49 U, grande initiale filigranée

Royal f. 47 O, grande initiale filigranée

LM7 f. 15 O, grande initiale filigranée

5. H III, 10 S I 164. 4 Pon 337.529 Or dit li contes que quant que quant li .v. messaie furent parti de leur dame... Rennes f. 57 U, Hippocrate guérit le neveu de l’empereur Auguste.

Rennes f. 54v O, Flegentine envoie des messagers à la recherche de son mari Nascien.

Royal f. 45v Rennes f. 52v aucune C, petite initiale bleue décoration sans filigranes sur 4 lignes

LM7 f. 12v C, grande initiale filigranée

Rennes 255 c. 1220

4. H II 533 S I 158. 23 Pon 328.515 Cist contes qui est apelez le conte del saint Graal deuise que quant li home le roi label orent pris celydoin si len menerent entre les roches...

Royal 19 C.xii c. 1220-50 ?

WLC/LM/7 c. 1220 ?

Texte d'après LM7

Le Mans f. 130v O, grande initiale champie Or dist li contes que voirs fu et lestoire des phylosophyes le tesmongne que Ipocras fu li plus souvrains clers...

Le Mans f. 126v O, Les messagers partent à cheval; terminal d’antenne en forme d’un joueur de cornemuse et d’un chevalier armé d’une épée et d’un bouclier.

Le Mans f. 123 O, Celidoine dans son bâteau lancé dans la mer par deux hommes; sur l’antenne un terminal en joueur de viel coiffé d’un chapeau pointu et un terminal en danceur levant les bras. Or dist li contes ici endroit del S. Graal et devise que quant...

Le Mans 354 c. 1285

BnF fr 770 f. 77v O, grande initiale filigranée Or dist li contes que voirs fu et lestoire des philosofes le tesmoigne que ypocras fu li plus sourains clerg...

BnF fr 770, f. 75v O, Les messagers chevauchant; sur l’antenne un chevalier coiffé d’un chapeau à pointe sur sa coiffe de mailles, portant un écu de gueules.

BnF fr.770 f. 73v miniature: Célidoine dans son bateau est lancé dans la mer par deux hommes sous l’ordre d’une troisième personne (le roi Laban). Or dit li contes ichi endroit qui est apieles li contes du saint Graal et devise que quant...

Paris, BnF fr 770 c. 1285

UN SCHÉMA D’EMPLACEMENT

427

Royal f. 60 O, grande initiale filigranée

LM7 f. 33v O, grande initiale filigranée

LM7 f. 34 C, grande initiale filigranée

8. H III, 123 S I 207. 24 Pon 410.645 Or dist li contes que quant nasciens se fu partiz de belyc si com il uos est ia deuise flegentine sa feme...

9. H III 126 S I 208. 21 Pon 412. 648 Ci endroit dit li contes que quant iosephes se fu partiz de sarraz il erra entre lui sa conpaignie meinte iornee... Royal f. 60v O, grande initiale filigranée Or dist li contes ci endroit que quant iospeph se fu partiz de sarraz...

Royal f. 52v C, grande initiale filigranée

LM7 f. 23 Q, grande initiale filigranée

7. H III 73 S I 182. 34 Pon 370.578 [0]Ci endroit dit li contes que quant li dui mesage e la damoisele qui auec lui estoient orent ore grant piece regarde...

Royal 19 C.xii c. 1220-50 ?

WLC/LM/7 c. 1220 ?

Texte d'après LM7

Rennes f. 68v O, Josephé et ses compagnons quittent la ville de Sarras. Or dit li contes ici endroit qe qant Joseph se fu partiz de la cite de sarraz entre lui et sa compaignie...

Rennes f. 68 O, Flegentine pleure son mari absent.

Rennes f. 60v C, Les deux messagers et la demoiselle devant le tombeau d’Hippocrate.

Rennes 255 c. 1220

Le Mans f. 152 O, initiale champie Or dist li contes que qant Josef se fut partis de Sarras...

Le Mans f. 151 O, Flegentine interroge les trois messages; terminal d’antenne: clerc lisant; sur l’antenne un jeune homme jetant une pierre; un éléphant portant au dos une selle (howdah) contenant deux chevaliers.

Le Mans f. 137 C, Les deux messagers et la demoiselle lisent les lettres gravées sur la maison d’Hippocrate (expliquant que les ruses de sa femme ont causé sa mort) ; terminaux d’antenne: tête de jeune homme, tête d’ours.

Le Mans 354 c. 1285

BnF fr 770 f. 90v miniature : Dieu au ciel indique du doigt Josephé et ses compagnons qui dorment. Or endroit dist li contes que quant Joseph se fit partis de Sarras...

BnF fr 770 f. 90 O, Flegentine pleure son mari absent ; terminal d’antenne: un chevalier armé d’un bouclier levant la main.

BnF fr 770 f. 81v miniature : La demoiselle et les messagers devant la maison d’Hippocras.

Paris, BnF fr 770 c. 1285

428

LM7 f. 45 O, grande initiale filigranée

LM7 f. 61 O, grande initiale filigranée

12. H III, 169 S I 231. 4 Pon 453. 720 Or dit li contes que quant josephes se fu partiz de Galafort...

13. H III, 235 S I 266. 3 Pon 519. 821 Or dit Ili contes que a celui point que Symeu en fu portez einsi come ge uos ai dit... Royal f. 80 O, grande initiale filigranée

Royal f. 68 O, grande initiale filigranée

Royal f. 64 C, grande initiale filigranée

LM7 f. 39v C, grande initiale filigranée

11. H III, 146 S I 219. 11 Pon 431. 682 Ci endroit dit li contes que quant li du se fu coucniet en son lit il comenca a penser...

Royal 19 C.xii c. 1220-50 ? Royal f. 62 O, grande initiale filigranée

WLC/LM/7 c. 1220 ?

10. H III, 133 Lm 7 f. 36 S I 212. 31 O, grande initiale Pon 419. 661 filigranée Or dit li contes que quant na (expunged) lì preudons qui a nascien avoit deuise del brief...

Texte d'après LM7

Rennes f. 89 O, Symeu emporté par deux hommes enflammés (initiale très abimée).

Rennes f. 89 O, Symeu emporté par deux hommes enflammés (initiale très abîmée).

Rennes f. 72v C, Le duc Ganor au lit réfléchit sur les paroles de Célidoine.

Rennes f. 70 O, Nascien, au bord de la mer parle avec un homme dans un bateau qui tient un livre.

Rennes 255 c. 1220

Le Mans f. 188 O, initiale champie

BnF fr 770 f. 110v O, Deux des compagnons de Josephé rencontrent Symeu dans sa fosse.

BnF fr 770 f. 98 miniature: Les compagnons de Josephé s’approchent d’une ville en Norgales, portant le Graal.

BnF fr 770 f. 94 C, Le duc Ganor au lit réfléchit; un homme- terminal d’antenne en prière lui fait face.

Le Mans f. 158v C, initiale champie

Le Mans f.188 O, initiale champie

BnF fr 770 f. 92 O, grande initiale filigranée

Paris, BnF fr 770 c. 1285

Le Mans f. 155v O, initiale champie

Le Mans 354 c. 1285

UN SCHÉMA D’EMPLACEMENT

429

WLC/LM/7 c. 1220 ?

LM7 f. 63 O, grande initiale filigranée

LM7 f. 68v E, grande initiale filigranée

LM7 f. 70 O, grande initiale filigranée

Texte d'après LM7

14. H III 245 S I 270. 1 Pon 527. 833 Or dit li contes que quant la nascele ou pierres estait se fu esloignie de la riue...

15. H III, 271 S I 281. 9 Pon 548. 864 En ceste partie dit li contes que quant iosephes se fu partiz de perron et de pharan...

16. H III, 279 S I 284. 21 Pon 555. 874 Or dit li contes que qant iosephes ot corone Galaaz son frere que il erra tant... manque

manque...

Royal f. 81V O, grande initiale filigranée f. 84v Royal termine incomplet, de cele request feistes ma volente .ge la uostre de ce dont vos me requirez/

Royal 19 C.xii c. 1220-50 ?

Rennes f. 96 O, Josephé bénit son frère Galaad après l’avoir couronné.

Rennes f. 94v O, Josephé part de Perron et Pharain en chevauchant. Or dist li contes en ceste partie que qant Josesphes se fu partis de Pieron...

Rennes f. 90v O, Perron, debout dans son bateau, fait ses adieux à Pharain et ses compagnons.

Rennes 255 c. 1220

Le Mans f. 199v O, initiale champie

Le Mans f. 197v O, initiale champie Or dist li contes en ceste partie que qant Josesphes se fu partis de Pieron...

Le Mans f. 190v C, La fille du roi Orcaut et deux demoiselles aperçoivent depuis la porte de leur château Pierre endormi dans son bateau ; deux terminaux d’antenne en forme d’ours et d’oiseau qui picore son propre cou. Chi endroit dit li contes ke quant la naciele u pieres estoit se fu eslongie de la rive...

Le Mans 354 c. 1285

BnF fr 770 f. 116v O, grande initiale filigranée

BnF fr 770 f. 115v O, Josephé et ses compagnons font leurs adieux à Perron. Or dist li contes en ceste partie que qant Josesphes se fu partis de Pieron...

BnF fr 770 f. 111v Miniature : Perron, debout dans son bateau, fait ses adieux à Pharain resté à la rive. Chi endroit dit li contes que quant la nacele ou pierres estait se fu eslongie de la riue...

Paris, BnF fr 770 c. 1285

430

S II 3. 1 MM 18. 1

Merlin LM7 f. 76 (feuille endommagée) M [une partie de grande majuscule est conservé) [olt] fu iriez.... fin du Merlin, incomplet.

LM7 f. 73v O, grande initiale filigranée

19. H III, 297 S I 291. 21 Pon 568.894 Or dit li contes que quant (Josephe rayé) celydoines (ajouté dans la marge) se fu partiz de (nar rayé) son pere il sen ala entre lui e narpus... manque

manque

manque

LM7 f. 73v O, grande initiale filigranée

18. H III 296 S I 291.7 Pon 567. 893 Or dit li contes que quant iosephes fut partiz de (celydoine rayé et gommé) siecle a son trespassement fu nascien et celydoine et narpus li filz celydoine...

Royal 19 C.xii c. 1220-50 ? manque

WLC/LM/7 c. 1220 ?

17. H III, 282 LM7 f. 70v S I 286. 1 O, grande iniPon 557. 878 Or dit li tiale filigranée contes que quant iosephes uit que il estoit el trespassement del siecle...

Texte d'après LM7

Rennes f. 101 M, La descente du Christ aux enfers. Mout fu iriet li enemis qant nostre sires ot est en enfer... f. 108 Vortiger trônant entre deux barons. Or dit li contes que Vertigiers fu elleuz a roj... Plus d’illustration dans le Merlin.

Rennes f. 98v O, Celidoine et le roi Mordrain font leurs adieux à Nascien.

Rennes f. 98v O, Nascien et le roi Mordrain devant le tombeau de Josephé..

Rennes f. 96v O, Alain pleure devant Josephé sur son lit de mort.

Rennes 255 c. 1220

Le Mans Merlin manque

Le Mans f. 205 O, initiale champie

Le Mans f. 204v O, initiale champie

Le Mans f. 200v O, initiale champie

Le Mans 354 c. 1285

BnF fr 770 f. 121 Le Merlin commence par une miniature provenant d’un manuscrit des environs de 1400 représentant la Vierge à l’enfant, collée sur une feuille de parchemin moderne, f. 127v O, Meurtre du roi Maines f. 128 O, Punition des bourreaux de Maines. Plus d’illustration dans le Merlin.

BnF fr 770 f. 119v-2 Miniature : Celydoine et ses compagnons à cheval. BNF fr. 770 f. 120v le texte termine incomplet à la fin du cahier...et en ce qui’il se fu abaissies pour boire li dus li vint par deriere/ réclame lespee traite/

BnF fr 770 f. 119v O, grande initiale filigranée

BnF fr.770 f. 117v O, Josephé sur son lit de mort transmet le Graal (une écuelle blanche) à Alain; un hybride-chevalier en terminal d’antenne tenant un bouclier de gueules à une bande cotisée argent [blanc] et une massue.

Paris, BnF fr 770 c. 1285

UN SCHÉMA D’EMPLACEMENT

431

XXI Seeing the Grail Prolegomena to a Study of Grail Imagery in Arthurian Manuscripts Or as veu ce que tu as tant desirré aveoir, et ce que tu as convoitié 1

T

he Quest for the Holy Grail is one of several significant themes in the courtly culture of the Middle Ages, a search for purity and perfection, for personal fulfillment and collective aspiration, in which only a chosen few of King Arthur’s knights will succeed.2 The Grail itself — a potent, mysterious, life-giving object, providing food, healing the wounded, bestowing sight, both physically and metaphorically, on the deserving few, yet blinding the unworthy — is never fully described in the texts that transmit its legends. Rather, it is reified by a word or phrase evoking some kind of receptacle:

Parts of this essay were presented at the Lancelot conference in Austin, Texas, in 1992, at the 17th International Arthurian Congress held in Bonn in 1993, at the Bible conference at the Hebrew University in Jerusalem in 1999, and at Magdalen College Oxford in 2001. 1 am especially grateful to †Elspeth Kennedy and Martine Meuwese for helpful discussion, and to Dhira Mahoney for her patient and pertinent editing. First published in The Grail, A Casebook, ed. D. Mahoney (New York: Garland, 2000), pp. 301–66. See now also ‘The Grail in Rylands French 1 and its Sister Manuscripts,’ Bulletin of the John Rylands University Library of Manchester, 81–83, 1999, pp. 55–95, reprinted in these essays. Since I published the present article many other authors have written about the Grail. I list their titles in the bibliography at the end of these essays: Ashe, Barber, Bryant, Combes, D’Arcy, Griffin, Lacy, Dixon, Goering, Meuwese, Ortenberg, Ramm, Séguy. 1 ‘So you have seen what you so desired to see, and what you coveted’: Pauphilet, Queste, (1965), pp. 270–71. I warn the reader that the pagination in previous and subsequent editions is not the same. 2 The classic studies are Weston, The Quest of the Holy Grail and Loomis, The Grail.

SEEING THE GRAIL

433

‘vessel,’ in the Perceval of Chrétien de Troyes (ca. 1175), the verse Estoire of Robert de Boron, and the prose Joseph (Modena version), ‘vessel’ or ‘escuele’ in the prose Lancelot-Graal (early thirteenth century), or something more elemental, like the stone of Wolfram von Eschenbach’s Parzìval (ca. 1200– 1210).3 Its pictorial presence in the illuminated manuscripts that transmit its legends is equally spare. Often it is excluded altogether in otherwise fully illustrated manuscripts. When it is depicted, its shape and degree of visibility follow several different patterns, whose other cultural referents offer visual clues beyond the words of the text as to how the Grail was perceived by those makers and patrons who felt its legends were worth copying and illustrating, selling and buying, reading, looking at, and owning, between the thirteenth and sixteenth centuries. By far the greatest number of Grail depictions occur in the manuscripts of the Lancelot-Graal or Vulgate Cycle of Arthurian romances in French prose.4 Composed soon after the turn of the thirteenth century, this five-part cyclical romance includes two branches, the Estoire and the Queste del Saint Graal, in which the early history of the Grail and its quest by King Arthur’s knights are the primary focus. Although these were not the first romances to emphasize the Grail in their texts, the illustrations devised to show it are particularly significant in the development of illustrated vernacular manuscripts: illustrated copies of the Lancelot-Graal cycle emerge early (by ca. 1220),5 last late (to the end of the fifteenth century),6 and survive 3 The textual sources — Matthew 27: 57–61, the Pseudo-Gospel of Nicodemus (6th C.), Vindicta Salvatoris (based in part on a manuscript of the 8th C), Chretien’s Perceval, Robert de Boron’s verse Estoire, the prose Joseph, the prose Estoire, and the Queste — are conveniently drawn together in Bertoni, Materiali. The most recent text edition of Chrétien is by Busby, Chrétien de Troyes, ‘Le roman de Perceval’. For Wolfram, see Wolfram von Eschenbach, Parzival und Titurel, ed. Bartsch; and Bumke, Die Wolfram von Eschenbach-Forschung. 4 Sommer, Vulgate Version. For Joseph, see O’Gorman, Robert de Boron. For the manuscript tradition, see Woledge, Bibliographie, nos. 93, 96, 114, and id., Supplément, nos. 93, 96, 114. 5 The earliest surviving Lancelot-Graal manuscript with cyclical illustration is the Estoire, Merlin, and beginning of Lancelot, Rennes, BM 255, ca. 1220, see Stones, ‘The Earliest’. This certainly antedates the earliest surviving manuscripts of Chrétien de Troyes that contain narrative illustrations, which are datable only in the third quarter of the thirteenth century: see Les Manuscrits de Chrétien de Troyes, particularly the Catalogue of Manuscripts by Terry Nixon, II, pp. 18–86, and the chronological sequence of illustrations, particularly figs. 103–435. Other ‘early’ French vernacular manuscripts with a single historiated initial at the beginning of each text are the Guiot manuscript of Chrétien’s works, Paris, BNF fr 794, and the prose Lancelot do Lac, Paris, BNF fr 768, the manuscrit de base of Kennedy’s edition. See Les Manuscrits de Chrétien de Troyes for full references and especially Stirnemann, ‘Some

434

in very large numbers (close to 200 copies),7 and their chronological and geographical distribution can be plotted with a fair degree of accuracy. By contrast, there is only a single Grail depiction in surviving copies of the Perceval of Chrétien de Troyes (composed ca. 1189) — Paris, BNF fr 12577, f. 18v — although the late thirteenth- and early-fourteenth-century copies of the Continuations added by later writers include more Grail depictions.8 Champenois Vernacular Manuscripts’ at pp. 204–05 and fig. 22 (Guiot), and p. 207 and fig. 24 (Lancelot). These two manuscripts also exclude illustrations of the Grail or related subjects. Early examples of the sporadic use of the historiated initial as such can also be found in the Roman de Troie, Paris, Bibl. de l’Arsenal 3340, copied in 1237, on which see Samaran and Marichal, Manuscrits datés, I, 159, pl. 12. Latin manuscripts offer some early examples of the presence of historiated initials, such as the classical Latin compilation. Paris, BNF lat 7936, on which see Avril, ‘Un manuscrit d’auteurs classiques’, and the possibly related compilation in Edinburgh University Library, MS 20 (D.b.VI.6), described in Borland, Catalogue of the Western Mediaeval Manuscripts in Edinburgh University Library, pp. 30–31. To my knowledge no scholarly opinion has so far been expressed about the likely date of BNF fr 748, mentioned below for its interesting opening initial; it is possible that it may also be included among these ‘early’ books. The illustrated compendium in Nottingham University, WLC/LM6, may be even earlier than Rennes: see now Stones, ‘Two French Manuscripts’, reprinted in these essays. 6 See Burin, ‘The Pierre Sala Manuscript’. The unillustrated Estoire, BNF fr 1427, was written in 1504, see Woledge, Bibliographie, p. 74. See also the editions of the Lancelot, Queste, and Mort Artu and the Tristan printed by Antoine Vérard from 1488 onward, discussed in Pickford, ‘Antoine Vérard’; see also Loomis, Arthurian Legends in Medieval Art, pp. 140–44; Woledge, Bibliographie, p. 75; Creating French Culture (exhibition catalogue), no. 53. For Vérard see now Winn, Anthoine Vérard. 7 A precise count depends on whether multiple volumes of the same set are counted separately or as part of the same cycle; see the Working List of Illustrated Manuscripts at the end of these essays. For approximate numbers of manuscripts per branch, see Stones, ‘BN fr 95 and Yale 229’, note 23. The most important survey is still Loomis, Arthurian Legends. See also Stones, ‘Arthurian Art Since Loomis’; Whitaker, Legends of King Arthur in Art and the articles on ‘Manuscripts. Illuminated’ by Stones and Stokstad in Lacy, The New Arthurian Encyclopedia, pp. 299–308. Most of the manuscripts discussed here were included somewhere in my unpublished Ph.D. thesis, Illustrations of Lancelot. Two other unpublished dissertations have been devoted to the iconography of the Estoire in two related Franco-Flemish manuscripts of the early fourteenth-century: Remak-Honef, Text and Image in the Estoire, and Meuwese, ‘L’Estoire del Saint Graal’, Een Studie over de relatie tussen miniaturen en tekst. Lists of the subjects in BNF fr 112, 113–116. 117–120, together with comparative iconography charts of the Lancelot, Queste, and Mort Artu branches of the cycle, are given in another dissertation, by Blackman, Manuscripts and Patronage of Jacques d’Armagnac. Blackman‘s comparative iconongraphical charts are also reproduced in ead., ‘Pictorial Sinopsis’. 8 See Baumgartner, ‘Les scènes du Graal’. A more general study of Grail iconography is Burdach, Der Graal, pp. 415–49; and, for Chrétien, see (with reservations) Hindman, ‘King

SEEING THE GRAIL

435

Illustrated copies of Wolfram’s Parzival (composed ca. 1200–1210) are few, and those that have survived rarely show the Grail.9 Robert de Boron’s verse Grail romance (ca. 1200) is unillustrated.10 The special prose version in Modena has historiated initials, but includes no depiction of the Grail.11 The Spanish and Galician/Portuguese Arthurian romances are also unillustrated, while texts in French copied in Spain are rare, and do not include the Grail. Arthurian texts in English or made in England are sparsely illustrated and have no Grail illustration. In Italy the illustrative Arthurian tradition occurs predominantly in copies of romances in French — with the important exception of the Tavola Ritonda, a special case that I consider below. The Perceval casket in ivory similarly lacks depictions of the Grail.12 Although there are illustrated copies of Peredur, Perlesvaus, Tristan, and Palamède, they Arthur, His Knights, and the French Aristocracy in Picardy’. The relation between the chalice/ attribute of Ecclesia and the Grail, overplayed in much of the popular literature, cannot be dealt with here. 9 In the late-thirteenth-century copy, Munich, Bayerische Staatsbibliothek, Cgm 19, f. 50v, the bottom register shows the Grail twice, each time borne by the Grail Queen who holds it at Feirefitz‘s baptism and again, before him. In both instances she holds it in a huge cloth in such a way that her hands are veiled by the cloth. This is a device to emphasize the holiness of the Grail that is borrowed from Christian iconography, where, from Early Christian times, holding something precious in veiled hands was a sign of reverence (for instance the crown of martyrdom in the apse mosaic of San Vitale, Ravenna, mid-6th C.). See Loomis, Arthurian Legends, pp. 131–32, fig. 358; see also the facsimile, Parzival, Titurel, Tageliede: Cgm 19 der Bayerischen Staatsbibliothek München. The four other illustrated manuscripts are summarily described in Benziger, Parzival in der deutschen Handschriften des Mittelalters, and in Schirok, Wolfram von Eschenbach: ‘Parzival.’ See also Stamm-Saurma (Saurma-Jeltsch), ‘Zuht und wicze’, and Curschmann, ‘Der Berner Parzival und seine Bilder’. Paradoxically, the Grail is not a major feature of the illustrations of these manuscripts. 10 Edited from the only surviving manuscript, Paris, BNF fr 20047, by Nitze, Le roman de l’Estoire dou Graal. See also Bertoni, Materiali, pp. 30–47, and Burdach. Der Graal, pp. 450–502. 11 It may well be that the special prose version of Arthurian romance transmitted in the Modena manuscript, Biblioteca Estense alpha 930 (E. 39), containing Joseph, Merlin, and Perceval, illustrated with fifteen small historiated initials, dates sometime between 1190 and 1210, as Roach surmised for its Perceval section: see The Didot Perceval, According to the Manuscripts of Modena and Paris, ed. Roach, p. 130. See also Woledge, Bibliographie, no. 1213, pp. 90–91, and Supplément, pp. 70–71. 12 For the one surviving example, see Koechlin, Les Ivoires gothiques françaises, I, pp. 513–16; II, no. 1310, 3, pl. CCXXIII–IV; Images in Ivory, ed. Barnet, no. 62 (entry by Danielle Gaborit-Chopin). See now the Gothic Ivories Project at The Courtauld Institute of Art, London, www.gothicivories.courtauld.ac.uk (accessed 10 March 2017) where 35 images of 4 caskets are listed under the heading Perceval.

436

do not, to the best of my knowledge, depict the Grail. The Lancelot-Graal cycle, then, provides the corpus of Grail illustrations discussed here. As text scholars and liturgists have long recognized, the legends of the Grail related in the Estoire and the Queste are permeated with Christian Eucharistic associations, which are particularly explicit in the Grail liturgies described at the beginning of the Estoire and at the end of the Queste. Christian liturgical and devotional practices relating to the transubstantiation of the Eucharistic elements — especially the elevation, veneration, and reservation of the Host,13 and the celebration of Eucharistic miracles, including the cult of the Holy Blood — inform the texts’ descriptions of how the Grail is perceived by the few knights who are privileged to see it. References to transubstantiation in the Fourth Lateran Council of 1215, for instance, show that such liturgical and devotional practices as these were already widespread,14 while the reservation of the Eucharist and the celebration of the feast of Corpus Christi emerged in the second half of the thirteenth century and became standard by the early fourteenth. Not surprisingly, liturgical vessels and containers for the celebration of the Mass and for the veneration and reservation of the Eucharistic elements are significant models for the depiction of the Grail. But they are not the only models. We shall see that Grail depictions often reflect Old Testament imagery instead of or as well as Eucharistic models, while concerns about the visibility of other sacred objects, especially the relics of saints and the relics of the Passion of Christ, also bear on the question of whether, and how, the Grail was revealed or concealed. Similarly, the depiction of liturgical vessels, in pictures of the Mass or as symbols in other contexts — especially at the Crucifixion of Christ — may be additional referents that inform the meaning of the shapes chosen to illustrate the Grail. We shall also see that this Holy Vessel is treated in a variety of ways in copies produced by the same craftsmen. I outline here the major appearances of the Grail as they were depicted in Estoire and Queste manuscripts, in the order of the texts, and comment briefly on some of the major variants in the treatment of the Grail itself and on some of the chief pictorial sources and related receptacles that may have informed the depictions. At the end of this chapter, I attach a working list of the illustrated Estoire and Queste manuscripts as a preliminary step 13 14

Cf. OED: the action or practice of retaining or preserving a portion of the Eucharist. See Tanner, Decrees of the Ecumenical Councils, I, pp. 227–303.

SEEING THE GRAIL

437

towards a comprehensive study of Grail illustration, which I reserve for more extensive treatment elsewhere (see now the list at the end of these volumes). Limitations of space allow me to reproduce only a small selection of the illustrations I discuss here. Just how the cycles of pictures in the Lancelot-Graal were selected,15 how the individual scenes were treated, what each manuscript tells us about its patrons and makers, how these manuscripts relate to what else the patrons owned and the makers made, are questions that are just beginning to be investigated in detail.16 Elsewhere I have used selected subjects in LancelotGraal cycle manuscripts, and a few particular copies of that version, to suggest some approaches to these questions.17 Here I lay the basis for a comparative study of the iconography of the Grail as a means to assess some medieval attitudes toward this symbol whose quest held so high a place in the medieval — and the modern — imagination.18

15 A useful survey of how medieval books were made and illustrated, and by whom, is Alexander, Medieval Artists and their Methods of Work. See also Stones, ‘Indications écrites et modèles picturaux.’ 16 I have investigated these questions in relation to some particular cases. For Rennes 255 see Stones, ‘The Earliest’. For London, BL Add. 10292–94 and Royal 14 E. Ill, together with Amsterdam, BPH 1/Manchester, Rylands Fr. 1/Oxford, Bodl. Douce 215, see ead.. ‘Another Short Note’. For Paris, BNF fr 95/New Haven. Yale 229, see ead., ‘BN, fr 95 and Yale 229’. See also my analysis of artistic contexts in relation to the manuscripts of Chrétien de Troyes, ‘The Illustrated Chrétien Manuscripts and their Artistic Context’, and, now to the Wollaton manuscript in Nottingham: ‘Two French Manuscripts’, reprinted in these essays. For descriptions of Amsterdam/Douce/Rylands and fr 95/Yale 229 see Stones, Gothic Manuscripts, Part I, vol. 2, Cat. nos. III–75 and 124 respectively. 17 For the Death of Arthur, see Stones, ‘Aspects of Arthur’s Death in Medieval Illumination’. For the depiction of the adulterous love between Queen Guinevere and Arthur’s valorous knight Lancelot, see Stones, ‘Arthurian Art Since Loomis’, at pp. 38–41; ead., ‘Images of Temptation. Seduction and Discovery,’ and ead., ‘Illustrating Lancelot and Guinevere’. A few other episodes — Lancelot finding the head of his ancestor, the Magic Carole, the liturgy of the Grail — are discussed in ead., ‘BN, fr 95 and Yale 229’. For an earlier note on the Grail, see ead., ‘Sacred and Profane Art’, pp. 102–03. For the iconography of bloodletting, see ead., ‘Indications écrites et modèles picturaux’, pp. 322–23, and for the hermit in the tree, see ibid., pp. 323–24, and ‘Arthurian Art Since Loomis’, pp. 39–41. 18 Surprisingly few studies have directly addressed the issue of how the Grail was depicted. Among the recent popularizing studies (to be read with caution) are von dem Borne, Der Gral in Europa; Matthews, Grail: Quest; Bouyer and Mentré, Les lieux magiques du Graal; Buschinger, Labia, and Poirion, eds., Scènes du Graal; Roquebert, Les Cathares et le Graal. See now Meuwese, ‘The Shape of the Grail’.

438

the grail depicted in the ESTOIRE Christ and the Hermit The opening scene in the Estoire usually shows Christ appearing to the hermit, handing him the book that relates the story of the Grail. Off to one side there is often an altar with a chalice — sometimes draped with a cloth, at other times undraped — upon it. Nothing suggests directly that this chalice is the Grail, since it is only after the hermit has opened the book that the Grail appears in the story. Furthermore, given the number of times in the Vulgate Cycle that visits to hermits by penitent knights include confession, it is most probable that the hermit himself, here and elsewhere, was also a priest, making the presence of a Eucharistic chalice upon an altar a commonplace. But it is also the case that the Grail is often given the shape of a chalice or goblet-shaped vessel, and so the depiction of this Eucharistic chalice could also be read as a visual anticipation of the Grail, whose story will follow. It is in the form described above that Christ’s appearance to the hermit is treated in the badly rubbed historiated initial that opens the text in the earliest extant copy, Rennes 255, made in Parisian royal circles ca. 1220 (Fig. 1), and a similar version of this subject can still be found in the midfourteenth-century copy written by Jean Deloles of Hainaut (?) in 1357, Yale 227 (Fig. 2).19 During the intervening 125 or so years, Estoire manuscripts show many variants on this subject, for instance, whether there is one scene or more, whether the chalice is present or not, and whether or not it is veiled with a cloth or corporal, whether the hermit is asleep in bed, sitting up, or kneeling in prayer, and whether the divine messenger is Christ, the Dove of the Holy Spirit, or an angel. An alternative opening illustration is simply an image of the Trinity by itself. One of the very earliest manuscripts, Paris, BNF fr. 748, f. 1, shows a ‘B’ initial with what may be a multi-Person Trinity at the top, and Christ blessing and holding an orb below, while BNF fr. 95 shows a Throne of Mercy (Gnadenstuhl) type of Trinity, where God the Father holds the 19 The colophon on f. 31 reads, ‘Cis livres fu parescript an mil.ccc.lvii. le premier samedi de guillet et le fist Jean Deloles escriven nés de hainnaut (haumaut ?) pries pour lui et ce que vous endires puissiez.’ See Shailor, Catalogue of Medieval and Renaissance, I, no. 227.

SEEING THE GRAIL

439

crucified Christ on the cross, with the Dove between.20 Occasionally both subjects are combined, so that in BNF fr. 749, f. 1, there is a Two-Persons and Dove Trinity as well as Christ appearing to the hermit (Fig. 3) and in Brussels, BR 9246, f. 2,21 the book is flown in by the Holy Ghost as dove, watched from the sky by God the Father and God the Son. The purpose of Christ’s appearance to the hermit is to dispel his fears about the mystery of the Trinity, and there is more explanation and discussion of the Trinity farther on in the text as well, so the depiction of a Trinity image here, rather than a scene of Christ and the hermit, has considerable textual justification.22 The subject would have been familiar to artists from a variety of liturgical contexts, such as the opening of the canon of the Mass in Missals, as the illustration of Psalm 109, ‘Dixit Dominus’ (‘The Lord said unto my Lord’),23 or at the opening of the hours of the Holy Ghost:24 for a medieval audience, a Trinity image in the Estoire would have resonated with allusions to these liturgical and devotional contexts — and perhaps vice-versa. The Grail at the Crucifixion The story goes on to relate how, after the Crucifixion, Joseph of Arimathea visited the room in which Christ celebrated the Last Supper with the disciples25 and took the vessel Christ used, the ‘escuele’,26 in order to collect Christ’s blood in it at the Entombment. Often, what is depicted here in Estoire manuscripts is a standard Crucifixion, without Joseph, no doubt

Reproduced in Stones, ‘BN, fr 95 and Yale 229’, fig. 8,16. Tenth-century wallpaintings in Norfolk (whose recent discovery is noted in the Newsletter of the International Center of Medieval Art, Spring 1997), apparently include a Gnadenstuhl Trinity that is among the earliest examples to have survived. For Trinity iconography, see Braunfels, Die heilige Dreifaltigkeit and Boespflug and Załuska, ‘Le dogme trinitaire’. 21 Reproduced in Loomis, Arthurian Legends, fig. 299, and in color in Matthews, Grail: Quest, p. 37. 22 Sommer, Vulgate Version, I, p. 4, lines 18–20 and 7, lines 17 and 36. A Gradenstuhl Trinity also occurs toward the beginning of Estoire in Paris, BN fr 344, f. 9, at the words, ‘Cil sains esperiz’ (Sommer, I, p. 25, line 6). 23 See Haseloff, Psalterillustration; Leroquais, Les psautiers. 24 Leroquais, Les livres d’heures. 25 For a rare instance of what I think is the influence of Grail imagery reflected in Last Supper iconography, see Stones, ‘Madame Marie’s Picture-Book: a precursor’. See also note 53 below, and Stones, Gothic Manuscripts, Part I, vol. 2, Cat. no. III–57. 26 Sommer, Vulgate Version, I, p. 13, line 27. 20

440

because the Crucifixion is by far the commonest subject of the Passion of Christ in art, and medieval artists would certainly have been called upon to depict it. It is also a subject in which, as early as the late ninth century, a vessel was often depicted at the foot of the cross, into which the blood of Christ could fall.27 Nothing is said in the Estoire of Joseph collecting the blood of Christ at the Crucifixion itself, although his presence at it, narrated in the Gospels, is confirmed in the Estoire. It is surprising, then, that two groups of manuscripts show him holding the Grail at the cross: the Additional, Royal and Amsterdam group of ca. 1315, and the two early-fifteenth-century manuscripts sold in Paris by Jacques Raponde, one of which was owned by Jean de Berry and inherited by Jacques d’Armagnac. In the copy now in Amsterdam (Fig. 4), Joseph is shown seated on the ground at the foot of the cross, holding the ‘escuele’, with Mary and John also present. The other two early-fourteenth-century copies show a similar depiction, including Joseph holding the ‘escuele’, and the two thieves as well as Mary and John (Add. 10292, f. 3v and Royal 14 E. III, f. 7). The fifteenth-century books show the Grail as a chalice. The shallow bowl used by Joseph in the Add./Roy./Amsterdam group corresponds in shape to the ‘escuele’ of the Estoire text, but is a shape that carries other resonances as well. A shallow bowl is also the vessel depicted in Old Testament illustration for collecting the blood of sacrificial animals, or other hallowed liquid, like the water Moses produced by striking the rock.28 Pertinent parallels for the ‘escuele’ in Old Testament imagery produced by the artists of Add./Roy,/Amsterdam/Ryl./Douce have not so far come to

27 The most comprehensive study is the unpublished dissertation by Malcor, The Chalice at the Cross, which, as the title implies, includes many dubious Grails. The earliest depiction of a vessel in association with the Crucifixion is in the Stuttgart Psalter, made in northern France in the early ninth century, and the subject was also popular in Anglo-Saxon England, Ottonian Germany, and Mozarabic Spain; I reserve a detailed examination for another occasion. 28 Illustrations abound, particularly in the Moralised Bibles of the second quarter of the thirteenth century: see Laborde, ed., La Bible moralisée conservée à Oxford, Paris et Londres; Haussherr, ed., Bible moralisée... Codex Vindobonensis 2554 der Österreichischen Nationalbibliothek, with reference to many articles by the same author; Guest, ed., Bible moralisée: Codex Vindobonensis 2554, Vienna, Österreichische Nationalbibliothek; see also Lowden, The Making. A Moses example is reproduced from Bodl. Bod. 270b. f. 50, in Stones, ‘Sacred and Profane’, fig. 8.

SEEING THE GRAIL

441

light, despite the prolific artistic output of these illuminators,29 but this type of vessel can be found in biblical illustration of the second quarter of the thirteenth century, in manuscripts stylistically related to the Rennes Estoire. Made ca. 1220, this is the copy that contains the earliest depiction of a Grail whose interpretation is certain.30 The Rennes manuscript lacks an image of Joseph at the Crucifixion, and so I return to it below, in my discussion of Josephé’s journey to Norgales, the one place in the illustrations of the Rennes manuscript where the Grail makes an appearance. In the other illustrated manuscripts that include Joseph at the Crucifixion, the vessel he holds is shown as a chalice. The versions in Paris, Ars. 3480, f. 483, and BNF fr. 120, f. 520, both made in Paris soon after 1400 (the latter owned by Jean de Berry), show a simpler Crucifixion with Mary and John, and Joseph of Arimathea as a much smaller figure robed in a long mantle kneeling on the left, a chalice in his outstretched hands. The Crucifixion is not depicted in Jacques d’ Armagnac’s copy of the Estoire of ca. 1475 (BNF fr. 113), but it is interesting to note that the opening miniature of the Queste part of the same set of volumes (BNF fr. 116, f. 607) is a Crucifixion where five angels catch the blood of Christ in chalices, following a model common in crucifixion iconography, and Joseph is not present. The Entombment Only occasionally is Joseph shown actually collecting the blood of Christ at the Entombment, and in the ‘escuele’. Both the Bonn manuscript made in 1285 and the closely related and possibly earlier manuscript, Paris, BNF fr. 19162, f. 631 include this as the final scene in the composite arrangement of images that opens the Estoire — in Bonn 526 the arrangement is six square miniatures grouped two deep across the three columns of the text and in BNF fr. 19162 there are three miniatures arranged vertically in the first of Discussed in Stones, ‘Another Short Note’, reprinted in these essays. For mention of earlier depictions of sacred vessels which eager writers have interpreted as ‘Grails’, see note 27 above. 31 This is stylistically closer to the miscellany, Paris, Bibl. Sainte-Geneviève 2200, part of which was written in 1277, than to Bonn 526, made a decade later (Stones, Illustrations of Lancelot, p. 459; and now ead., Gothic Manuscripts, Part I, vol. 2, Cat. no. III–116). It is interesting that neither of the two other Lancelot-Graal manuscripts that these craftsmen also made — BNF fr 110, close to Bonn 526, and BNF fr 24394, close to BNF fr 19162 — includes a picture of this scene. Indeed, neither shows the Grail at all. For Bonn 526 see now Stones, Gothic Manuscripts, Part I, vol. 2, Cat. no. III–121. 29 30

442

two columns and one at the top of the second column. To my knowledge, the only other manuscript to include this motif is the set made for Jacques d’Armagnac, Paris, BNF fr. 113, f. 7, ca. 1475, where the Grail is a flat, paten-like receptacle (Fig. 5).32 The Ark Joseph builds a wooden ark to house the Grail.33 Representations of the ark in Estoire manuscripts are again rare, and there are interesting links between the various depictions of the Grail-ark in Estoire and the ark of the Covenant in Old Testament iconography, of which north French and Flemish manuscripts offer some interesting depictions in the second half of the thirteenth century.34 In BNF fr. 344, f. 11 v (made in Lorraine, ca. 1300) it is shown as a pink chest on four legs, with a lock, its contents not revealed until later (f. 26) when Joseph’s son, Josephé, outside the ark, shows King Eualach and Queen Sarracinte what is within — a gold ‘escuele’ and three gold nails (Fig. 6).35 In the Le Mans manuscript, Josephé, at Christ’s invitation, enters the ark and sees Christ Himself surrounded by five angels: in the image, the structure of the ark is dispensed with, and we see Josephé surrounded by what he sees. The text says the angels hold three bleeding nails, a bleeding lance, a red cloth, a bloody sponge held upright, and a bleeding scourge, and describes these objects as ‘les ames par coi li gugieres qui chi est, uainqui la mort et destruist’ (‘the arms with which the judge of the world

See Blackman, The Manuscripts of Jacques d’Armagnac, p. 505. Sommer, Vulgate Version, I, p. 20, lines 33–34, ‘de cest bois feras tu a mescuele une huche’ (‘from this wood you will make an ark for my vessel’). 34 Again, examples abound in the Moralised Bibles, on which see note 28 above. There are also several notable examples in the Old Testament Picture Bible, New York, Pierpont Morgan Library and Museum, M.638 with leaves in the BNF, Paris, and the J. Paul Getty Museum, reproduced in full in Old Testament Miniatures (on which see also Weiss et al., The Morgan Crusader Bible Commentary; Stones,’Questions of Style and Provenance’); other examples are in the Psalter of St. Louis, Paris, BNF lat 10525, reproduced in Le psautier de saint Louis, and the hours with a calendar of Arras in the Musée de Lille, SA 367, reproduced in Stones, ‘BN, fr 95 and Yale 229’, fig. 8.40; and ead., Gothic Manuscripts, Part II, vol. 2, Table of Books of Hours. Bibles provide another obvious source of comparisons, a topic too large to outline here. Some provocative examples are reproduced in Rosenau, Vision of the Temple, to which Martine Meuwese kindly drew my attention. For more on links between the Grail chapel and the Temple, see note 57 below. 35 Not in Sommer. 32 33

SEEING THE GRAIL

443

conquered and destroyed death’).36 The illustration shows four angels, not five, and only the cross (not mentioned in the Le Mans text version), nails, and lance are shown, while a man on horseback holding a white shield and raising his sword against a snail provides an inverted play on the notion of ‘arms’.37 Even so, as I have noted elsewhere,38 both text and picture represent an extremely early39 version of the ‘arma Christi’ the devotional image based on emblematically depicted Instruments of the Passion that only in the early fourteenth century came to be widespread, often with accompanying prayers, as a focus for private prayer and meditation.40 A different, in some ways simpler, view of the ark and its contents is shown in BNF fr. 105, f. 19 (a Parisian product of the 1320s–40s),41 where Josephé stands outside and 36 Hucher, Le Saint Graal, is a complete edition based on this manuscript. This passage, in Hucher, II, pp. 174–75, corresponds with some variants to Sommer, Vulgate Version, I, p. 32. For related manuscripts, see note 39 below. For a reproduction, see Stones, ‘BN, fr 95 and Yale 229’, fig. 8.20. This manuscript is now reproduced on the Lancelot-Grail Project site http://www.lancelot-grail/pitt.edu and on BVMM; see also Stones, Gothic Manuscripts, Part I, vol. 2, Cat. no. III–26. 37 For other examples of the motif of knight fighting snail, see Randall, ‘The Snail in Gothic Marginal Warfare’. 38 ‘BN fr 95 and Yale 229’, p. 224. 39 For discussion of Le Mans 354 in relation to BNF fr 342 of 1274 and BNF fr 770, see Stones, The Illustrations of Lancelot, ch. 3 and ead., Gothic Manuscripts, Part I, cat. no. III–26. The same scribe, Walterus de Kayo, signed his name and the date 1282 in a copy of the Image du Monde, Paris, BN fr 14962, a discovery made by Terry Nixon and cited in Stones, ‘The Illustrated Chrétien manuscripts and their artistic context’, I, pp. 237–38. BNF fr 770 and BNF fr 342 are included in Bräm, ‘Ein Buchmalereiatelier in Arras um 1274‘, a study which omits Le Mans 354 and numerous other books I have identified as part of this stylistic group; and the attribution to Arras is certainly open to challenge since the use of the liturgical and devotional books is of Douai (in the diocese of Arras, though not Arras itself ). I think Le Mans 354 was probably made around 1285. 40 For ‘Arma Christi’, see Berliner, ‘Arma Christi’; Suckale, ‘Arma Christi’; and my summary in Wace, La Vìe de sainte Marguerite, pp. 195–96. See also Sandler, ‘Jean Pucelle and the Lost Miniatures’. The lance, nails, and crown of thorns appear earlier in monumental art, together with the cross, in the context of the Last Judgment, for instance in the sculpture of the west facade at Saint-Denis (1137–1140), which had relics of the Passion given in the eleventh century but quickly claimed to be a donation of Charles the Bald. See Gerson, ‘Suger as Iconographer’, especially notes 28 and 31, with references to Beaulieu and Conques. The Apocalyptic Vision tympanum of 1186 at Santiago de Compostela has been linked with Grail iconography and an interest in Perceval: see Moralejo Álvarez, ‘Entre el Grial y la Divina Commedia’. 41 It is illustrated throughout by a painter closely related to, if less competent than, the Master of the Roman de Fauvel, Paris, BNF fr 146, made ca. 1316 in Paris. I call him the

444

looks into the ark, here painted gold and resting on silver (or tin) legs, seeing Christ crucified flanked by two seraphim inside it (Fig. 7). We shall see that the exterior/interior views in BNF fr. 344 and 105 are significant in relation to other ways artists found to screen the Grail from view, while the more fully developed vision of the interior shown in the Le Mans manuscript also finds echoes in the depiction of the liturgy in other manuscripts. The screen functions as a protection for the viewer as well as for the Holy Vessel. Looking into the ark is perilous. Nascien is so bold as to try to see what is in the Holy Vessel itself, having removed the ‘platine’ which covered it — and is blinded.42 An angel restores his sight by anointing his eyes with blood, collected in a ‘boiste’ (box), from the Holy Lance.43 The blinding, so far as I know, is not included in the visual tradition, and the curing is also rare. The Bonn manuscript, however, includes the latter (f. 16v), showing the angel’s ‘boiste’ as a round container with a round finial on the lid; the presence of the Grail, shown as a chalice, in the top right comer of the miniature behind Nascien serves as a reminder of what caused his blindness in the first place (Fig. 8). The Grail Liturgy The liturgy of the Grail follows, celebrated at great length in the text by Josephé, son of Joseph of Arimathea and first Christian bishop, according to the Estoire. Particularly notable about the text of Estoire, composed at least as early as c. 1220 (because of the likely date of Rennes 255 made at about that time), is the extremely detailed description of the elevation and transubstantiation, including explicit mention of the ‘cors autre tel comme d’un enfant’ (‘actual body like that of a child’) that Josephé finds in his hands, and that Christ orders him to tear apart.44 As we shall see, this is also a feature of the Grail liturgy in the Queste, probably composed Sub-Fauvel Master, and discuss his artistic personality in ‘The Artistic Context of le Roman de Fauvel’. Interestingly, the iconographic selection in BNF fr 9123, illustrated in part by the same artist, is different, and omits this scene of Joseph and the ark. See now Stones, Gothic Manuscripts, Part I, vol. 2, Cat. no. I–61 and, for BNF fr 9123, 105 and BR 9246, Stones, ‘L’Estoire del saint Graal dans la version pour Jean-Louis de Savoie’ and www.lancelot-project. pitt.edu. 42 Sommer, Vulgate Version, I, p. 79. 43 Ibid., I, p. 80. 44 Ibid., 1, p. 40.

SEEING THE GRAIL

445

at about the same time or shortly before the Estoire, and depictions of the transubstantiation in Queste illustration are considerably more literal than those that accompany the Grail liturgy in Estoire. At the time of the composition of these texts, the elevation of the Host during the Mass was a relatively recent phenomenon, decreed at the synod of Paris between 1205 and 1208, under Bishop Eudes de Sully, and surrounded by considerable debate about the precise moment of transubstantiation in relation to the two elements of the Eucharist.45 Debate about transubstantiation as such was already under way very much earlier, as shown in the opposing positions taken by Berengar of Tours († 1088), in favor of the symbolic presence of Christ, and Lanfranc of Bec, advocate of the real presence, whose view prevailed.46 The illustrations in an eleventhcentury copy of the Life of St Aubin of Angers (a text composed in the Merovingian period), show an early depiction of a Eucharistic miracle. The Life includes scenes where the saint is compelled to bless unconsecrated wafers (eulogia) that are to be given to an excommunicate, who then dies at the sight of the objects. The illustrator has shown a transformation that goes beyond the text, to reflect the contemporary Eucharistic debate, by inscribing the eulogia with the letters IHS and XPC, the Name of Christ, transforming them thereby into transubstantiated Hosts.47 We shall see that the Host at the Grail liturgy in Estoire manuscripts is sometimes similarly inscribed, carrying similar liturgical connotations. See Kennedy, ‘The Moment of Consecration,’ esp. pp. 146–47, and Dumoutet. Le désir de voir l’hostie, noting also the critical analysis of Dumoutet in Wilmart, Auteurs spirituels, p. 371. n. 2. See also Maccarrone, ‘Innocenzo III teologo dell’Eucharistia’, and Browe, S.J., Verehrung, esp. pp. 70–88. The custom of elevating the Host was already attested by Hildebert of Le Mans († 1133), cited by Carrasco, ‘St Albinus of Angers’, p. 339, with reference to Migne, PL. 46 See Montclos, Lanfranc et Bérenger, cited by Carrasco, p. 338. 47 Carrasco, pp. 337–41. A second eucharistic miracle depicted in an eleventh-century manuscript is the final communion of St Denis, in which the host is given the saint by Christ, depicted in Paris, BNF lat 9436, f. 106v. This is reproduced and discussed in Vloberg, L’Eucharistie dans l’art, II, p. 186, also cited by Carrasco, p. 339. Further depictions of Eucharistic miracles cited by Carrasco are listed in Rohault de Fleury, La Messe, IV, pp. 1–36; de Boom, ‘Le culte de l’eucharistie’, pp. 326–32; Elbern, ‘Über die Illustration des Messkanons’. To these may be added the miracle in which Guillaume d’Aquitaine is reconciled to the church by a consecrated Host held out before him by St Bernard of Clairvaux, depicted in the pair of illustrated antiphonaries probably made for the Cistercian Abbey of Cambron (Diocese of Cambrai), and now divided between the J. Paul Getty Museum, Los Angeles, MSS 83.ML.99 (Ludwig VI. 5) and MS 44, and the Hill Monastic Manuscript Library, 45

446

The appearance of Christ as a Child in the Host, described unambiguously in both Estoire and Queste liturgies, also has parallels in earlier accounts of Eucharistic miracles and in their depiction. Among the earliest accounts is the one given by Guibert de Nogent († 1124),48 and among the earliest depictions is a miracle witnessed by St Edward the Confessor.49 A generation later, growing devotion to the Eucharist was marked by the founding of a feast of Corpus Christi in the diocese of Liège in 1246. By the early fourteenth century, the reservation of the Eucharist had become a general practice and Corpus Christi had become a universal feast, and by 1337– 1339 the Bolsena miracle of the Bleeding Host, and the commissioning of the Corpus Christi liturgy (once attributed to Thomas Aquinas) had been depicted in a series of eight scenes on the reliquary made by Ugolino di Vieri in 1337–1338 to house, in the Cathedral of Orvieto, the relic of the corporal (cloth) on which the Host bled; 50 and three explicit renderings of Collegeville, Minnesota, MS 8 (Bean MS 3), with fragments in Stockholm, New Mexico, and elsewhere: see Stones and Steyaert, Illuminations, Glass and Sculpture in Minnesota Collections, no. 4; von Euw and Plotzek, Sammlung Ludwig, I, pp. 280–84; ‘Acquisitions/1992,’ The J. Paul Getty Museum Journal, 21 (1993), p. 111; and, now, Stones, Gothic Manuscripts, Part I, vol. 2, Cat. no. III–48. The only other known representation of this subject is in the early fourteenth-century Hungarian Legendary divided among the Morgan Library, the Vatican, and Berkeley, see Magyar Anjou Legéndarium, ed. Levärdy; Ungarisches Legendarium, ed. Morello, Stamm, and Betz; Vaticana (exhibition catalogue), ed. Plotzek, no. 48, and Corsair. 48 Cited by Carrasco, St Albinus of Angers, p. 339, with reference to De pignoribus sanctorum, 1. 2: PL 156: 616, and by Dumoutet. Le désir, pp. 46–47. Browe, Die Verehrung, pp. 100–01, cites a ninth-century miracle described by Paschasius Radbertus († 864). For other textual accounts see Corblet, Histoire du sacrament de l’Eucharistie, I, pp. 447–515. See also Vloberg, ‘Les miracles eucharistiques’. 49 He witnessed Christ’s real presence in the consecrated host; the miracle had been depicted by ca. 1250 in La estoire de seint Aedward le rei, Cambridge, UL Ee.3.59, p. 37. See Rubin, Corpus Christi, pp. 118–42, and the full description of the manuscript in Morgan, Gothic Manuscripts, II, cat. no. 123. 50 The reliquary is preserved in the Cappella del Corporale, situated off the north side of the north transept in the Cathedral of Orvieto, whose rebuilding on a grand scale appears to have been closely associated with the occurrence of the miracle close by at Bolsena and its authentication at Orvieto by Pope Urban IV. The miracle and surrounding events are also depicted on the east wall of the chapel, in frescoes by Ugolino di Prete Ilario (1357– 64), including a scene showing the pope commissioning the office of Corpus Christi. The corporal reliquary is fully reproduced in Dal Pogetto, Gli Smalti di Orvieto; see also de Castris, ‘Il reliquiario del corporale a Orvieto’, and the not very scientific study by Testa, La Cattedrale di Orvieto, pp. 80–83. Another famous example of the miracle of Bolsena coupled (indirectly, since they are in different stanze) with the transubstantiation (showing the Host in a monstrance) is among the frescoes by Raphael and followers, commissioned by

SEEING THE GRAIL

447

the transubstantiation, also by Ugolino, had been painted on the walls of the reliquary chapel. 51 The Host is shown as Christ in the form or a small naked boy, holding (in two of the three images) a cross staff and facing frontally, visible to viewers both within the scene and outside it, unlike the side-on view presented in the Life of St. Edward, which limits the perception of the event to those in the composition who witness it. Both the Estoire and the Queste must be read against this background, which is also significant in relation to the illustrations devised for them. 52 We shall see that illustrations of the Grail and the liturgy of the Grail in Queste manuscripts sometimes include a depiction of transubstantiation that is just as explicit as it is in the Life of St Edward and in Orvieto, and there are even a few instances where, conversely, aspects of Grail iconography based on the Estoire or the Queste seem to find reflection in devotional or liturgical depictions of the Last Supper and the Mass.53 Each of the manuscripts that depict the Grail liturgy in Estoire — and only a few of them do — shows a slightly different aspect of the events described, and emphasizes a position in relation to the elevation and transubstantiation to a greater or lesser degree. They are perhaps rather reticent by comparison with the Life of St Edward and the Bolsena reliquary and frescoes, but at least one illustration datable in the middle of the thirteenth century clearly suggests a link between the growing cult of the Eucharist and the Grail, and Pope Julius II ca. 1507–17 for the Vatican stanze. See Hersey, High Renaissance, pp. 129–76 (accepting the institution of the feast in 1264 as a result of the miracle, p. 150). I thank David Wilkins and Matthew Roper for these references and for helpful discussion of depictions of transubstantiation in Italy. 51 Although these frescoes were completely repainted between 1855 and 1860, the sinopia, discovered during the restorations of 1977–1978, clearly shows that the details of transubstantiated Host as Christ/boy holding a cross are accurate in the restoration (Testa, pp. 198–99). One of the transubstantions of Orvieto is reproduced (but not actually mentioned in the text) in Camporesi, ‘The Consecrated Host’, which focuses on perceptions about the swallowed transubstantiated Host assimilating and being assimilated by the human body during the process of digestion. 52 See particularly Dumoutet, Corpus domini; and Anitchkof, ‘Le saint Graal et les rites eucharistiques’. Other critics have emphasized the rôle of a Cistercian ethic (Pauphilet, Études sur la Queste) or the concept of grace (Gilson, ‘La mystique de la grâce dans la Queste’, and id., Les Idées et les lettres, pp. 59–91). 53 A case in point is a depiction of a celebration of mass in the early-fourteenth-century Breviary of the Cathedral of Saint-Lambert, Liège, Darmstadt, Hessische Landes-und Hochschulbibliothek 394, to which I shall return elsewhere. Another case is the Last Supper miniature in the Livre d’images de Madame Marie, cited in note 25 above.

448

by the end of the century, several ways of depicting the Grail liturgy had been developed. Probably the earliest representation — since the earliest surviving manuscript, Rennes 255, does not illustrate the liturgy, or the ark, at all — is the small historiated initial made perhaps in Normandy about 1250, in UCB Berkeley 106, f. 126, where Josephé stands before an altar on which is a draped chalice. A golden Host hovers in midair, and the Head of God looks down from a cloud above. This is a depiction that certainly suggests the veneration of the Eucharist, and perhaps alludes also to its reservation — neither of which is depicted in liturgical illustration at a comparable date. But not all Estoire depictions herald these new trends in contemporary liturgical practice. Much more commonplace is the illustration in BNF fr. 770, f. 19v (made on the Artois-Flanders border, ca. 1280), which simply depicts an altar with a chalice on it and a group of figures kneeling before it. But by the end of the century, Estoire illustration shows several parallels with the iconography of transubstantiation. In BNF fr. 95, f. 18 (made in the diocese of Thérouanne. ca. 1290) the Grail liturgy illustration is similar to that of BNF fr. 770, but what is shown is the moment of the elevation of the Host, with Josephé bareheaded and his followers kneeling on the ground behind, 54 a depiction similar to what is commonly shown at the beginning of the canon of the Mass in missals. 55 Josephé elevates a Eucharistic wafer, not the human figure, but the Host is inscribed with the letters IHS of the Name of Jesus, and so contains a cryptographic reference to the transubstantiation. On the altar is the partly veiled Grail, shown as a chalice. The right-hand part of the opening miniature of Estoire in Yale 227 (Fig. 2), is similar: Josephé, in the presence of three kneeling knights, celebrates Mass at an altar in which arc a draped chalice and a large Eucharistic Host with a cross between dot motifs inscribed on it. But at the place in the text where the liturgy is described, there is no corresponding illustration. The most extensive pictorial treatment of Josephé’s liturgy in the Estoire occurs in the Amsterdam and Royal copies, both made in Flanders by the same craftsmen ca. 1316.56 Both manuscripts include two scenes of

54 Reproduced in Stones, ‘BN, fr 95 and Yale 229,’ fig. 8.15. See now ead., Gothic Manuscripts, Part I, vol. 2, Cat. no. 124. 55 Leroquais, Sacramentaires et missels. 56 They are reproduced and discussed in Stones, ‘BN fr 95 and Yale 229’, figs. 8. 18 and 8. 19.

SEEING THE GRAIL

449

Josephé’s liturgy, but the third manuscript made by this team, Add. 10292, omits illustrations of the liturgy altogether. The first image in Amsterdam, f. 18, shows Josephé in the ark, robed as a bishop, accompanied by an angel holding a silver ewer and two more angels kneeling to the left of the ark, one holding an incense-boat and swinging a thurible. As his father Joseph looks into the ark, shown as a Gothic church, 57 he sees a small altar, covered with a white cloth and a red cloth of ‘samit’ (precious fabric, of uncertain translation, omitted in the miniature); on the altar are three bleeding nails and the top of a bleeding lance (held, in the text, by hands which are omitted in the image),58 a silver chalice-like vessel on a stem with a lid, mostly covered by a cloth painted red in the miniature but described as white in the text, and partially veiled from our sight but revealed to Joseph: and the silver bowl-shaped ‘escuele’ — the Grail. The presence of both the ‘escuele’ of the text and also a ciborium suggests a conflation of sacred objects to include both what is called for in the text and what was familiar from contemporary liturgical practice; at the same time this image is reminiscent of the depiction of Josephé in the ark as shown in the Le Mans manuscript. The Royal 14 E. III version is similar but much simpler in treatment, omitting the elaborate architectural ark, also omitting the hands holding candles and the lance and nails; but it does include a red cross held by a hand above the altar (Fig. 9), and so is not merely a simplification of the Amsterdam image. The hands holding the instruments of the Passion (the three nails and lance) are rarely shown in Estoire illustration: one isolated example occurs in a later manuscript, BNF fr 113, f. 18v, one of those made for Jacques d’Armagnac. There, the hands holding cross and candle emerge from the reredos behind the altar, on which rest the three nails and the tip of the lance, a round ciborium-like vessel, and a flat, dish-like paten. Is the Grail the ciborium or the paten? The more elaborate vessel is undoubtedly the ciborium, yet in the rare depiction of the collecting of Christ’s blood at the Entombment found earlier in this manuscript, Joseph used a shallow dish that is remarkably similar to the paten shown on this altar. So this time the

This is quite different from the centrally planned ‘Grail Temple’, based on the Temple of Jerusalem and/or the Church of the Holy Sepulchre, described in Albrecht’s Jüngere Titurel, composed in Middle High German ca. 1270, which is the starting point for Ringbom’s wideranging study: Ringbom, Graltempel und Paradies. See also Rosenau, Vision of the Temple. 58 Sommer, Vulgate Version, I, pp. 33–34. 57

450

proliferation of sacred vessels is not simply based on the multiple vessels of the Eucharist, but also links back to Joseph’s pious act earlier in the story. Josephé, wearing civilian clothes and bareheaded, kneels, alone, before the altar; back in the doorway behind him stand four wingless angels, one of whom holds a situla (bucket) and aspergillum (holy-water sprinkler), while another holds a processional cross. In BNF fr. 113, f. 18v, both picture and rubric focus on Josephé contemplating the Grail in a chapel ‘C(i) iosephez estoit devant le saint Graal a genoulx’ (‘Here Josephé knelt before the Grail’). Four angels stand behind him in the doorway, one of whom holds a situla and aspergillum; and on f. 21 v, three angels administer the communion to Josephé and his company.59 The second miniature in Amsterdam, BPH 1, on f. 21, shows Christ and Josephé administering the sacrament — a Eucharistic Host, to the kneeling assembly. Although Christ is present in the text, and is described as wearing priestly robes, the communion is administered, in the text, by Josephé alone.60 Undoubtedly Christ’s actions as shown in this image can be explained by reference to the Grail liturgy and its illustrations in the Queste, in which it is Christ, not Josephé, who distributes the Host to the knights, as I show below; but the addition of Christ here is all the more interesting as none of these three manuscripts includes an illustration to the liturgy in the Queste.61 The Royal 14 E. III version is simpler, showing the altar draped with a cloth on which are a gold chalice and wafers; Christ holds the Grail, shown as a gold bowl-shaped vessel, and blesses a kneeling man, probably Joseph, while Josephé, robed as bishop, administers the Host to the rest of the assembly, and two angels swing thuribles above. Josephé’s Journey to Norgales The Grail is subsequently taken by Josephé and his companions to Norgales, a journey miraculously accomplished on the hem of his garment.62 It is in this context, surprisingly, that the earliest depiction of the Grail is found, in 59 Blackman, The Manuscripts, p. 505. No image of the liturgy is in BNF fr 112 or BNF fr 117–20, the other Vulgate Cycle manuscripts owned by Jacques d’Armagnac, BNF fr 112 is the special version he commissioned; BNF fr 117–20 is the set he inherited from his greatgrandfather Jean de Berry. 60 Sommer, Vulgate Version, I, p. 34. See Stones, ‘BN, fr 95 and Yale 229’, fig. 8.19 . 61 Illustrations of the Queste liturgy in other manuscripts are discussed below’. 62 Sommer, Vulgate Version, I, p. 211.

SEEING THE GRAIL

451

the Rennes mansucript, made probably in Paris, ca. 1220. 63 The journey is simply depicted in Rennes 255, where one of Josephé’s followers, wearing a round-crowned skullcap-style Jewish hat rather than a mitre, carries the silver (?) ‘escuele’ in his hands (no depiction of the hem of the garment is included).64 In BNF fr 749, f. 98, the silver ‘escuele’ has been endowed with a gold cover with a round finial on the top.65 In BNF fr 344, f. 65v (Fig. 10), the Grail is borne in procession, much like the ark of the covenant (a link with the ark shown earlier in BNF fr. 344 and a few other manuscripts), or a reliquary, and even more like a monstrance (a vessel for the display of the consecrated Host), since a circle has been lightly sketched within the tabernacle-like shrine that is carried on rods on the shoulders of two men. The earliest surviving example (second quarter of the thirteenth century?) of a monstrance for the reservation of the Eucharist is the one preserved in the treasury of the church of Notre-Dame at Saint-Omer, from the Cistercian abbey of Clairmarais near Saint-Omer (diocese of Thérouanne). It is a vessel on a stem in which the Eucharistic element could presumably be seen through the horizontally disposed openings around the perimeter.66 But, as See Stones, ‘The Earliest’ (above, note 5), where the date is argued on the basis of stylistic parallels with royal manuscripts of c. 1220. The carrying of the Grail to Norgales is fig. 3a. 64 BNF fr 770. f. 98 and BNF fr 9123 f. 79 show a very similar, non-text-specific, depiction. UCB Berkeley 106, Le Mans 354, BNF fr 95, and Yale 227 have no illustration of this; it is in BNF fr 113, f. 88, but without the Grail. See Blackman, ‘The Manuscripts’, p. 506. 65 This manuscript has a large number of marginal notes for the illuminator: the one on this page says ‘Josef q(ui) enporte le sent vaisel (et) gens avec lui’. It is by the same painters as London, BL Harley 4979, discussed in my ‘Three Illuminated Alexander Manuscripts’. See also ead. and Ross, ‘The Roman d’Alexandre’; and now Pérez-Simon, Mise en roman et mise en image, passim. 66 Gauthier, Les routes de la foi, no. 66, in the treasury of the church of Notre-Dame, Saint-Omer (Pas-de-Calais); see also Trésors des églises de l’arrondissement de Saint-Omer, (exhibition catalogue), no. 3, pp. 42–45, where it is identified as a pyx, a type of vessel that is not normally perforated. Another pertinent example is the fully visible lunar stand, made to support the Host, within rock crystal surrounded by an architectural framework, seen on the monstrance made in Paris for another Cistercian house, the nunnery of Herkenrode (near Hasselt, province of Limburg, diocese of Liège) in 1286. An inscription indicates it was commissioned for Herkenrode by prioress Adelheid von Diest in 1286, and it bears the earliest dated Parisian silver stamp. See Andrieu, ‘Aux origines du culte du Saint-Sacrement’, and Schatz aus den Trümmern (exhibition catalogue), no. 27. It now belongs to the church of Saint-Quentin in Hasselt (Belgium) but is on loan to the Stellinwerff-Waerdenhof Museum in Hasselt. 63

452

with the ‘Arma Christi’, the depiction of a Corpus Christi procession with the Host in a monstrance would seem not to appear in liturgical manuscripts for another generation or so.67 The version of Josephé’s miraculous journey shown in Amsterdam BPH 1 reverts to the depiction of the Grail found elsewhere in this manuscript, and shows Josephé, robed as a bishop, carrying the ‘escuele’, and the assembled company traveling on the hem of Josephé’s alb; this is also included in mirror image in Add. 10292. In Royal 14 E.III (f. 66v), the journey is depicted, showing Josephé holding the alb on which stand the traveling figures, but no Grail is shown. The Grail Transmitted by Josephé to Alain Thereafter, shown again as the ‘escuele’, the Grail is handed by the dying Josephé to his follower Alain.68 What Josephé hands to Alain in BNF fr. 344, f. 78v, is a veiled vessel with a cross on top of the veil and, as in the procession to Norgales, a circle lightly drawn on the veil, suggesting again the analogy between the Grail and a monstrance for the display of the Host (Fig. 11).69 BNF fr. 770, f. 117v, Amsterdam BPH 1, i, f. 114v, and BNF fr. 105, f. 122,70 show the Grail again as the ‘escuele’, whereas in Royal 14 E. III, and Add. 10292, f. 73, it has become a covered ciborium with a stem, painted entirely in gold, as also in Yale 227. In Le Mans 354, this might be the episode represented as the last of the scenes in the composite opening miniature on f. 1, although the curious hand gesture of the standing figure holding the ‘escuele’ and the presence of a companion suggests that this is a curing scene rather than the transmission of the Grail, and I return to it below. One variation on the transmission theme occurs in BNF fr. 749, where it is the shield argent a cross gules,71 rather than the Grail, which is handed by Josephé to his followers. 67 For Christ as priest, and the Host displayed in a monstrance, see Avril, ‘Une curieuse illustration de la Fête-Dieu’, cited in Stones, ‘BN, fr 95 and Yale 229’, note 118. See also the Eucharistic miracles referred to in note 47 above. 68 Sommer, Vulgate Version, I, p. 286. The Rennes manuscript shows Alain weeping at Josephe’s deathbed, but no Grail is shown. 69 Not according to Loomis, The Grail, p. 280, where Josephé hands Alain an ‘escuele’. 70 It is worth noting that in this manuscript Josephé is shown standing, not lying on his deathbed as in the other depictions of this episode. 71 These are standard heraldic terms: argent=silver (here white is used as a substitute), gules=red.

SEEING THE GRAIL

453

King Alphasem The final episodes in Estoire include miraculous acts performed by the Grail concerning King Alphasem: the cure of his leprosy, effected by the Grail after his baptism, and his wounding in the groin, as punishment for having presumed to sleep in the Grail Castle.72 Neither is commonly illustrated, ceding place rather to concluding incidents about the tomb of King Lancelot, ancestor of King Arthur’s knight of the same name,73 or Nascien and King Mordrain at Josephé’s tomb,74 or Celidoine defeating his enemies,73 or Celidoine and his son bidding farewell to Nascien.76 None of these other episodes involve the Grail. A few exceptional miniatures do depict the cure of King Alphasem or his wounding: Bonn 526, Royal 14 E. Ill, and Add. 10292 all combine the baptism and cure of the king into a single scene, in which the Grail is present, shown in Bonn as a chalice (without an accompanying cross) and in Royal and Add. as a covered ciborium. The wounding alone is shown in Amsterdam BPH l,77 where a cleric stands by the king’s bed, wearing a maroon cope over his alb, holding in his left hand the Grail — shown as the usual silver bowl — and in the other a lance; in the corresponding miniature in Add. 10292, f. 74, a bearded man wearing a hooded robe and painted orange (to represent flames), pierces the king through the bedclothes with a spear, but the Grail is not present,78 while in BNF fr. 113, f. 113v, an angel hovers above the wounding scene, holding the Grail.79 Neither picture quite corresponds to the text, where the Grail has disappeared before the appearance of ‘vns hons ausi comme tous enflammes’ (‘a man as if all in flames’) who does the wounding, and there is nothing in the text that justifies the representation of this figure as a cleric. The previous scene in BNF fr. 113, on f. 112v, shows King Alphasem and his followers standing before Alain, having been Sommer, Vulgate Version, I, pp. 288–89. As in Amsterdam, BPH 1 and Royal 14 E. III. 74 Rennes 255, f. 98v; Tours 951, f. 156v. 75 Add. 10292, f. 74v; Yale 227, f. 138. 76 Rennes 255, f. 98v; Yale 227, f. 13 lv. 77 Sommer, Vulgate Version, I, p. 289. In the parallel miniature in Add. 10292, f. 74, the wounding is also shown but the Grail is omitted. 78 Reproduced in Matthews, Grail: Quest, p. 76. 79 Blackman, The Manuscripts, p. 508. 72 73

454

cured, without the Grail depicted. One further image might be interpreted as showing a combination of the Grail’s appearance and the wounding, again not quite justified by the text: in Le Mans 354, f. 1, an uncrowned figure in bed is admonished by two men who stand behind the bed raising accusing index fingers at him; one of them holds the Grail, shown as a bowl, and the other, taller, is clad in an orange tunic, corresponding somewhat to the ‘hons tout enflammes’ of the text. An alternative interpretation, proposed above, is the transmission of the Grail from the dying Josephé to Alain. It is worth noting the contrast between the paucity of illustrations of the healing of King Alphasem in the manuscript tradition as a whole, and the prominence accorded it in popular twentieth-century studies of Arthurian legend, such as the much-read Ritual to Romance by Jessie L. Weston.80 The healing episode was allusively drawn upon by T. S. Eliot in The Waste Land, as his notes to the poem indicate.81 Harvard’s copy of the 1920 edition of Weston’s book is inscribed by T.S. Eliot: ‘This is the copy I had before writing The Waste Land. T.S.E.’ 82 We shall see that the parallel event in the Queste, the curing of the Roi Mehaignié, enjoyed an equally spare iconographic tradition.

the grail in QUESTE manuscripts In the Queste del saint Graal, depictions of the Grail are in general even less numerous than they are in the illustrations of the Estoire. They occur in just a few key scenes, shown in a small selection of otherwise quite fully illustrated manuscripts. The Appearance of the Grail at Arthur’s Court Towards the beginning of the Queste, the Grail makes its first appearance before King Arthur and his knights as they sit at table. Emitting marvelous odors, it miraculously refreshes them with whatever meat (‘del tel viande’) each knight desires.83 Five manuscripts depict this episode, with differences Weston, Ritual to Romance. Eliot, Collected Poems 1909–1935, p. 91. 82 I thank Oliver Benjamin for discussion of this point. 83 Pauphilet, Queste, p. 15. 80 81

SEEING THE GRAIL

455

1. Rennes, BM 255, L’Estoire del saint Graal, f. 1, Christ appears to the hermit-author, handing him a book (photo: author)

2. New Haven Yale University, Beinecke Library 227, L’Estoire del saint Graal, f. 12, An angel hands a scroll to the hermit-author; Josephé celebrates the Grail mass before his followers (photo: Yale University)

456

3. Paris, BnF fr 749, L’Estoire del saint Graal, f. 1, The Trinity; the hermitauthor reading; Christ appears to the hermit-author; a game of bowls (photo: Bibliothèque nationale de France)

SEEING THE GRAIL

4. olim Amsterdam, BPH 1, L’Estoire del saint Graal, f. 6v, Joseph collects Christ’s blood at the Crucifixion (photo: Lancelot-Grail Project)

5. Paris, BnF fr 113, L’Estoire del saint Graal, f. 7, Joseph collects Christ’s blood at the Entombment (photo: Bibliothèque nationale de France)

457

458

6. Paris, BnF fr 344, L’Estoire del saint Graal, f. 26, Josephé shows the ark and its contents to King Evalach and Queen Sarracinte (photo: Bibliothèque nationale de France) 7. Paris, BnF fr 105, L’Estoire del saint Graal, f. 19, Josephé looks into the ark and sees Christ on the cross between two seraphim (photo: Bibliothèque nationale de France)

SEEING THE GRAIL

8. Bonn, LUB 526, L’Estoire del saint Graal, f. 16v, An angel cures Nascien’s blindness (photo: HMML)

9. London, BL Royal 14 E.III, L’Estoire del saint Graal, f. 15v, Bishop Josephé urges his father to leave the ark (photo: British Library)

459

460

10. Paris, BnF fr 344, L’Estoire del saint Graal, f. 65v, The Grail borne to Norgales (photo: Bibliothèque nationale de France)

11. Paris, BnF fr 344, L’Estoire del saint Graal, f. 78v, Josephé entrusts the Grail to Alain (photo: Bibliothèque nationale de France)

SEEING THE GRAIL

12. Florence, Bibl. Medicea-Laurenziana, Ash. 121 [48], La Queste del saint Graal, f. 6, The Grail appears at Arthur’s Round Table (photo: author)

13. Florence, Bibl. Medicea-Laurenziana, Ash. 121 [48], La Queste del saint Graal, f. 86, Josephé celebrates a Grail Mass before the Grail Knights (photo: author)

461

462

14. Paris, BnF fr 343, La Queste del saint Graal, f. 103v, Christ administers the sacrament to the Grail Knights (photo: Bibliothèque nationale de France) 15. Paris, BnF fr 343, La Queste del saint Graal, f. 103, Galaad heals the Roi Mehaignié with the Grail (photo: Bibliothèque nationale de France)

SEEING THE GRAIL

16. Paris, BnF fr 116, La Queste del saint Graal, f. 672, The Knights carry the Grail Table and the cripple is cured (photo: Bibliothèque nationale de France)

463

464

17. Paris, BnF fr 112, La Queste del saint Graal, f. 181v, Galaad dies before the Grail (photo: Bibliothèque nationale de France)

SEEING THE GRAIL

465

in pictorial emphasis. In the two versions commissioned by Jacques d’Armagnac, BNF fr. 112, vol. II, f. 5 (ca. 1470), and BNF fr. 116, f. 610v (ca. 1475), Arthur and the knights sit at the Round Table (shown as annular), contemplating the Grail, which appears in the open circular space at the center of the table. Shown as a ciborium, it is surrounded by rays of light and supported by two angels; this is also how the scene is shown in BNF fr. 120, f. 524.84 In the Rylands manuscript and in the closely related copy Royal 14 E. III (ca. 1315), Queen Guinevere is also present,85 sitting with King Arthur and his knights behind a rectangular table, in front of which is the Grail, painted gold and shaped like a ciborium with a closed lid. It is borne by a figure who stands on the near side of the table, a monk/cleric who is encircled by a wavy cloud motif — so that he is visible to us, but not to Arthur and his court, while the Grail that he holds out before the assembled company lies beyond the perimeter of the cloud. This paradox of visibility and invisibility is suggested by the text, ‘mes il n’i ot onques nul qui poïst veoir qui le portoit’ (‘but there was not a soul who could see who carried it’),86 and the cloud motif also refers back visually to the opening scenes in Estoire where, in these two manuscripts and in the related Add. 10292, Christ hands the book to the hermit. Curiously, however, this Grail-scene at the beginning of the Queste is omitted in Add. 10294.87 The other book that shows it is the one made in 1319 in Avignon, Florence, Laur. Ash. 121, f. 6 (Fig. 12).88 where the Grail, without a Grail-bearer, and depicted as a veiled chalice, is poised on a rod (the Holy Lance?) high above the heads of the diners, who (without Queen Guinevere) sit behind a rectangular table. The elevated placing of the Grail here may allude to another important liturgical practice concerning the Eucharist that can be traced to the early thirteenth century — that of

See Baumgartner, ‘La couronne et le cercle’, all illustrated; Matthews, Grail: Quest, pp. 81, 84 for reproductions; Blackman, Pictorial Synopsis‘, p. 38. 85 Not according to Pauphilet, Queste, p. 15, nor in the previously described episodes at the beginning of the Queste where Arthur and his knights are at table. 86 Pauphilet, ibid., p. 15. 87 Nor is this Grail-scene illustrated in Brussels, BR 9627–8, Paris, BNF fr 339, Oxford, Bodl. Digby 223, Paris, BNF fr 342 (a banquet scene is shown at the beginning of the Queste, but without the Grail), BNF fr 344, Bonn 526, BNF fr 110, BNF fr 123, Yale 229, Paris, BNF fr 1422–4, Oxford, Bodl. Rawl. Q.b.6, Douce 199, Paris, Bibl. de l‘Arsenal, 3482. 88 See Breillat, ‘La Quête du Saint-Graal en Italie’, pp. 296–300; Stones, ‘BN, fr 95 and Yale 229’, pp. 213–19; Walters, ‘Wonders and Illuminations’. 84

466

its reservation (as opposed to its open display in a monstrance), which was accomplished by suspending the ciborium containing the Host above the altar by means of pulleys.89 Lancelot Sleeps While a Wounded Knight Is Cured by the Grail Lancelot, because of his sin of adultery, is denied entry to the Grail chapel and sleeps outside by a cross, while a wounded knight is cured by the Grail.90 This episode is a second instance where a Grail-scene is depicted (by two different painters) in Royal 14 E.III and Rylands, but omitted altogether in the related manuscript, Add. 10294.91 The setting of this miraculous cure, in both text and pictures, is described and depicted in these two manuscripts in great detail. On the left is the chapel, from which the six-branch candlestick with lighted candles, the silver Grail Table, and the Grail itself (shown as a gold ciborium) have emerged into the open air. The wounded knight (wearing a knotted headscarf in Ryl) kisses the altar. In the background is his litter, drawn by two horses. Lancelot is shown asleep. In the illustration in BNF fr. 342 (whose text was copied a generation earlier, in 1274, by a female scribe),92 the Grail itself is omitted

89 The tabernacles on pulleys cited in Rubin, Corpus Christi, p. 47, are English and date between the late fourteenth and sixteenth centuries: so also the hanging tabernacle depicted on a triptych of the sixteenth century at Arras Cathedral, reproduced in Vloberg, L’Eucharistie, I, p. 73. Several considerably earlier (early thirteenth century) French examples are extant, made of Limoges enamel and taking the form of a dove; see Vloberg, ibid., I, p. 72 and Enamels of Limoges 1100–1350, no. 106, Paris, Musée du Louvre. Département des Objets d’art (OA 8104), catalogue entry by Elisabeth Taburet-Delahaye with full bibliography. There is no reflection of this form in Estoire or Queste iconography. See Braun, Das christliche Altargerät; Elbern, ‘Der eucharistiche Kelch’, for chalices: figs. 2–144: patens (non-bowl-shaped), figs. 145–164; pyxes (three of them stemless), figs. 165–178; ciboria, figs. 179–197, of which figs. 186, 187, 192, 193, 194 are fifteenth-century examples: monstrances, figs. 226–296; and Skubiszewski, ‘Die Bildprogramme der romanischen Kelche und Patenen’. See too the important article on inscriptions on liturgical vessels which tell much about how the theological writings were interpreted (Favreau, ‘Inscriptions médiévales’). 90 Pauphilet, Queste, p. 59. A few examples are illustrated in Stones, ‘Sacred and Profane’, pp. 100–12, figs. 4, 5, 6. 91 This scene is also omitted in Brussels, BR 9627–28, Paris, BNF fr 339, 110, 123, Oxford, Bodl. Digby 223, Rawl. Q.b.6, Douce 199, Yale 229, BNF fr 344, Ars. 3482 (listed in approximate chronological order). 92 The colophon says, ‘pries pour ce li (not ‘celui’) ki l’escrist’, cf. Pope, From Latin to Modern French, p. 325, para. 845.

SEEING THE GRAIL

467

altogether,93 whereas in Bonn 526 (written in 1286), it is given particular visual emphasis by being shown as a chalice with a cross inside, hovering in midair. Since the scribe of the Bonn manuscript, Arnulphus de Kayo, tells us in his colophon that he was in Amiens at the time of writing (‘qui est ambianis’, as though he were not normally there), it is interesting to note the presence of a closely similar chalice with cross in it, shown on a shield held by the virtue ‘Faith’ among the virtues carved on the west façade of the Cathedral of Amiens.94 In other ways, however, as I have shown elsewhere, the style of Bonn 526 has more to do with manuscripts associated with Thérouanne or Cambrai than it does with Amiens books.95 The chalice-with-cross motif can occasionally be found elsewhere, as in the early twelfth-century psalter made in England for the anchorite Christine of Markyate and commonly known as the St Albans Psalter,96 or, ca. 1250, on the cover of the Sainte-Chapelle Gospels, Paris, BNF lat. 17326, where the chalice and cross, placed on the right of Christ, symbolize Ecclesia, a parallel for the Tablets of the Law on His left.97 Other examples are rare.98 This episode is also prominent in the special version made in Lombardy, ca. 1380, BNF fr. 343, f. 18, where Lancelot sits off to the left, separated 93 The Grail is also omitted at this point in the Avignon copy of 1319, Florence, Laur. Ash. 121 (48). 94 Reproduced in Vloberg, L’Eucharistie, II, p. 232. 95 See Stones, The Illustrations of Lancelot, ch. 5, and ead., ‘Sacred and Profane’, pp. 108–10 and fig. 5. See also the image of the knights at the Grail Table with chalice and cross in the Bonn manuscript, discussed below. 96 Preserved at the church of St Godehard, Hildesheim. See St Albans Psalter, ed. Wormald, Pächt, and Dodwell, figs. 25 and 26, illustrating pages 39 and 40 of the manuscript, the Agony in the Garden and Christ waking the sleeping apostles (the manuscript is paginated, not foliated. The commentary on pp. 61 and 69 recognizes that the inclusion of the ‚cup of bitterness‘ in these scenes is unusual, but makes no mention of the still less common presence of the cross in the chalice. See also Geddes, The St Albans Psalter and https://www.abdn.ac.uk/ stalbanspsalter/ (consulted 15 March 2017). 97 For a reproduction, see Art and the Courts, no. 37, p. 55. I thank Adelaide Bennett for suggesting this and the examples mentioned in the following note. 98 Very few other examples show the combination cross and chalice: a painted panel of c. 1400, now in the Staedelmuseum, Frankfurt, shows an angel holding a cross descending toward a chalice containing a wafer, before which Christ kneels. A similar scene is included in the Belles Heures of Jean de Berry at the Metropolitan Museum, The Cloisters, New York. acc. no. 54.1.1, f. 123, illustrating Matins of the Hours of the Passion. A painting attributed by van Os to Taddeo di Bartolo at the Thorvaldsenmuseum, Copenhagen, shows an angel holding a cross-surmounted chalice, descending toward the kneeling Christ (‘Lippo Vanni’, p. 82, no reproduction given).

468

from the healed knight by the litter, placed in the center of the composition, while the knight, now cured, kneels on the right in the chapel where the Grail is depicted on the altar.99 The Grail is shown as a gold chalice with a round cup and a polygonal base, and has a paten, also in gold, covering the top. In Paris, Bibl. de l’Arsenal 3479 and BNF fr. 117, both made in Paris ca. 1405, this episode is given unprecedented prominence. Lifted out of its narrative context, it is the last of four key events in the life of Lancelot that are placed in a four-part composite miniature at the very beginning of the set of three volumes that transmit the five-part cycle complete. Florence, Laur. Ash. 121, the Avignon copy, also gives considerable emphasis to this subject, showing it (at the normal place in the text) in two separate scenes, ff. 20v, 21v. Jacques d’Armagnac’s manuscripts, BNF fr. 116, f. 621. and BNF fr. 112, vol. II. f. 15v, both include this curing scene in the body of the text, elaborating upon the treatment of the Grail, which in these books has been transformed into a highly ornamented and bejeweled ciborium.100 The Grail Liturgy101 The depiction of the Grail liturgy that occurs toward the end of the story is relatively rare,102 being limited to just a few examples, which I examine below. Other manuscripts depict one or more alternative scenes at the end, instead of the Grail liturgy or in addition to it: the death of King Mordrain,103 Galaad joining together the broken sword with which Josephé had been wounded,104 Galaad curing the Roi Mehaignié,105 the three knights Galaad, 99 Reproduced in Loomis, Arthurian Legends, fig. 333. See also the brief catalogue entry by Avril in Dix siècles, no. 84, pp. 98–99 and now Pastoureau and Gousset, Lancelot du Lac et la quête, p. 45 (f. 18). I thank Marie-Thérèse Gousset for her generous assistance with this manuscript. 100 Blackman, The Manuscripts of Jacques d’Armagnac, p. 23, and ead., ‘Pictorial Synopsis’, p. 39. 101 The Grail liturgy is discussed in more detail in Stones, ‘BN, fr 95 and Yale 229’, pp. 213–21. 102 Pauphilet, Queste, pp. 269–72. 103 ibid., p. 263: illustrated in BR 9628, BNF fr 339, 123, 342, Yale 229, Oxford. Bod. Rawl. Q.b.6, BNF fr 1424, Add. 10294, Royal 14 E.I1I, Rylands Fr. 1, Florence, Laur. Ash. 121 (48), BNF fr 111, 112, 116 (listed in chronological order). 104 ibid., p. 266: illustrated in BNF fr 342, Add. 10294 (but not in Royal 14 E. III or Rylands Fr. 1); BNF fr 112 (but not BNF fr 116). 105 ibid., p. 271, discussed below.

SEEING THE GRAIL

469

Perceval, and Bohort, kneeling before the Grail,106 the three knights together with a cripple, carrying the Grail Table at Sarraz, or the crowning and death of Galaad.107 Some of these alternative scenes also include a depiction of the Grail, and I return to them below. But most manuscripts of Queste, including Jean de Berry’s and Jacques d’Armagnac’s manuscripts, leave the climactic Grail liturgy to the imagination — a deliberate ploy, also used in the illustration of the manuscripts of Chrétien de Troyes’s Perceval and its Continuations, the other great story about the Grail.108 Identified in the context of the two earlier Grail episodes as ‘li Sainz Graal’ or ‘li Sainz Vessel’ (or a variant thereupon), it is only in the context of the liturgy that the Grail, in the Queste, is defined more precisely: ‘Ce est, fet il [Christ], l’escuele ou Jhesucriz menja l’aignel le jor de Pasqes o ses deciples. Ce est l’escuele qui a servi a gré toz çax que j’ai trovez en mon servise; ce est l’escuele que onques hons mescreanz ne vit a qui ele ne grevast molt. Et por ce que ele a si servi a gré toutes genz doit ele estre apelee le Saint Graal’ (‘This is, he [Christ] said, the bowl from which Jesus Christ ate the Pascal lamb with his disciples. This is the bowl which has satisfied all those in my service; it is the bowl which no unworthy man set eyes upon without it doing him grievous harm. And because it has so served the desires of all people it should be called the Holy Grail’).109 From this description one would expect that depictions of the Grail in Queste illustration would show it looking just like the ‘escuele’ that is so common in Estoire illustration. But we have seen that the Grail in Estoire illustration also takes other forms, derived from Christian liturgy, and those liturgical forms have so far been the ones preferred for Grail depiction in Queste scenes as well. None of the depictions of the Grail in the context of the Queste’s Grail liturgy follow these words. Nor is there any parallel in Queste illustration for the ‘escuele’ mentioned here in the text and illustrated in the Estoire manuscripts. A sacred liturgy is enacted several times in the text of Queste: The first could be considered a proto-Grail liturgy, since the Holy Vessel is not explicitly part of the episode. The three Grail knights, Galaad, Perceval, Bohort, and Perceval’s sister, see, as a priest celebrates mass of the Holy ibid., p. 277, discussed below. ibid., pp. 277–79. The crowning and death are illustrated as two separate scenes in Florence, Laur. Ash. 121(48); just the death, as a single scene, in BNF fr 112 (but not BNF fr. 116, see Blackman, ‘Pictorial Synopsis’, p. 42). 108 For Chretien’s Perceval see Baumgartner, ‘Les scènes du Graal’. 109 Pauphilet, Queste, p. 270. 106 107

470

Ghost, the miraculous White Stag transformed into Christ in Majesty, and the four accompanying lions become the symbols of the four evangelists, the man, eagle, lion, and ox.110 Two illustrations depict this in the Avignon manuscript, one showing the four companions kneeling behind the priest and acolyte at an altar on which is a partially veiled chalice; the second is a historiated initial Q enclosing Christ in Majesty and the four evangelist symbols. While the stag and lions are depicted at this spot in Add. 10294 and the Udine manuscript (f. 91),111 and several manuscripts show two knights in a chapel, 112 there is, to my knowledge, only one parallel for the illustration of this episode in the Avignon copy — in the special version of the Queste made in Lombardy ca. 1380, BNF fr. 343. There, the two scenes are combined, to show the three knights — without Perceval’s sister — outside a chapel, addressed by a priest; inside the chapel, Christ is enthroned on the altar, sitting on a beige wooden throne decorated with ebony and ivory inlay, surrounded by the four evangelist symbols. At Christ’s feet are two cushions and a book. This is a second instance of a rare subject illustrated in the Avignon and Lombard manuscripts, and only there, so far as I know.113 Next, the Grail liturgy is viewed through the open door of the Grail chamber by Lancelot, who has been forbidden to enter the room: it is almost an anti-liturgy, comparable with and referring back to Lancelot’s presence at, but exclusion from, the Grail’s healing miracle related and illustrated earlier.114 Only one manuscript includes a depiction of what Lancelot sees as he witnesses the Grail liturgy from afar: it is the Avignon copy. There is no 110 ibid., pp. 234–35. Standard elements of Christian iconography, the association of the evangelists with these creatures, which emerged in the Early Christian period as man=Matthew, lion=Mark. ox=Luke, eagle=John, is based on the Visions of Ezekiel (Ezek. l, 4–28), on the vision of St John in the Apocalypse (Apoc. 4. 7–9), and on commentaries on those passages. 111 Bergamini and Menis, Miniatura in Friuli, no. 22, p. 79, and the facsimile edition. La grant Queste del Saint Graal. 112 Oxford. Bodl. Digby 223, Paris, BNF fr 110, BL Royal 14 E. III, BNF fr 116. f. 643v, Florence, Laur. Ash. 121(48). The White Stag also features in the major Grail liturgy in the Florence Tavola Ritonda, discussed below. For more on the White Stag see now Stones, ‘Le merveilleux dans le Lancelot-Graal: l’exemple du cerf accompagné de quatre lions,’ reprinted in these essays. 113 The other such scene occurs just before this, and shows Bohort seeing a Pelican in her Piety, a symbol of Christ drawn from the Bestiary’ tradition and a common iconographie motif (Pauphilet, Queste, p. 168). The Florence example is reproduced in Stones, ‘BN, fr 95 and Yale 229,’ as fig. 8. 5. 114 Pauphilet, Queste, p. 255.

SEEING THE GRAIL

471

attempt to render the doorway through which Lancelot looks, although the entire scene takes place under an arched frame. Lancelot is shown standing to the left, next to the two figures whom the text says handed the priest a third person whom the priest elevated before the assembled company. The two figures — positioned, according to the text, above the priest, ‘desus les mains au preudome en haut’ — are here shown standing on the ground, on Lancelot’s side of the altar, with the third person raised above the chaliceGrail on the altar, while the priest and two onlookers are on the right of the altar. This is an allusion to the Trinity, linking thereby with the opening of the Estoire — although the Avignon manuscript transmits only the Queste and Mort Artu branches and no companion volumes have survived. By including the transubstantiation, this miniature also anticipates the much longer passage describing the liturgy that the three chosen knights will witness later. The miniature in the Avignon manuscript has been severely defaced, as have several other illustrations, or parts of illustrations, in the manuscript. Perhaps in this instance someone felt it was inappropriate to show Lancelot seeing the transubstantiated Host. The earliest surviving Queste manuscripts do not depict any liturgy at all, so the earliest illustration we have of the major climactic liturgy in the Queste is the one in Yale 229, made ca. 1295 in the diocese of Thérouanne, perhaps for one of the sons of Guy de Dampierre, whose iconography I have analyzed elsewhere.115 Several of its notable features do not depend on the text: the presence of five knights (not just Galaad, Perceval, and Bohort), the celebrant shown as Bishop Josephé, wearing a mitre (not Christ), and the extraordinary shrine-like, or tabenacle-like, container, painted in gold, which conceals the Grail from the viewer.116 According to the text, Josephé disappears, and Christ emerges from the Grail and Himself administers the sacrament to the knights:117 the artist (or planner) has misinterpreted the text. Josephé and Galaad are also the major participants, though shown without an audience of other followers, in the illustration showing the liturgy in the late-thirteenth-century Italian copy, Udine, Bibl. Archivescovile 177, illustrated with the tinted line drawings that characterize a large number of French manuscripts copied probably in Genoa for Italian patrons in the late See Stones, ‘BN fr 95 and Yale 229’, pp. 214–15, 219–20. Numerous parallels with shrines and tabernacles of the late thirteenth century can be drawn: see ibid., pp. 225–27. 117 Pauphilet, Queste, pp. 269–70. 115 116

472

thirteenth and early fourteenth centuries.118 There is only one Grail scene, showing Josephé, mitred and holding a cross staff, borne on a throne by four angels before a single knight wearing civilian dress who stands beside a table on which is placed the Grail, showrn as an amazing pear-shaped vessel with a wide rim and a shallow base and two huge rounded handles. This depiction of the Grail is without precedent: it reminds one of the vessels common in Early Christian mosaics from which flowering plants and fruit grow as Paradise motifs; two-handled vessels of this sort also figure among the vessels at the foot of the cross in ninth- and tenth-century Crucifixion scenes.119 Possibly associations like these were intended. In the copy made in Avignon in 1319, Florence, Laur. Ash. 121 (48), the Grail liturgy is treated as a single-column miniature (Fig. 13) showing Josephé raising his hands in prayer as he celebrates Mass at an altar with a chalice-shaped vessel on it; on the left, behind Josephé, stand angels with two candles behind them: one of the angels holds the lance, resting its foot on the floor. On the other side of the altar kneel three knights. It is in Pierart dou Thielt’s famous depiction in Paris, Bibl. de l’Arsenal 5218, f. 88, made in Tournai in 1351,120 that the emphasis on transubstantiation corresponds most precisely to the textual description of the Host transformed into a child — for which Eucharistic miracles like that of Edward the Confessor and those painted on the walls of Orvieto Cathedral are parallels.121 Placed earlier in the text, at the spot where the Yale manuscript and others show the death of King Mordrain, it combines

118 For the attribution to Genoa, see La grant Queste, pp. 31–47, and Avril, Gousset, and Rabel, Manuscrits enluminés d’origine italienne. II, Xllle siècle, pp. 25, 32–52. The Udine manuscript would seem to be the only surviving illustrated copy of Queste from this workshop, whereas there are several copies of the Lancelot proper and several copies of the prose Tristan: see Branca, ‘Tradizione italiana dei testi arturiani’. An Italian (or Cypriot or Levantine) artist participated in the illustration of Tours BM 951 (Estoire, Joseph, Merlin), but neither he nor his French collaborator included the Grail. See Folda, Crusader Manuscript Illumination, pp. 122–23; for the text see O’Gorman. The Italian Joseph and Merlin in Florence. Bibl. Riccardiana 2759, is unillustrated, and the Italian Queste, Oxford. Bodl. Rawl. D. 874, is minimally illustrated and does not include a depiction of the Grail. 119 See note 27 above. 120 For the artistic context see Avril in Fastes du Gothique, no. 301, pp. 348–49, and for the iconography, Walters, ‘Wonders’. This was also reproduced in Loomis, Arthurian Legends, fig. 341, and in Stones, ‘Sacred and Profane,’ fig. 7; in color in Matthews, Grail: Quest, pp. 52–53 and on the cover of Weston, Ritual. 121 See notes 50–51 above.

SEEING THE GRAIL

473

the Grail procession with the transubstantiation.122 The central pivot of the miniature is the figure of Christ in the chalice-shaped vessel, which floats in midair, between the two kneeling, candle-bearing angels, in front of the horizontal bleeding lance; both lance and chalice stand out against the white cloth on the table, behind which are grouped the eleven knights. Josephé sits, holding his crozier, on a throne at the left short end of the table, and the space on the right of the composition is occupied by two standing angels, one of whom carries an object draped with a red cloth, while the other holds a lance. The composition is ingeniously arranged so that the narrative climax, the transubstantiation, is enclosed within the rest of the pictorial depiction of the procession, as though the preceding action were taking place around it.123 In one long image, Pierart dou Thielt has apparently rearranged the items carried by the angels so that the Grail, covered by the ‘touaille de vermeil samit’ (‘cloth of red precious fabric’), is carried by one of the angels on the right, while the lance is borne separately by the other, and does not drip into the vessel. In the middle of the same image is another depiction of the Grail, shown as a chalice with the figure of Christ inside it, and the lance is again placed so that the Grail does not receive the dripping blood — this time it is horizontal, as if floating, a position perhaps governed by the desire to make it stand out against the long horizontal white of the tablecloth. The floating effect (not in the text) may at the same time underline its supernatural quality. Both Italian manuscripts, the special French version of the Queste made in Lombardy, BNF fr. 343, and the Italian compilation based on the Queste, and known as the Tavola Ritonda, Florence, BN, Pal. lat. 556,124 include a depiction of the Grail liturgy: as a single, unfinished, scene in BNF fr. 343,125 Pauphilet, Queste, pp. 269–70. ibid., p. 269. 124 Gardner, Arthurian Legend, p. 153, reproductions facing pp. 2, 112, 122, 174, 180. 190, 208. 268. 272, 286; Loomis, Arthurian Legends, p. 121, figs. 337–39; Mostra dei codici romanzi, pp. 119–20. For the text, see La Tavola Ritonda; Branca, I Romanzi italiani di Tristano e la ‘Tavola Ritonda’, esp. pp. 204–07; see also Kleinhenz, ‘Tristan in Italy’, pp. 145–58; Heijkant, La tradizione de ‘Tristan’. 125 Many of the illustrations in BNF fr 343 remain as drawings, including the opening portrait of the patron on f. 1, although the miniature on the same page was completed in color, and others sporadically throughout the manuscript are also unfinished. There appears to have been more than one painter at work, and perhaps more than one draftsman, a question I cannot pursue here. 122 123

474

and as a series of no fewer than nine tinted drawings in Pal. lat. 556. The scene in BNF fr. 343, f. 103v, shows Christ, standing on a table, holding a closed cylindrical pyx (vessel for consecrated Hosts) with a round finial on its lid in His left hand, while handing a Host inscribed with a cross to a large group of kneeling knights (Fig. 14). The drawing was left uncolored and the Head of Christ, which is surrounded by rays of light, remains featureless. Produced in 1446 by two craftsmen from Cremona, the scribe or editor (or both ?) Zuliano di Anzoli,126 and, in all likelihood, the artist Bonifacio Bembo,127 the Mantua manuscript now in Florence depicts nine different stages in the enactment of the liturgy, a phenomenon remarkable as an unprecedented number of Grail images in general, and as a pictorial concentration of images extraordinary within this manuscript.128 The Grail liturgy begins on the lower half of f. 147v with the appearance of Josephé, in bishop’s robes, holding a crozier and supported by three angels, before the knights. The Grail procession enters on f. 148 (upper miniature), passing alongside Josephé: the first angel carries two candles, the second a cloth with a band of decoration and a fringe at the end: the third holds the lance, held at an angle over the angel’s shoulder, so that its tip is poised directly above the Holy Grail, which the fourth angel raises high in veiled hands in order to catch in it the blood of the lance. The Grail is extraordinary in shape: like a huge round ceramic pot with a heavy torus molding at the rim and base, each decorated with a band of small circles; the rounded body of the vessel is patterned with a ring of cinqfoil flowers, matching in shape and size the ones on the walls of the room ! 129 The lower miniature on the same page shows the angels kneeling beside and behind Josephé, two of them now holding the candles, while the Grail rests on the table with the lance standing 126 His name and the date are in the colophon: he is known as a Cremonan from his Filocolo, see De Robertis, ‘Centesimo dei manoscritti,’ p. 276, cited by Delcorno Branca, ‘Tradizione italiana,’ p. 243, n. 78. 127 For the attribution to Bembo, see Rasmo, ‘Il Codice palatino 556’; Salmi, ‘Nota su Bonifacio Bembo’; Mulazzani, I tarocchi; Woods-Marsden, The Gonzaga of Mantua, pp. 28, 184 n. 81; see also Stephani, ‘Per una storia della miniatura italiana,’ II, p. 859. A useful summary of the problems is Branca, ‘Rassegna’. I thank Daniela Delcorno Branca, Marie-José Heijkant, Christopher Kleinhenz, and Joanna Woods-Marsden for their generous assistance with the manuscripts of the Italian tradition. 128 Breillat, ‘Le manuscrit Florence Palatin 556 et la liturgie du Graal. Other aspects of the iconography are equally unusual, see Stones, ‘Aspects of Arthur’s Death’, p. 71. 129 The closest parallel is the two-handled urn-shaped Grail showm in the single Grail illustration in the Udine manuscript (see notes 111 and 118 above).

SEEING THE GRAIL

475

vertically inside it; the knights watch the moment of transubstantiation as Josephé holds up high the small blessing figure of Christ.130 Then Josephé, having replaced the transubstantiated bread in the vessel (in which the lance still stands), bends over to give Galaad the Kiss of Peace (f. 148v, top), which Galaad should then transmit to the other knights; the four angels stand behind Josephé, two holding candles, the other two holding their arms crossed over their chests like the knights. As in the Queste, Josephé then disappears (not represented in the illustration), and Christ appears and administers the sacrament to the knights. Here (f. 148v, bottom) Christ, lifesize, revealing the wounds of the Passion, actually stands on the table next to the Grail (no lance in the Grail this time), and blesses the knights. After that (f. 149, top) Christ leans over, his hand on the rim of the vessel, about to lift it up for Galaad to drink from it; Galaad bends over the vessel, still with crossed arms. Then (f. 149, bottom) Christ flies up toward the window. Outside, Lancelot, not permitted to witness these events, sits on the ground by the door. The scene shifts altogether to the outside of the Grail palace on f. 149v (bottom) as Lancelot, still seated on the ground facing the closed door, looks up at Christ, now transformed into a seated stag, who has emerged from the window above the door and is borne away on a cushion by three angels.131 In the next scene (f. 150) Lancelot sees the four angels carry away the two candles and the lance, still dripping blood into the Grail, which is borne by the last angel who is still emerging through the window above the door of the castle. Although Josephé consecrates the bread in the Tavola Ritonda, the substance that is meted out to the knights is not flesh, as in the Queste,132 but blood. As Breillat pointed out, much more attention is given to blood as an underlying motif throughout the Tavola Ritonda. Not only does the description of the Grail liturgy make clear that the blood of the lance actually falls into the vessel, as is also shown in the illustrations, but the Christ is the Christ of the Passion, revealing his wounds in both text and picture much more explicitly than in the Queste. 133 130 This corresponds to Polidori, La Tavola Ritonda, pp. 473–74. For the citations, see Stones, ‘BN fr 95 and Yale 229’, pp. 216–19. 131 Reproduced in Matthews, Grail: Quest, p. 89. 132 ‘la piece en semblance de pain’, Pauphilet, Queste, p. 270. 133 Cf. Pauphilet, Queste, p. 270: ‘un hom ausi come tout nu. et avoit les mains saignanz et les piez et le cors’; there is no mention of the Wounds as such.

476

Bloodletting and the Element of Blood In the Tavola Ritonda as in the Queste, Perceval’s sister dies from donating her blood to cure a maiden with leprosy. But in the Tavola Ritonda version of this episode, the leper is cured by drinking the blood, not by having it splattered on her breast as in Queste. 134 Only three manuscripts illustrate the bloodletting and curing: the Avignon manuscript of 1319, Florence, Laur. Ash. 121, the Bonn manuscript of 1286,135 and the special version made in Lombardy ca. 1380, BNF fr. 343. In the Bonn manuscript’s version, the tourniquet, wound, blood, bowl, and supporting staff lend an air of medical reality to the depiction, while striking a discordant note. For the scrupulously supervised bloodletting is the sacrifice that will lead to death, whereas the simultaneous application of the blood to the breast of the leper — a symbolic act — is what will lead to a miraculous cure. The Avignon version shows the surgeon making the incision as Perceval’s sister holds the bowl to catch the blood, watched by three standing figures; the leper (her face rubbed out — because it was disfigured?) lies on a patterned bed in front of the scene. BNF fr. 343, f. 59v, shows the assembled company watching Perceval’s sister swooning in a chair, supported by the hooded physician, her right arm resting on a cushion, bared to show the incision, a silver ‘escuele’shaped bowl placed at the edge of the red-and-gold cushion to catch the blood. The cured damsel stands to the right, accompanied by a knight.136 Although these bloodletting scenes occur in the text without the accompaniment of the Grail, the illustrations in Bonn 526 and the Avignon and Lombard copies may allude to its healing powers through the shape of the bowl in which the donated blood is collected — the same as the Grail so

Pauphilet, Queste, pp. 240–41; Polidori, Tavola Ritonda, p. 472. For connections with medical illustrations of this subject, see Stones, ‘Indications écrites’, pp. 322–23, figs. 1–3. One of Jacques d’Armagnac’s manuscripts, the special version, BNF fr 112, has on f. 172 a scene showing the three knights Galaad, Bohort, and Perceval fighting the knights who want a dishful of blood from Perceval’s sister. This scene is not otherwise represented, and BNF fr 112 does not include the bloodletting scene itself. See Blackman, The Manuscripts, p. 238 and ead.,‘Pictorial Synopsis’, p. 41. 136 Reproduced in Loomis, Arthurian Legends, fig. 329, and, in color, in Matthews, Grail: Quest, p. 49. The miniature is rubbed in the area around Perceval’s sister’s arm, and the bowl is unclear in black-and-white reproductions, although it is perfectly distinct in the original and can be seen in the color reproduction. 134 135

SEEING THE GRAIL

477

common in the Estoire tradition, occasionally in the Last Supper,137 and the blood of Old Testament sacrifice or of healing water.138 In Queste illustration the Grail itself is always depicted as a chalice or a ciborium, the former shape evoking the Eucharistic blood, the other the Host. I would suggest that the occasional depiction of a chalice to represent the Grail in Estoire — as in its one appearance in the illustrations of BNF fr. 770 — seems merely to reflect the Eucharistic connotations of the Grail in general, while the consistent use of a chalice to represent the Grail in the Avignon manuscript — where the Grail as such is shown more times than in any other manuscript — does suggest a particular desire on the part of those who devised the illustrative program to place special emphasis on the element of blood. This emphasis on the element of blood links this manuscript thematically with the text and the images of the Tavola Ritonda. The cult of the Holy Blood at Mantua, one of the earliest documented in Western Europe,139 was a governing factor in the Gonzaga’s choice of an Arthurian subject about Bohort, the only one of the Grail heroes to produce any progeny, for the painting by Pisanello in their castle at Mantua.140 The Florence Tavola Ritonda, whose illustrations have been attributed to Bembo, is thought to belong to the same artistic current as the Camera paintings,

See note 25 above. See note 28 above. 139 Discovered in 804 in a vessel contained in a chest of leather inscribed Jesu Christi Sanguis, noted in Annales Regni Francorum, p. 119; also mentioned in Sumption, Pilgrimage, pp. 46–48. Pope Sixtus IV granted indulgences in three briefs of 1475 to Lodovico Gonzaga and Barbara of Brandenburg and their descendants: see Woods-Marsden, pp. 54, 63–64, 212 n. 126, citing Donesmondi, Dell’istoria ecclesiastica di Mantova, pp. 4–12 (which was not available to me): Peebles, The Legend of Longinus; Horster, ‘Mantuae sanguis preciosus’. The relic is in the crypt under the dome at the church of San Andrea in Mantua, but the crypt is not accessible to the public. In the main church, an enormous bronze polygon, containing explanatory inscriptions and surrounded by a marble railing, marks the spot below which the relic is preserved. Part of the Mantua relic came in 1055 into the hands of Emperor Henry III, then at his death the following year to Count Baldwin of Flanders. It was taken by Baldwin’s daughter Judith to the Welf court at her marriage in 1071 to Duke Welf IV, and came at her death to the abbey of Weingarten. The story is recounted on an altar-piece of 1489 now at the Württembergisches Landesmuseum, Stuttgart. See Kruse, Die Weingartner Heilig-Blut-Tafel von 1489. 140 What is shown is the victory of Bohort over King Brangoire in a tournament, followed by his deception, thanks to a potion, by the king’s daughter, which resulted in the conception of Helain le Blanc, future emperor of Constantinople. See Woods-Marsden, The Gonzaga, pp. 13–20. 137 138

478

and Benedetti has identified the Queste manuscript of Italian manufacture, Udine Bibl. Arcivescovile 177, discussed above, in the Gonzaga inventory of 1407.141 Although Yale 229 has no known medieval provenance, apart from what can be deduced stylistically and from the heraldry in its decoration,142 it has been suggested that it or another copy from the same workshop might well have also been in the Visconti-Sforza library (where many Arthurian manuscripts are listed in terms too general for particular copies to be identified) and could have been available to Bembo during one of his visits to Milan to work for the dukes of Milan. There he might also have seen BNF fr. 343, made ca. 1380 for someone in the circle of the Visconti dukes of Milan and identifiable in their fifteenth-century inventories.143 Some of the French manuscripts with particularly interesting depictions of the Grail were also made in regions that were also important for the cult of the Holy Blood. I have argued that BNF fr. 95 and Yale 229 were most likely made in the region of the diocese of Thérouanne.144 Boulogne-surMer, in the diocese of Thérouanne, and Bruges, in the neighboring diocese of Tournai,145 also boasted Holy Blood relics; and Ars. 5218 was made in Tournai. The Bruges relic was allegedly acquired by Thierry of Alsace from the Patriarch of Jerusalem in 1148, in appreciation of his valor in the Second Crusade, and housed in a crystal cylinder with late-thirteenth-century metal terminations;146 the Boulogne relic is a piece of cloth supposedly donated to his native Boulogne-sur-Mer by Godefroi de Bouillon in 1101 and now preserved under rock crystal in the splendid translucent enamel disk made to house it in the last decade of the thirteenth century, its workmanship attributed to the Parisian goldsmith Guillaume Julien, and its patron

Benedetti, ‘ “Qua fa’ un santo e un cavaliere...” ’, cited by Branca, ‘Rassegna (1985– 1992)’, p. 470. 142 See Stones, ‘BN, fr 95 and Yale 229’, pp. 232–33. 143 See note 124 above. In addition to the unfinished state of the portrait initial at the beginning, perhaps intended to show the patron, there is no Visconti-Sforza heraldry. It did however enter the Visconti-Sforza library, where Pellegrin identified it with inventory no. A 908 (La Bibliothèque des Visconti et des Sforza, p. 274), and Delcorno Branca suggests that it can also be identified with an entry in the Sforza inventory of 1488. See Branca, ‘Tradizione italiana’, p. 241 n. 69, citing Cavagna, ‘Il libro desquadernato’, item no. 617. 144 Stones, ‘BN, fr 95 and Yale 229’, pp. 229–30. 145 The most reliable maps are the ones in Moreau, Histoire de l’Église en Belgique; see also Cottineau, Répertoire topo-bibliographique and Supplément, ed. Poras. 146 Toussaert, Sentiment religieux, pp. 259–67. 141

SEEING THE GRAIL

479

Philippe le Bel, king of France.147 BNF fr. 95 and Yale 229 were made at about the same time as this reliquary. In the Queste, the Grail cult is essentially based on the bread of the Eucharist rather than on the wine-blood. In the Estoire, both the elements are present, though the emphasis on transubstantiation gives greater weight to the bread than to the wine.148 In the Tavola Ritonda the more important element is the blood, an emphasis that is anticipated visually in the illustrations of the Queste made in Avignon and in Italy (there are no surviving copies of Estoire from Italy), where the Grail is consistently shown as an open chalice — a vessel to be drunk from — rather than a closed ciborium, a vessel for reservation and concealment. The Healing of the Wounded King Even in the Queste, the conclusion to the liturgy of the Grail emphasizes the element of blood, since Christ’s final instructions to Galaad are that he cure the Roi Mehaignié by anointing his wound with the blood of the lance.149 A parallel to the wounding of King Alphasem in Estoire, the healing of the Roi Mehaignié by Galaad is depicted considerably less frequently than one might expect. Only a single manuscript includes a scene of it, the special version in French made in Lombardy, BNF 343, f. 103 (Fig. 15), and the illustration depicts a shift of emphasis in the healing element. Rather than use the Holy Lance dipped in blood to heal the king, as the Queste text says, Galaad effects the cure by placing the Grail itself, shown as a chalice which he holds in the middle of the stem, at the groin of the seated king, suggesting that the vehicle of healing is the vessel itself, regardless of which element of the sacrament it contains.150 This healing scene is anticipated in BNF fr. 343 147 It is housed today at the church of Saint-François-de-Sales in Boulogne-sur-Mer. See Gauthier, Les émaux, no. 158; Trésors des Églises de France, no. 42, p. 111. The reliquary was a gift from Philippe le Bel to Notre-Dame de Boulogne in commemoration of the marriage in 1308 of his daughter Isabelle to Edward II of England. Vloberg. L’Eucharistie. II. 149, refers to Malou, Du culte du Saint Sang de Jésus Crist. Vloberg, L’Eucharistie, II, pp. 149–51, also discusses the Fécamp relic of the Holy Blood, allegedly contained in a vial and hidden in a hollow fig tree by Isaac, nephew of Nicodemus, then carried by water to Fécamp (Fici Campus). 148 Sommer, Vulgate Version, I, p. 40. 149 Pauphilet, Queste, pp. 271–72. 150 The scene is reproduced in Loomis, Arthurian Legends, fig. 334, and Matthews, Grail Quest, p. 59. Perhaps BNF fr 343 is the copy that inspired the thinking of Jessie L. Weston;

480

by an earlier episode, showing an emaciated King Mordrain, sitting on an invalid’s bed on wheels, receiving the Host from a priest who holds a chalice covered by a folded corporal (cloth). The Grail at Sarraz The final healing episode in the Queste also links the cure with the Holy Vessel. After a miraculous sea voyage in the boat on which they had earlier found the sword with the mysterious hangings (‘estranges renges’), where now they find the silver Grail table and the Grail itself, the three knights Galaad, Perceval, and Bohort arrive at the city of Sarraz and are instructed to carry ashore the silver altar and the Grail.151 Finding the table heavy, Galaad asks a cripple to assist them, pronouncing him cured of his affliction. Illustrated in all three early-fourteenth-century Flemish manuscripts. Add. 10294, Royal 14 E. III. and Rylands Fr. 1, this scene is also in Ars. 3482, f. 537, and in BNF fr. 112, f. 181, but none of these illustrations shows the Grail on the table. In the Avignon manuscript, Florence Laur. Ash. 121, f. 88v, the Grail is shown, as a chalice, on the table which the three knights carry, while the standing cripple occupies a separate compartment to the right.152 In Jacques d’Armagnac’s copy, BNF fr. 116, f. 672, the knights carry the table on which is the Grail, shown as a multilobed circular ciborium with a knop and tall spirelike finial, while the cripple is shown rising, his hands in prayer (Fig. 16).153

one wonders whether she saw the manuscript itself or perhaps Loomis’s photographs of the manuscript. There is one other possible example: although unlikely given the absence of this episode elsewhere in the French illustrative tradition, it may be that the scene on f. 525 in Ars. 3482, showing a knight and a servant before a king who stands in front of a throne, depicts this episode. What is going on is not clear, and another explanation is that this might depict the death of King Mordrain. It is also worth noting that BNF fr 112, f. 281, has a blank space for an illustration of Perceval and the Roi Mehaignié. 151 Pauphilet, Queste, pp. 273–75. 152 The miniature is partially rubbed, and the cripple erased, as was the case with Lancelot observing the Grail liturgy from the doorway, discussed above. Here, however, there is no obvious reason why the cripple’s presence should have been objected to. 153 See Blackman, The Manuscripts, p. 239, and ead., ‘Pictorial Synopsis’, p. 42.

SEEING THE GRAIL

481

The Final Events: The Knights Before the Grail, The Crowning of Galaad, The Last Grail Liturgy, The Death of Galaad and The Disappearance of the Grail In the final appearance of the Grail in the Tavola Ritonda (f. 151), the nails of the Passion and the Crown of Thorns are also included; they all accompany the dead body of Perceval’s sister on the boat (bottom scene) and are transferred by the knights to the altar in the Grail chapel, before which Perceval’s sister is buried (top scene) and where Galaad, having ruled as king for a year, will also be buried (a scene not illustrated in Pal. 556).154 In the story as the Queste relates it, Galaad, Perceval, and Bohort, having buried Perceval’s sister (an episode not illustrated in any of the surviving manuscripts), are imprisoned by King Escorant. In prison they are succored by the Grail (shown in the Avignon manuscript). At the death of King Escorant — an episode also illustrated in the Avignon copy, now badly rubbed, showing the king on his deathbed with figures behind — the three are released from prison and Galaad is crowned as successor to Escorant,155 again illustrated in the Avignon copy (f. 89v). Galaad builds a tabernacle (‘arche’) of gold and precious jewels to cover the Holy Vessel, a textual link with the ark in which the Grail is seen by Josephé in the Estoire, and a cross-media link with the shrine-like Grail tabernacle shown in the Yale manuscript’s depiction of the liturgy. But neither the Yale copy nor any other includes an image of this ark at this point in the text. After this comes the final Grail liturgy, celebrated for the three knights by Josephé, followed by the death of Galaad and the return of the Grail to heaven, transported by a miraculous hand. There is sometimes ambiguity in the illustrations as to which of these episodes is intended in the picture, as the placing of the final miniatures in the text is often far removed from the event shown in the image. In BNF fr. 342, f. 145v, there is a rather generic final scene of knights before a table, with no Grail present;156 while in the final scene in Bonn 526, f. 451v, Galaad, Perceval, and Bohort stand behind a laden table, Galaad (presumably) holding a chalice with a cross in it, like the one before which,

Polidori, Tavola Ritonda, pp. 475–76. This corresponds to Pauphilet, Queste, p. 276. Pauphilet, Queste, p. 277. 156 The miniature has been placed at Pauphilet, ibid., p. 268. 154 155

482

in this manuscript, Lancelot is shown sleeping earlier in the story.157 The Avignon manuscript unequivocally shows the three knights in prison receiving communion from Christ, who holds the Grail, shown as a large chalice-like vessel,158 while Josephé’s final appearance is not shown. In Ars. 3482, f. 538v, the final image shows Galaad, alone, receiving the sacrament from Josephé, robed as a bishop. Jacques d’Armagnac’s special version, BNF fr. 112, f. 179v, shows the knights kneeling, their backs to the viewer, before an altar on which rests a spectacular jeweled polygonal ciborium-like Grail with a knop in the center of the stem and surmounted by a polygonal turret with a jeweled cross on top.159 As Galaad dies before the Grail (f. 181v, Fig. 17), supported by Perceval and Bohort, the vessel in BNF fr. 112 has become a round ciborium on a splayed stem, with scalloped motifs on bowl and cover, and beading on the rim of the foot, the base of the lid, and the top of the lid, below a finial of four balls supporting another ball on top of which is a cross. As described in the text,160 a hand appears and miraculously carries off the Grail and the lance. The Avignon manuscript has a simpler version: Galaad, crowned, lies beside the altar on which is the chalice-like Grail from which the lance projects vertically. The miniature is badly rubbed, like many in this manuscript, and one cannot be sure whether Galaad has a halo or whether his head rests on a cushion.161

conclusion As befits its mysterious nature, the depiction of the Grail does not reflect a single direction or trend, nor is it depicted in a single way in the illustrations of Estoire and Queste. Manuscripts produced by the same team of craftsmen show inconsistencies as to which Grail scenes are included and how the Grail is treated, suggesting that the rules were flexible at best, and that there was 157 Reproduced in Ringbom, Graltempel und Paradies, fig. 132, p. 473 (where the manuscript is wrongly identified as a ‘Parzival-Handschrift’) and in von dem Borne, Der Gral, fig. 16. This scene is placed at Pauphilet, Queste, p. 262, at a point where most illustrated manuscripts show King Mordrain dying in the arms of Galaad. 158 Placed at Pauphilet, ibid., p. 276. 159 Not in Jacques d’Armagnac’s other manuscripts, BNF fr 116 and BNF fr 120: see Blackman, The Manuscripts, p. 239, and ead., ‘Pictorial Synopsis’, p. 42. 160 Pauphilet, Queste, p. 279. 161 ibid., p. 278.

SEEING THE GRAIL

483

no single line of development.162 The complexities of the iconography of each branch of the Lancelot-Graal suggest that painters — or, more likely, their patrons and/or the directors of operations — took upon themselves either to exclude or include the subject on an ad hoc basis, and, if the latter decision was made, to be guided by what seems to have been quite a complex network of factors, the rationale for which is not altogether clear. The adaptation of models from the context of Christian art was an obvious route to follow, since the events surrounding the Crucifixion, Deposition, and Entombment are directly narrated in Estoire, and since Christ Himself plays a mystical part in the Grail liturgies in the Queste. In both texts the Grail liturgy explicitly includes consecration, elevation, transubstantiation, and communion, which link it directly to the Christian sacrament of the Mass. But the vessel used to depict the Grail depends only partially on parallels drawn from Christian liturgy, deriving its pictorial form not only from the chalice and paten used in the celebration of the Eucharist, but also from the ciborium and monstrance that during the thirteenth century came to be used in its reservation and veneration. Certain iconographic patterns do emerge: the predominance of the Grail as a shallow bowl, the ‘escuele’, in the first century of Estoire illustration (Rennes, Le Mans, Add., Royal, Amsterdam); the displacement of the ‘escuele’ in Estoire and its replacement by the chalice or the paten in the manuscripts of the fifteenth century (Ars. 3479–80, BNF fr. 117–20, BNF fr. 112, BNF fr. 113–16); a preference for the chalice as the Holy Vessel in Queste illustration in the thirteenth and fourteenth centuries (Bonn, Avignon, Ars. 5218), the emergence of the ciborium as Grail in Queste manuscripts of the early fourteenth century (Add. Royal, Rylands) and the displacement of the chalice by the ciborium in the fifteenth century (Ars. 3479–80, BNF fr. 117–20, BNF fr. 112, BNF fr. 113–16). At the same time there are instances where the shape of the Grail seems to reflect particular interests and present isolated solutions to the question of how to depict it — notably the presence of the Host shown ca. 1250 in UC Berkeley 106, the cross-in-chalice in Bonn 526 in 1286, the tabernacleshrine in Yale 229 (ca. 1295), the allusions to Corpus Christi processions 162 The major groups are (in chronological order): Le Mans 354 and BNF fr 770, to which BNF fr 342 is also related; BNF fr 95 and Yale 229; Amsterdam/Rylands/Douce, Royal 14 E. III, Add. 10292–4: BNF fr 105, 9123, Ars. 3482; BNF fr 117–120. Ars. 3479–80, BNF fr 112, 113–116; for the approximate dates, see the Working List of Illustrated Manuscripts at the end of these essays.

484

and the depiction of the Host in BNF fr. 344 (ca. 1310–1320), and to the reservation of the Eucharist and the emphasis on the element of blood in the Avignon manuscript of 1319, the presence of the Child in the chalice in Ars. 5218 in 1351, the peculiar urn-like shapes given to the Grail in the Udine (ca. 1300) and Florence manuscripts (1446), and the paten shown in BNF fr. 113 (ca. 1475). Some of these are cases where representations of the Grail and its liturgy in the Estoire and the Queste keep pace with, perhaps even anticipate, what is shown in contemporary liturgical and devotional books. There are also some instances where a model drawn from Christian iconography is inappropriate to the context: the Grail-bearing Joseph included at the Crucifixion; Josephé robed and mitred distributing the sacrament rather than Christ, and the corollary of this scene: Christ as well as Josephé administering the sacrament. Although the Christianization of the Grail plays such an important part in the development of its iconography, the healing and nourishing properties of the Holy Vessel itself are an aspect of the Grail that have no parallel in the theology of the Eucharist. From this point of view, the most interesting depictions are those that illustrate the Grail itself, rather than its contents, functioning as the instrument of healing and nourishing — in the curing of King Alphasem in the Estoire, and of the Roi Mehaignié in the Queste-, the nourishing of the knights at the Round Table, the curing of the wounded knight in the presence of Lancelot, and of the cripple who assists Perceval, Bohort, and Galaad in carrying the Grail Table. Yet the illustrative tradition — if such it can be called — makes no distinction among the various types of vessel depicted, so that, depending on which manuscript and which text is concerned, these nourishing or healing aspects of the Grail’s supernatural power are effected either by the Grail as ‘escuele’, or as chalice, or as ciborium — and, in the case of the curing of the Roi Mehaignié in BNF fr. 343, the power of the Holy Vessel itself, shown as a chalice, displaces that of the blood it contains.

XXII The Grail in Rylands MS French 1 and its Sister Manuscripts 1

T

he Lancelot-Graal manuscript, The John Rylands University Library of Manchester, MS French 1, made in Flanders ca. 1315 together with its other components (Amsterdam Bibliotheca Philosophica Hermetica MS 1, and Oxford, Bodleian Library, MS Douce 215),2 is a particularly interesting witness to the transmission of the iconography of this popular Arthurian 1 This is the first of several articles I plan to write on the interrelations among the three sister manuscripts, based in the first instance on direct study of the manuscripts in the original, and then on the complete sets of slides of each that were purchased with funds provided by the Central Research Fund of the Office of Research at the University of Pittsburgh. My work is part of a collaborative study of these manuscripts called the LancelotGraal project in which the other participants are Susan Blackman, Keith Busby, Elspeth Kennedy, Martine Meuwese, Roger Middleton, Kenneth Sochats and Gourai Cai, assisted by graduate students at the University of Pittsburgh. An important aspect of the project is the analysis of scanned manuscript pages in Geographic Information Systems software now under development at the University of Pittsburgh. The present article draws upon the images of the project and on the expertise of the other participants. I also acknowdege the support of the directors and keepers of the libraries housing the manuscripts of the project, especially F.A. Janssen at the Bibliotheca Philosophica Hermetica, Amsterdam; Pamela Porter at the British Library, London; Peter McNiven at The John Rylands University Library, Manchester; Bruce Barker-Benfield at the Bodleian Library, Oxford. Since I wrote this, funds from the National Endowment for the Humanities and the Andrew W. Mellon Foundation enabled us to obtain high resolution scans, photographed by DIAMM in Amsterdam and Manchester, and purchased from the British Library for Royal 14 E.III and Add. 10292–4. This essay was first published in the Bulletin of the John Rylands Library, 81–83, 1999, pp. 55–95. 2 Hereafter Rylands, Amsterdam, Douce, respectively. For the linking of Rylands and Douce, see Pickford, ‘An Arthurian manuscript’, and for the illustrations, Stones, ‘Short note’, reprinted in these essays. For Amsterdam, see Kraus, Cimelia, Catalogue 165, pp. 12–15, and Sotheby’s 7.XII.2010, lot 33, now in private hands. For the date, provenance, and stylistic cognates, Stones, ‘Another short note’.

486

romance.3 Among the many illustrations in Rylands/Amsterdam/ Douce are several scenes that are rarely, if ever, treated elsewhere, even in the two sister copies (British Library, MSS Additional 10292–4 and Royal 14 E. III), which were made by the same team of scribes, decorators, and illuminators. In general, the selection, placing, and treatment of the illustrations among the three related copies are often similar, suggesting that they were produced at about the same time, in very similar circumstances, and shared common models. There are also important differences in treatment between the three copies, which give rise to questions about models, copies, and intentions, on the part of patrons, creators, or both; about the sequence in which these copies were produced; and about the relative precedence of each over the others. The episodes surrounding the Grail and its miraculous appearances are among the scenes in which Rylands/Amsterdam/ Douce and its sister manuscripts are particularly explicit in their rendering of the Holy Vessel in ways which set them apart both from earlier and from later manuscripts, while there are also important differences among the three copies. Here I examine the treatment of the Grail in the Estoire and Queste del saint Graal branches of the five-part romance, in all three related sets of the LancelotGraal.4 3 The manuscripts of the Lancelot-Graal (otherwise known as the Vulgate Cycle or simply as the Lancelot) as a whole are listed in Woledge, Bibliographie, nos 93, 96, 114, and id., Supplément, nos 93, 96, 114, and in Micha, ‚Les manuscrits du Lancelot en prose’, all emended by Stones, ‘The earliest illustrated’, and ead., ‘Aspects of Arthur’s death’, pp. 87–95; see now the list at the end of these essays. The text editions I shall refer to are Le saint Graal, ed. Hucher, based on Le Mans, Médiathèque Louis Aragon, MS 354 (hereafter Le Mans 354); The Vulgate Version, ed. Sommer (hereafter Sommer), based on British Library, MSS Additional 10292–4 (hereafter Additional); La Queste del saint Graal, ed. Pauphilet (1965); L’Estoire del saint Graal, ed. Ponceau, based on Amsterdam and Rennes; La Version post-vulgate, ed. Bogdanow. Complete citations are in the Bibliography at the end of these essays. While writing this article I did not have access to Seynt Graal or the Sank Ryal, ed. Furnivall, based on London, British Library MS Royal 14 E. III (hereafter Royal), or to Yvon, L’Illustration. Studies of Estoire iconography are: Remak-Honef, Text and image in the Estoire del saint Graal; Meuwese, L’Estoire del saint Graal; lists of the subjects in Bibliothèque Nationale de France, MSS fr 112, 113–16, 117–20, together with comparative iconography charts of the Lancelot, Queste and Mort Artu branches of the cycle, are given in Blackman, The Manuscripts and Patronage of Jacques d’Armagnac. The charts are also reproduced in ead., ‘Pictorial synopsis’. For the artistic context, see Loomis and Loomis, Arthurian Legends in Medieval Art ; Stones, The Illustrations of Lancelot’; ead. ‘Arthurian art since Loomis’. 4 For Grail appearances as treated in BNF fr 95 and Yale University, Beinecke Library, 229, see Stones, ‘The illustrations of BN fr 95 and Yale 229’; and for a general outline of the topic, see ead., ‘Seeing the Grail’, both reprinted in these essays. In those articles I explored

THE GRAIL IN MS RYLANDS FRENCH 1

487

In the Lancelot-Graal text, it is only in the Queste del saint Graal, the fourth branch of the five-part cycle, that the Grail is defined and its meaning explained.5 Indeed, a characteristic feature of the Holy Vessel in earlier literature is its imprecise nature: ‘vessel’, in the Perceval of Chrétien de Troyes (c. 1175), the verse Estoire of Robert de Boron, and the prose Joseph (Modena version); ‘vessel’ or ‘escuele’ in the prose Estoire, part of the Lancelot-Graal (early thirteenth century), or something more elemental, like the stone of Wolfram von Eschenbach’s Parzival (c. 1200–10).6 After introducing the Grail as ‘li sainz Graal’ or ‘li sainz Vessel’ (or a variant thereupon), it is in the context of the liturgy in the Queste that the nature and significance of the Grail is explained: ‘Ce est [...] l’escuele ou Jhesucriz menja l’aignel le jor de Pasqes o ses deciples. Ce est l’escuele qui a servi a gré toz çax que j’ai trovez en mon servise; ce est l’escuele que onques hons mescreanz ne vit a qui eie ne grevast molt. Et por ce que ele a si servi a gré toutes genz doit ele estre apelee le Saint Graal’ (Pauphilet, p. 270). What is striking about Grail illustrations in the Queste, however, is that the shallow-bowl shape evoked by the word ‘escuele’ is never (among the extant manuscripts) used to represent the Grail; rather, it is in illustrations of the Estoire that this form is preferred. Even then, the Grail as a bowl is found only in certain groups of manuscripts, although the word ‘escuele’ also figures prominently — if without explanation — for the Grail in the Estoire text. In what follows, I compare the treatment of the Grail in the three sister manuscripts, taking each illustrated episode in the order in which it occurs in the text of the Estoire and the Queste. I include reference to the treatment of the Grail elsewhere in the illustrative tradition in order to demonstrate the ways in which the three sister manuscripts differ from, or are similar to, earlier and later copies to which they are unrelated stylistically.

parallels for the Grail and its cult in contemporary liturgical and devotional practices, especially devotions to the sacrament, relics of the Passion, and eucharistie and relic miracles. Here I concentrate upon the interrelations among the three sister copies. 5 See also the wording of the rubrics to the Crucifixion in Royal and Amsterdam, cited in note 10 below. A useful comprehensive study of the Grail is Burdach, Der Graal; more popular studies are Matthews, Grail: Quest; Bouyer and Mentré, Les lieux magiques du Graal; Scènes du Graal, eds. Buschinger, Labia, and Poirion. 6 See Bertoni, Materiali.

488

The Grail at the Crucifixion The Grail makes its first appearance in the Lancelot-Graal towards the beginning of Estoire, where, after the Crucifixion, Joseph of Arimathea visits the room in which Christ celebrated the Last Supper with the disciples,7 and takes the vessel Christ used, the ‘escuele’ (Sommer, I, p. 13; Ponceau, I, p. 25) in order to collect Christ’s blood in it at the Entombment. Often what is depicted here in Estoire manuscripts is not Joseph’s return to the upper room — an episode that is never, to my knowledge, depicted in the Estoire tradition — or even the Entombment, but a Crucifixion scene showing Christ between the Virgin Mary and St John the Evangelist, without Joseph — a scene that would have been among the standard repertoire of any medieval artist. Nothing is said at this point in the text about Joseph collecting the blood of Christ at the Crucifixion, yet the Additional, Royal, and Amsterdam manuscripts (Fig. 1) form one of two groups of manuscripts that show him doing so then, rather than at the Entombment. The second group of manuscripts that show Joseph collecting Christ’s blood at the Crucifixion were sold in the early fifteenth century by the Parisian book-dealer Jacques Raponde. These include the set of volumes bought by Jean de Berry in 1405 and inherited ca. 1465 by his great-grandson Jacques d’Armagnac (BNF fr 117–120, Fig. 2).8 Only occasionally is Joseph shown actually collecting the blood of Christ at the Entombment as the text specifies: ‘si concueilli le degout du sanc tant com il en puet avoir, si lemist en l’escüele, puis reporta l’escüele en sa maison’ (Sommer, I, p. 14; cited from Ponceau, I, p. 25). The Bonn manuscript made in 1286 (Bonn, Landes-und-Universitätsbibliothek, MS 526 (hereafter Bonn 526), f. 1) includes this as the final scene in the composite arrangement of images that opens the Estoire, and Bonn 526 and a closely related and possibly earlier manuscript (BNF, fr 19162, f. 6) both show Joseph holding a bowl-shaped receptacle. To my knowledge, the only other manuscript to include this motif is the set made for Jacques

7 For a rare instance of what I think is the influence of Grail imagery reflected in Last Supper iconography, see Stones, ‘Madame Marie’s picture-book: a precursor’, p. 52. See also the Breviary of the Cathedral of Saint-Lambert, Liège, discussed below. 8 Meiss, French Painting in the time of Jean de Berry, I, p. 252; for the repainting done under Jacques d’Armagnac, and for Jacques d’Armagnac’s own commissions, see Blackman, Manuscripts, pp. 182–245, and ead., ‘Pictorial synopsis’, p. 4.

THE GRAIL IN MS RYLANDS FRENCH 1

489

d’Armagnac ca. 1475 (BNF fr 113, f. 7), where the Grail is a shown as a flat, paten-like receptacle.9 In Amsterdam’s Crucifixion (f. 6v, Fig. 1), Joseph is shown seated on the ground at the foot of the cross, holding the ‘escuele’, with Mary and John also present, while Additional (f. 3v) and Royal (f. 7) show a more elaborate version: Joseph holding the ‘escuele’, and the two thieves as well as Mary and John.10 In Additional there is a further unusual anecdotal detail: the fate of the souls of the dead thieves is included, that of the good thief received by an angel, that of the bad thief hooked by a devil.11 The three related copies are consistent, and alone, in showing the Grail as a shallow dish, painted in silver,12 reminiscent of the vessel in which the blood of sacrifice is collected in Old Testament iconography.13 In the other illustrated manuscripts that include Joseph at the Crucifixion, the vessel he holds is shown as a chalice: the Raponde manuscripts from ca. 1405 (Paris, Bibl. de l’Arsenal, MS 3480, f. 48314 and BNF fr 120, f. 520, Fig. 2) show a simple Crucifixion with Mary and John, and Joseph of Arimathea as a much smaller figure robed in a long mantle kneeling on the left, holding out a chalice in his outstretched hands. The Crucifixion is not depicted in the copy of the Estoire made for Jacques d’Armagnac ca. 1475 (BNF fr 113), but it is interesting to note that the opening miniature of the Queste part of the same set of volumes (BNF 9 See Blackman, Manuscripts, p. 505; Meuwese, Estoire, fig. 36. Blackman’s iconographical tables (Manuscripts, pp. 213–43) show numerous instances of iconographical similarity between Amsterdam/Douce/Rylands and BNF fr 113–116, but the selection of Grail scenes and their treatment are not uniformly comparable. 10 The Royal rubric comments: ‘Ensi que Josephes recoilli le degout du sanc qui issoit des plaiies n[ost]re seigneur qui puis fu apeles li s[ains] g[ra]alz’. Here as elsewhere, spellings and punctuation have been normalized, and square brackets indicate missing letters. Amsterdam’s rubric is less graphic, and is written in part over erasures: ‘Chi est ensi q[ue] Joseph rechut [erasure] le sanc n[ost]re sign[eur] qui puis fu apeleis graalz.’ It is interesting that the rubric in Additional is considerably more generic: ‘Ensi q[ue] Joseph recuelle les goûtes de sa[n]c e[n] .j. esq[u]ele’. 11 Martine Meuwese compared the three Crucifixions in lectures in Pittsburgh in 1995 and at the Leeds International Medieval Conference in 1997. See ead., ‘Three Illustrated Prose Lancelots’, and ead., ‘Inaccurate Instructions’. 12 Without technical analysis I cannot determine whether this is really silver, or a metal substitute, such as tin. The same problem exists wherever ‘silver’ is used — I shall mention it several times in the description of these miniatures — and the possibility that another metal was used should always be borne in mind. 13 I have commented upon the Old Testament parallels in Stones, ‘Seeing’, n. 27. 14 Meuwese, Estoire, fig. 35.

490

fr 116, f. 607) is a Crucifixion where five angels catch the blood of Christ in chalices, following a model common in Crucifixion iconography, and Joseph is omitted. The association of a chalice-shaped vessel with the Crucifixion is a common topos in Christian iconography, linking Christ’s death on the Cross with the sacrament of his body and blood in the Eucharist,15 and the emerald or onyx chalice of the Last Supper was allegedly preserved by the sixth century in Jerusalem at a site adjacent to that of Golgotha, a locational juxtaposition that aroused much interest a century later in Britain through the description and drawing of the site by bishop Arculf of Gaul, transmitted by Adamnan of Iona and Bede.16 It is not unlikely that the artists who depicted Joseph with a chalice at the Crucifixion were drawing upon these traditions, at least in part. So it is all the more significant that the three sister manuscripts, and they alone, follow a different direction by showing the Grail as the ‘escuele’ of the text. The Grail Liturgy The next Grail appearance to be illustrated in the Rylands/ Amsterdam/ Douce set of volumes is the Grail liturgy, celebrated by Josephé, son of Joseph of Arimathea and first Christian bishop (Sommer, I, pp. 33–34; Ponceau, I, pp. 86–88). None of these manuscripts shows the ark that Joseph had made to house the Grail, nor do they illustrate the blinding of Nascien, punished for looking into the Grail, and his subsequent healing — rare scenes in the illustrative tradition as a whole.17 The account of the Grail liturgy given in the Estoire is divided into several parts, and the three manuscripts offer varying pictorial emphases depending on which part — if any — is selected for illustration.

The most comprehensive study is Malcor, The Chalice at the Cross, which, as the title implies, includes several spurious Grails. See also the standard iconographical reference works, Schiller, Ikonographie, 2 II, pp. 98–162; and Kirschbaum, ed., Lexikon der christlichen Ikonographie, II, col. 606–42. I reserve detailed treatment of the motif of the chalice at the Crucifixion and its potential Grail associations for another occasion. 16 The various accounts are drawn together in Breviarius de Hierosolima, and Wilkinson, Jerusalem Pilgrims before the Crusades. I return elsewhere to the link between these early references to the chalice of the Last Supper and the so-called ‘Grails’ of Valencia and Genoa. 17 I discuss the few illustrations of these scenes in Stones, ‘Seeing’, nn. 41–2. 15

THE GRAIL IN MS RYLANDS FRENCH 1

491

There are also differences in the illustrative tradition as a whole as to which of the various scenes in this episode are selected and how they are illustrated. The entire liturgy is omitted in the earliest surviving manuscript (Rennes, BM, MS 255, ca. 1220; hereafter Rennes 255). The opening scene is represented only once, to my knowledge: ca. 1280 in Le Mans 354 (f. 30, Fig. 3). What is shown is the moment when Josephé, at the invitation of Christ, opens the door of the ark and sees a man (Christ Himself ) robed in red brighter than burning lightning (‘foudres ardans’), his face, hands and feet of the same colour, and five angels (four in the image) also the same colour, with six wings of burning fire (‘de fu ardant’), two holding the three nails, all bleeding, the third holding the bleeding lance, the fourth holding to Christ’s mouth a bleeding sponge, and the fifth holding a bleeding scourge; in their hands the angels hold scrolls inscribed ‘çou sunt les armes par coi li gugières, qui chi est, vainqui la mort et detruist’ (cited from Hucher, II, pp. 174–75; see also Sommer, I, p. 32; Ponceau, I, pp. 72–73). I have commented elsewhere on the textual identification of the Instruments of the Passion as the Arma Christi, a motif which as an independent devotional image does not appear before 1300, although of course the Instruments themselves are depicted much earlier in the context of the Passion and the Last Judgement.18 At the conclusion of this episode, Josephé falls into a deep reverie, from which he is roused by a voice, and he sees Christ dying on the cross, the nails, lance, and sponge all in place and the blood of Christ flowing into the vessel which his [Josephé’s] father had placed in the ark, filling it to overflowing (Hucher, II, p. 176; Sommer, I, p. 32; Ponceau, I, p. 74). This description certainly echoes the miniature shown earlier, accompanying the account of the Crucifixion, in Amsterdam, Royal, and Additional (but not in Le Mans), which showed Joseph collecting Christ’s blood in the Grail — although the analogy is not complete as no mention is made of Joseph holding the Grail at this point, or even being himself present at this vision of the Crucifixion, which is seen only by his son Josephé. Josephé’s vision, so far as I know, is illustrated only in Paris, BNF, MS fr 105, f. 19 [and also in Brussels, BR 9246, f. 24v]. The text’s account of it may contain a hint as to its possible source and provide another link to the accounts of the holy places in Jerusalem discussed above. A curious line in the description of the vision asserts that it was impossible to tell from what kind of wood the cross 18

Stones, ‘BN fr 95 and Yale 229’, pp. 224–45, and ‘Seeing’, n. 39.

492

of the Crucifixion was made. Most of the pilgrim accounts discussed above mention the crosses preserved in the basilica that Constantine constructed over the spot where his mother Helena had excavated them. One of the accounts, that of the Piacenza Pilgrim (c. 575), states that the wood of the cross was of the nut tree.19 After the description of Josephé’s vision comes the episode where Joseph joins Josephé outside the ark and looks in. What Joseph sees is shown in one of the miniatures illustrating the Grail liturgy in Amsterdam and Royal. Additional omits illustrations of the liturgy altogether, a very suprising divergence from the selection made in the other two copies, but it is by no means the only instance in which the pictures, and picture-cycles, in the three manuscripts offer marked differences.20 The first image in Amsterdam (f. 18, Fig. 4) shows Josephé in the ark, robed as a bishop, accompanied by an angel holding a silver ewer and two more angels kneeling to the left of the ark, one holding an incense-boat and swinging a thurible. As his father Joseph looks into the ark, shown as a Gothic church,21 he sees a small altar, covered with a white cloth (Sommer, I, pp. 33–34; Ponceau, I, p. 75).22 On the altar are three bleeding nails and the top of a bleeding lance; a silver chalice-like vessel on a stem with a lid, mostly covered by a cloth painted red in the miniature (though white in the text), and partially veiled; and the silver bowl-shaped ‘escuele’ — the Grail. The Instruments of the Passion (shown in Amsterdam’s image), and also a hand holding a cross and two hands holding candles (none of which are shown in Amsterdam’s image) are specified in the text (Sommer, I, pp. 33–34; Ponceau, I, p. 75). The Royal version (f. 15v, Fig. 5) is similar but much simpler in treatment, omitting the elaborate architectural ark, 19 Wilkinson, Jerusalem, p. 83, paragraph 20, with references citing Cyril of Alexandria’s interpretation of Cant. 6: 10, ‘I went down to the garden of nuts’ (in hortum nucum), as the garden in which Christ appeared risen. 20 Papers given by Elspeth Kennedy and Martine Meuwese at the Leeds International Medieval Conference in 1997 drew attention to divergencies in the Lancelot and in other scenes in Estoire. Publications are forthcoming (see n. 11 above and †Kennedy and Stones, ‘Signs and Symbols’). 21 This is quite different from the centrally planned ‘Grail Temple’, based on the Temple of Jerusalem (or on the Church of the Holy Sepulchre), described in Albrecht’s Jüngere Titurel, composed in Middle High German ca. 1270, which is the starting-point for Ringbom’s wideranging study, Graltempel und Paradies. See also Rosenau, Vision of the Temple, a reference for which I thank Martine Meuwese. 22 The text also mentions a red cloth, not shown here.

THE GRAIL IN MS RYLANDS FRENCH 1

493

also omitting the hands holding candles and the lance and nails; but it does include the red cross held by a hand above the altar, and so is not merely a simplification of the Amsterdam image. Furthermore, this is not the first image in the liturgy sequence in Royal, but the second. On f. 14v is a lead-in scene, showing Josephé’s followers, both male and female, kneeling at a draped altar on which is a partially covered chalice; the head of God emerges from a cloud above (Fig. 6) (Sommer, I, p. 30; Ponceau, I, p. 69). This is unique, to my knowledge, in the illustrative tradition of the Estoire. The hands holding the cross and candles are rarely shown in Estoire illustration: an example occurs one of Jacques d’Armagnac’s manuscripts (BNF fr 113, f. 18v, Fig. 7). There, the hands holding cross and candle emerge from the reredos behind the altar, on which rest the three nails and the tip of the lance, a round ciborium-like vessel, and a flat, dish-like paten. Is the Grail the ciborium or the paten? The more elaborate vessel is undoubtedly the ciborium, yet in the rare depiction of the collecting of Christ’s blood at the Entombment found earlier in this manuscript, Joseph used a shallow dish that is remarkably similar to the paten shown on this altar. So this time the proliferation of sacred vessels is not simply based on the multiple vessels of the Eucharist, but also links back to Joseph’s pious act earlier in the story. Both picture and rubric focus on Josephé contemplating the Grail in a chapel: ‘C[i] Josephez estoit devant le saint Graal a genoulx’. Josephé, wearing civilian clothes and bare-headed, kneels alone before the altar; in the doorway behind him stand four wingless angels, one of whom holds a situla and aspergillum, while another holds a processional cross. The emphasis in BNF fr 113 is different from the Amsterdam scene, where the psychological interaction between Josephé and Joseph is a significant part in the miniature that is especially underlined in the rubric. Clearly added after the miniature, as its letters avoid the pinnacles of the church structure which project into its space, the rubric states ‘Chi canta Josephus messe et mist son pere hors du tabernacle’ (Fig. 4).23 The Josephé-Joseph interaction As on most pages in this manuscript and the others made by the same team, notes for the rubricator, in a compressed cursive hand, and a second set of notes, for the illuminator, in a larger, less compressed script, are found in the bottom margin beneath the picture; see Stones, ‘Indications écrites’, figs. 19–20, and Alexander, Medieval Illuminators, pp. 112–15. [The exclusion of Joseph recalls the exclusion of the laity from the altar in the Decretals of Gregory IX, Book III: see for instance, Cambrai MM 619(571), f. 123, probably from Toulouse (Stones, Gothic Manuscripts, Part II, vol. 1, Cat. no. VII–17, ill. 379). I do not know of a copy made by the artists of the Amsterdam manuscript.] 23

494

is still further enhanced in Royal by the exclusion of the architectural setting and the censing angels, giving greater emphasis to the two figures and the altar. Space was left for a rubric that was never added above the miniature, so we do not know how its wording would have compared with that of Amsterdam. After this follows a lengthy explanation of the meaning of the vestments which Josephé is wearing as first bishop, after which Christ asks Josephé to celebrate his first mass (Sommer, I, p. 40; Ponceau, I, p. 87). Four manuscripts depict the moment at which Josephé consecrates the host. The small historiated initial made perhaps in Paris [or Jumièges] at an uncertain date, perhaps as early as 1250 (University of California, Berkeley, MS UCB 106 (hereafter Berkeley 106), f. 126, Fig. 8), may be the earliest surviving depiction: Josephé, bareheaded, stands before an altar on which is a draped chalice. A golden host hovers in mid-air, and the Head of God looks down from a cloud above. This is a depiction which certainly suggests the veneration of the Eucharist, and perhaps alludes also to its reservation — neither of which are depicted in liturgical illustration in the West before ca. 1300, although by 1215 transubstantiation was already an assumption in the statues of the Fourth Lateran Council, and the Christ Child in the host had been depicted in twelfth-century wall-painting in the East.24 Much more commonplace is the illustration in BNF, MS fr 770 (hereafter BNF fr 770), made on the Artois-Flanders border ca. 1280, in close stylistic association with Le Mans 354, which simply depicts an altar with a chalice on it and a group of figures kneeling before it (f. 19v). By the end of the century, however, Estoire illustration shows several parallels with the iconography of transubstantiation.25 In BNF fr 95, made in the diocese of Thérouanne ca. 1290,26 the Grail liturgy illustration is similar to that of BNF fr 770, but what is shown is the moment of the elevation of the host, with Josephé bareheaded and his followers kneeling on the ground behind (f. 18, Fig. 24 Tanner, Decrees of the Ecumenical Councils, I, pp. 227–303. For a fuller discussion of the implications of this see Stones, ‘BN fr 95 and Yale 229’, p. 222 and nn. 94–96. Christ as Child in the host is found at Kurbinovo in 1191, and Kastoria and Prilep, all in the archdiocese of Ochrid (former Yugoslavia); see Hadermann-Misguich, Kurbinovo: les fresques, and Babić, ‘Discussions christologiques et décor’. I thank Areti Papanastasiou for these references. 25 For a more general outline of the topic, with reference to eucharistie miracles, the reservation of the sacrament, Corpus Christi, the cult of the Holy Blood, bloodletting, and the Italian Grail traditions, see Stones, ‘Seeing’. 26 See Stones, ‘BN fr 95 and Yale 229’, p. 221.

THE GRAIL IN MS RYLANDS FRENCH 1

495

1. olim Amsterdam, BPH 1, i, L’Estoire del saint Graal, f. 6v, Joseph collecting Christ’s blood at the Crucifixion (photo: Lancelot-Grail Project)

2. Paris, BnF fr 120, La Queste del saint Graal, f. 520, Joseph collecting Christ’s blood at the Crucifixion (photo: Bibliothèque nationale de France)

496

3. Le Mans, MM 354, L’Estoire del saint Graal, f. 30, Josephé celebrating a Grail Mass in the Ark accomanied by Christ and angels bearing the Instruments of the Passion (photo: Lancelot-Grail Project)

THE GRAIL IN MS RYLANDS FRENCH 1

497

4. olim Amsterdam, BPH 1, i, L’Estoire del saint Graal, f. 18, Josephé bidding his father leave the Ark (photo: Lancelot-Grail Project)

5. London, BL Royal 14 E.III, f. 15v, L’Estoire del saint Graal, Josephé bidding his father leave the Ark (photo: British Library)

498

6. London, BL Royal 14 E.III, f. 14v, L’Estoire del saint Graal, Joseph and his company before the Grail; Head of God emits rays (photo: British Library)

7. Paris, BnF fr 113, L’Estoire del saint Graal, Josephé kneels before the Grail (photo: Bibliothèque nationale de France)

THE GRAIL IN MS RYLANDS FRENCH 1

8. University of California, Berkeley, 106, L’Estoire del saint Graal, f. 126 Josephé before the Grail (photo: University of California, Berkeley)

9. Paris, BNF fr 95, L’Estoire del saint Graal, f. 18, Josephé elevates the Host, blessed by the Hand of God (photo: Bibliothèque nationale de France)

499

500

10. olim Amsterdam, BPH 1, i, L’Estoire del saint Graal, f. 21, Christ and Josephé administer the sacrament to the faithful (photo: Lancelot-Grail Project)

11. London, BL Royal 14 E.III, L’Estoire del saint Graal, f. 17v, Christ and Josephé administer the sacrament to the faithful (photo: British Library)

THE GRAIL IN MS RYLANDS FRENCH 1

12. Darmstadt, Hessische Landes-und Hochschulbibliothek 394, Breviary of the Cathedral of Saint-Lambert, Liège, vol. 2, f. 209v, Bishop celebrating mass, blessed by Christ (photo: author)

501

502

13. Rennes, BM 255, L’Estoire del saint Graal, f. 76v, The Grail borne to Norgales (photo: Lancelot-Grail Project)

14. olim Amsterdam, BPH 1, i, L’Estoire del saint Graal, f. 88v, Josephé bearing the Grail to Norgales on the hem of his alb with followers and unbelievers (photo: LancelotGrail Project)

THE GRAIL IN MS RYLANDS FRENCH 1

15. London, BL Add. 10292, L’Estoire del saint Graal, f. 55v, Josphé transporting the faithful on the hem of his alb, the unfaithful in the water (photo: BL)

16. olim Amsterdam, BPH 1, i, L’Estoire del saint Graal, f. 114v, Josephé on his deathbed giving the Grail to Alain (photo: LancelotGrail Project)

503

504

17. London, BL Royal 14 E.III, L’Estoire del saint Graal, f. 86, Josephé on his deathbed giving the Grail to Alain; King Alphasem cured of leprosy by the Grail and baptised (photo: British Library)

THE GRAIL IN MS RYLANDS FRENCH 1

18. Paris, BnF fr 105, L’Estoire del saint Graal, f. 122, The Grail given by Helein to King Arphasam (photo: Bibliothèque nationale de France) 19. olim Amsterdam, BPH 1, i, L’Estoire del saint Graal, f. 115v, King Alphasem wounded by a hermit holding the Grail (photo: Lancelot-Grail Project)

505

506

20. Paris, BnF fr 113, L’Estoire del saint Graal, f. 112v, King Arphasam, cured before Alain (photo: Bibliothèque nationale de France)

THE GRAIL IN MS RYLANDS FRENCH 1

21. Paris, BnF fr 116, La Queste del saint Graal, f. 610v, The Grail appears at Arthur’s Round Table (photo: Bibliothèque nationale de France)

22. Manchester, The John Rylands University Library, French 1, La Queste del saint Graal, f. 184v, The Grail borne before King Arthur, Queen Guinevere, and three knights all at table (photo: Lancelot-Grail Project)

507

508

23. London, BL Royal 14 E. III, La Queste del saint Graal, f. 99, Lancelot sleeps while a wounded knight is cured by the Grail (photo: British Library)

24. Manchester, The John Rylands University Library, French 1, La Queste del saint Graal, f. 195v, Lancelot sleeps while a wounded knight is borne before the Grail (photo: Lancelot-Grail Project)

THE GRAIL IN MS RYLANDS FRENCH 1

25. Paris, BnF fr 117, L’Estoire del saint Graal, f. 1, Episodes from the Life of Lancelot du Lac (photo: Bibliothèque nationale de France)

509

510

26. Paris, BnF fr 112, Lancelot-Graal, f. 15v, Lancelot asleep as the wounded knight gives thanks before the Grail for his cure (photo: Bibliothèque nationale de France)

THE GRAIL IN MS RYLANDS FRENCH 1

511

9), a subject that is just beginning to be depicted in liturgical illustration in the thirteenth century.27 Josephé elevates a eucharistie wafer, not the human figure specified in the text, but the host is inscribed with the letters IHS of the Name of Jesus, and so contains a cryptographic reference to the transubstantiation. On the altar is the partly veiled Grail, shown as a chalice. This eucharistie emphasis continues into the fifteenth century, for instance in part of the opening miniature of the Estoire in Yale 227, written by Jean de Loles in Hainaut in 1357, showing Josephé celebrating mass at an altar on which are a draped chalice and a large eucharistie host with a cross between dot motifs inscribed on it (f. 12). But at the place in the text where the liturgy is described, there is no corresponding illustration. None of the surviving Estoire manuscripts explicitly illustrate the text’s description of the transubstantiation where the bread is literally transformed into a child, which Christ orders Josephé to eat — details depicting this are avoided.28 The second miniature in Amsterdam (f. 21, Fig. 10) shows Christ and Josephé (wearing his mitre) administering the sacrament — a eucharistic host — to the kneeling assembly (Sommer, I, p. 40–41; Ponceau, I, pp. 87–88). Christ is framed by the open transept of the church-like ark, with a crossed halo, robed as a priest, and even tonsured; Josephé, also robed as a priest, still wears his mitre. They each hold a shallow bowl painted in silver, like the one on the altar in the previous miniature; two angels are also included, one holding another shallow bowl (his incense boat) and swinging a thurible, while the second angel, in a stooping pose on the left next to Christ, uses both hands to support a huge silver chalice. This angel, wearing an orange robe, ought to be the one described in the text as ‘tous daarains’, but if so he should be carrying the ‘escuele’, not the chalice. The Royal version (f. 17v, Fig. 11) is simpler, showing the altar draped with a cloth, upon which are a gold chalice and a huge pile of wafers; Christ is again included, holding the Grail, shown as a gold bowl-shaped vessel, and blessing a kneeling man, probably Joseph, while Josephé, robed as bishop, administers the host to the rest of the assembly. In the text, the sacrament is then distributed to the faithful by an angel — not by Josephé, as in these miniatures, and certainly not by Christ 27 See Vloberg, L’eucharistie dans l’art, I, p. 63, and Kennedy, ‘The moment of consecration’, pp. 146–7. 28 I describe in Stones, ‘BN fr 95 and Yale 229’ and ‘Seeing’, how the illustrations of Queste manuscripts, whose text is equally literal in its description, treat the representation of the liturgy, and show the extent to which the pictures may be paralleled or anticipated in scenes showing Eucharistic miracles.

512

— so Christ’s presence, and his and Josephé’s actions, are at variance with the text and require explanation.29 They most likely refer back to the episode just after Joseph looks into the ark and just before Josephé’s mass — but which is nowhere illustrated — where Christ makes an appearance robed as a priest and gives Josephé a lengthy explanation as to the meaning of each of his episcopal vestments and his responsibilities as bishop (Sommer, I, p. 34; Ponceau, I, p. 77). The presence of Christ probably also reflects knowledge of the text and illustrative tradition of the Queste, in which it is indeed Christ who distributes the host to the knights. But the addition of Christ here is all the more interesting as none of these three manuscripts includes an illustration of the liturgy in the Queste.30 A second Grail liturgy image is also included in BNF fr 113, showing, in a version much closer to the text of the Estoire, three angels administering the communion to Joseph and his company, without Christ or Josephé (f. 21v).31 The pairing of Bishop Josephé with Christ makes an interesting reappearance in the context of the mass in a liturgical book that is closely related stylistically to the Royal, Additional, and Amsterdam/Manchester/ Oxford Lancelot-Graal manuscripts. Several devotional and liturgical books were produced by this team, the stylistic variants of which I reserve for more detailed treatment elsewhere.32 Among them is the breviary of the Cathedral of Saint-Lambert, Liège, in Darmstadt (Hessische Landes-undHochschulbibliothek, MS 394; hereafter Darmstadt 394). On f. 209v of vol. II is a miniature depicting a bishop, still mitred, at an altar with a chalice on it (Fig. 12). On the other side of the altar stands Christ, leaning over towards the bishop and blessing him; in an arch frame above, a priest elevates the host. This illustration accompanies the antiphon at Vespers of the Sacrament. Nothing proves that the bishop in this miniature ‘is’ Josephé, and the vessel on the altar is a chalice, not the ‘escuele’, but the scene is sufficiently unusual in liturgical illustration to suggest that some kind of visual association with the Grail liturgy may well be intended here. A second instance is one I have discussed elsewhere, suggesting that the Last Supper image in the devotional See Stones, ‘BN fr 95 and Yale 229’, figs 8, 19. Illustrations of the Queste liturgy in other manuscripts are discussed below. 31 Blackman, Manuscripts, p. 505. No image of the liturgy is in BNF fr 112, the special version made for Jacques d’Armagnac, or BNF fr 117–120, the copy he inherited from Jean de Berry. 32 For preliminary observations, see Stones, ‘Another short note’, and, for a different view, Oliver, Liège, II, pp. 201–02, pl. 210–21. 29 30

THE GRAIL IN MS RYLANDS FRENCH 1

513

picture-book made ca. 1285–1300 for a lady of Mons (probably Marie de Rethel, dame d’Enghien), BNF, MS naf 16251, may be another rare instance where Grail imagery is reflected in a sacred context.33 There, St Peter hands Christ a shallow bowl which he has just filled from a jug, which Christ will use for the conseration of the wine of the Eucharist.34 The rarity of the motif, the fact that the vessel is clearly to be used by Christ, and the formal links with the ‘escuele’ of Grail depictions, are factors that suggest a deliberate parallel with Grail iconography, even though, in this instance, no LancelotGraal manuscript illustrated by the painter, Master Henri, has so far come to light. Josephé’s Journey to Norgales The Grail is subsequently taken by Josephé and his followers to Norgales, a journey miraculously accomplished on the hem of his alb (Sommer, I, p. 211; Ponceau, II, pp. 416–17). Amsterdam, Royal, and Additional all include a miniature showing this episode, which is common throughout the illustrative tradition from its beginnings: it is in this context, surprisingly, that the earliest depiction of the Grail is found, in Rennes 255 (f. 76v, Fig. 13). The Rennes manuscript omits the Crucifixion, Entombment, and other scenes that might have afforded an opportunity to show the Grail earlier in the story.35 The journey is simply depicted in the restricted space of the historiated initial in Rennes 255. A man wearing a round-crowned Jewish hat in the style of a skull-cap carries the silver ‘escuele’ in his hands, and is followed by supporters. The Grail-bearer is one of Josephé’s followers, as in the text, rather than Josephé himself, but the group should be walking See note 7 above. The standard iconographical dictionaries are unhelpful in tracing this detail. While an ‘escuele’-shaped bowl is extremely rare at the Last Supper, there are several parallels for an apostle holding a vessel with a stem, and for the pouring of liquid into it, in twelfth and thirteenth-century English, Walloon, and German manuscripts, and in metalwork, which I reserve for detailed treatment elsewhere. 35 See Stones, ‘Earliest’, where the date is argued on the basis of stylistic parallels with royal manuscripts of ca. 1220. 36 BNF fr 770, f. 98, and fr 9123, f. 79 show a very similar, non-text-specific, depiction. Berkeley 106, Le Mans 354, BNF fr 95, and Yale 227 have no illustration of this; it is in BNF fr 113, f. 88, but without the Grail. See Blackman, Manuscripts, p. 506. Other copies made in the second half of the thirteenth century show the Grail in a variety of other ways: as a vessel with a gold cover, as a covered reliquary or monstrance, anticipating representations 33 34

514

upon the water, not clear in the image.36 In Amsterdam (f. 88v, Fig. 14), Royal (f. 66v), and Additional (f. 55, Fig. 15), the scene focusses on Josephé, shown robed and mitred as a bishop, in the company of his followers, and on the water. There are several differences: Amsterdam shows Josephé standing on the left and his followers standing on the hem of his alb which is spread out over the water, as in the text. Royal and Additional again align with each other against Amsterdam by placing Josephé on the right of the composition, holding a separate cloth on which stand his followers. Martine Meuwese has suggested that this variant may derive from biblical illustration such as the story of Elijah in the Book of Kings, striking the water of the River Jordan with his mantle to divide it so that he and Elisha may cross.37 This is depicted, for instance, in the early thirteenth-century Bible moralisée. Additional and Royal both exclude the Grail in this episode — correctly in relation to what the text says, since Josephé’s followers had already set off with it before he himself started his journey — but in Amsterdam, Josephé holds the Grail, shown again as a bowl-shaped vessel in silver. As Meuwese has also pointed out, the inclusion of the Grail in Amsterdam is justified in relation to the importance of his rôle as its guardian. Helein chosen, and the Grail transmitted to him by Josephé At its next appearance, Helein (Alain), having caught a large fish, is asked by Josephé to pray before the Grail for his fish to be multipled (Sommer, I, p. 251; Ponceau, II, p. 493). What is shown in Amsterdam (f. 102v) is not Helein before the Grail, but the preceding scene where Helein, in a boat, catches the fish, and no Grail is present. Additional (f. 65) and Royal (f. 77v) both show the Grail, as a gold ciborium, placed on the laden table in front of which kneels Helein and behind which sit Josephé and his company. This is the only Grail scene which is omitted in Amsterdam but present in the other two copies, neither of which, however, shows the fishing scene. This is also the first time the Grail is shown as a ciborium, and anticipates the choice of ciborium for the Grail that will characterize subsequent Grail depictions of Corpus Christi processions that would not be shown in a sacred context for another generation or so. I discuss these variations, and suggests parallels in Christian art, in Stones, ‘Seeing’, nn. 87–9. For further discussion, see Meuwese, ‘Three illustrated prose Lancelot manuscripts’. 37 See the unpublished lecture given by Meuwese at the International Arthurian Congress in Garda in 1996.

THE GRAIL IN MS RYLANDS FRENCH 1

515

in Additional and Royal, against Amsterdam/Rylands. There, the ‘escuele’ continues as the preferred shape to the end of the Estoire illustrations, giving way to the chalice only in the work of the second painter who did the Queste Grail scenes in Rylands. The transmission of the Grail from Josephé on his deathbed to Helein is an episode illustrated in all three copies (Sommer, I, p. 286; Ponceau, II, p. 557).38 Amsterdam (f. 114v, Fig. 16), aligning with an earlier copy (BNF fr 770, f. 117v), shows the Grail again as the ‘escuele’, whereas in Royal (f. 86, Fig. 17) and Additional, it has become a gold covered ciborium with a stem. This version of the Grail is similar to what is shown in another later copy, Yale 227 (dated 1357),39 but the ‘escuele’ shape can also be paralleled in a variant on this scene in a manuscript of ca. 1330 (BNF, MS fr 105, f. 122, Fig. 18), where the Grail, shown as the ‘escuele’, is given by Helein to King Arphasam (Alphasem) (Sommer, I, p. 286; Ponceau, II, p. 562).40 The Curing and Wounding of King Arphasam The final episodes in the Estoire include miraculous acts performed by the Grail concerning King Arphasam: the cure of his leprosy, effected by the Grail after his baptism (Sommer, I, p. 288; Ponceau, II, p. 561), and his wounding in the groin, as punishment for having presumed to sleep in the Grail Castle (Sommer, I, p. 289; Ponceau, II, p. 563). Neither scene is common. Royal (f. 86, Fig. 17) is the only one of the three related copies to include the cure of the king, but the Grail is omitted in the illustration of the wounding (f. 87). As in the earlier copy, Bonn 526, the baptism and cure are combined into a single scene in which the Grail is present; the latter is shown in Bonn as a chalice (without an accompanying cross, a motif commonly used in conjunction with the Grail in Bonn), and in Royal as a covered ciborium, as introduced in the scene of the multiplication of Helein’s fish. The wounding alone is shown in Amsterdam (f. 115v, Fig. 19), where a cleric stands by the king’s bed, wearing a maroon cope over his alb, holding in his left hand the Grail — the usual silver bowl — and in the other a lance. In the corresponding miniature in Additional (f. 74), a bearded man, wearing a Rennes 255 shows Helein weeping at Josephe’s deathbed, without the Grail (f. 96v). See Stones, ‘Seeing’, nn. 90–2. 40 For the stylistic context of this manuscript, a product of the prolific output of the Fauvel Master and his associates, see Stones, ‘The artistic context of the roman de Fauvel’. 38 39

516

hooded robe and painted orange, pierces the King (through the bedclothes) with a spear, but the Grail is not present.41 One of Jacques d’Armagnac’s manuscripts (BNF fr 113) shows variant versions of this episode: firstly, King Arphasam and his followers standing before Helein, having been cured, without the Grail depicted (f. 112v, Fig. 20); and secondly, an angel hovering above the wounding scene, holding the Grail (f. 113).42 The version in Additional is closest to the text, where the Grail has disappeared before the appearance of ‘uns hons [...] ausi comme tous enflammés’ who does the wounding; but showing the man with the lance robed as a cleric requires additional explanation. He represents a pictorial conflation of the wounding scene and the immediately preceding episode, where the king sees a priest celebrating mass just before the appearance of the flaming man; however, in the text this priest and the man who does the wounding are not one and the same person. The Grail in the ‘Queste’ Most striking about the illustrations of the Queste in Rylands, Additional, and Royal is the omission of the Grail liturgy, which makes such an important visual impact in both earlier and later copies and which was depicted in such striking detail in the Estoire in Amsterdam and Royal (Pauphilet, pp. 269– 70). All three manuscripts also omit several of the seven associated episodes in which transformation or healing are the foci. The first such episode is the proto-liturgy: as a priest celebrates mass of the Holy Ghost, Galaad, Perceval, Boort, and Perceval’s sister see the miraculous White Stag transformed into Christ in Majesty, and the four accompanying lions become the symbols of the four Evangelists, the man, eagle, lion, and ox (Pauphilet, pp. 234–35). The stag and lions are depicted at this spot in Additional 10294 (f. 45v), but not their transformation nor the subsequent mass; the image is absent in Royal, and in Oxford there is a lacuna at this point. The other proto-Grail episodes are: Lancelot’s exclusion from a Grail liturgy (Pauphilet, p. 255), the bloodletting incident (Pauphilet, pp. 240–41), the healing of the Wounded King (Pauphilet, pp. 271–72) — for which the healing of King Arphasam, parts of which are shown in all three copies in Estoire, is a parallel — the 41 Reproduced in Matthews, p. 76. I have suggested in Stones, ‘Seeing’, n. 68, that the puzzling final scene in the opening miniature of Le Mans 354 might represent either this episode or the transmission of the Grail to Helein. 42 Blackman, Manuscripts, p. 508.

THE GRAIL IN MS RYLANDS FRENCH 1

517

final Grail liturgy, and the death of Galaad before the Grail (Pauphilet, p. 277). The healing of a cripple by the Grail and the carrying of the Grail table at Sarraz are included in all three copies (Additional, f. 138v; Royal, f. 211v; Rylands, f. 52), but none of the pictures show the Grail. In its first appearance in the Queste, the Grail comes before King Arthur and his knights as they sit at table. Emitting marvellous odours, it miraculously refreshes them with whatever food (‘de tel viande’) each knight desires (Pauphilet, p. 15). This is one of the best-known scenes in Arthurian illustration, because the extremely striking images in the fifteenth-century copies depict the Round Table as annular and display the Grail hovering at its centre.43 In the two versions commissioned by Jacques d’Armagnac (BNF fr 112, vol. II, f. 5 (c. 1470) and BNF fr 116, f. 610v (c. 1475), Fig. 21), Arthur and the knights sit at the Round Table, contemplating the Grail which appears in the open circular space at the centre of the Table. Shown as a ciborium, it is surrounded by rays of light and supported by two angels; this is also how the scene is depicted in Jean de Berry’s copy of ca. 1405 (BNF fr 120, f. 524) and the related Arsenal manuscript. The other manuscript in which this is a prominent scene is that made in 1319 in Avignon (Florence, Laurenziana, MS Ash. 121 [48], f. 6).44 Here the Grail, without a Grail-bearer, and depicted as a veiled chalice, is poised on a rod (the Holy Lance?) high above the heads of the diners, who (without Queen Guinevere) sit behind a rectangular table.45 Otherwise this scene is widely omitted altogether.46 Rylands (f. 184v, Fig. 22) and Royal (f. 91v) both include this scene, which is omitted in Additional. The Royal version is simpler, showing the Grail-bearer on the left, approaching King Arthur and three courtiers at table; in the Rylands version, the right-hand part of the miniature is similar in composition, showing the Grail-bearer facing three courtiers at table; but they are watched on the left by Queen Guinevere and King Arthur, who gestures in surprise as he sees the Grail. Queen Guinevere’s presence is notable, and the importance of the royal couple is further emphasized by the 43 See Baumgartner, ‘La couronne et le cercle’; Matthews, pp. 81, 84; Blackman, ‘Pictorial synopsis’, p. 38. 44 See Breillat, ‘La Quête du Saint-Graal en Italie’, pp. 296–300; Stones, ‘BN fr 95 and Yale 229’, pp. 213–19, and Walters, ‘Wonders and illuminations’. 45 There may be an analogy here with tabernacles suspended above altars on pulleys: see Stones, ‘Seeing’, n. 89, citing Rubin, Corpus Christi, p. 47. 46 Stones, ‘Seeing’, n. 87.

518

tall arch with turret and pinnacles that projects above them into the upper margin. Guinevere is not mentioned at this point in the text, nor in the preceding episodes at the beginning of the Queste where King Arthur and his knights are at table (Pauphilet, pp. 1–15). To my knowledge, Rylands is the only copy to include her. One might therefore speculate as to whether her presence might in some way be a reflection of the interests of a female patron.47 The Grail in Rylands (Fig. 22) is borne by a figure who stands on the near side of the table, a monk or cleric encircled by a wavy cloud motif — indicating that although we can see him, to Arthur and his court he is invisible. In contrast, the Grail that he holds out before the assembled company lies beyond the perimeter of the cloud and is thus seen by those at the table: a phenomemon suggested by the text (‘mes il n’i ot onques nul qui poïst veoir qui le portoit’ (Pauphilet, p. 15)). The cloud motif also refers back visually to the opening scenes in Estoire where, in these two manuscripts and also in Additional, Christ is surrounded by a cloud as he hands the book to the hermit so that, again, he is visible to the viewer but not to the hermit. In Royal, the monk/cleric holds the vessel closer to his body, so that it is within the cloud, and not as clearly visible therefore to Arthur and his company. This is a case where the Rylands version is preferable. The Grail is painted gold, and shaped in Rylands as a chalice and in Royal as a veiled vessel with a pointed top — no doubt a ciborium. Both represent a notable choice of shape for the Grail. In Rylands/Amsterdam the chalice form represents a rejection of the bowl which had up to this point been the preferred choice, used consistently in all the depictions of the Grail in Estoire, even in the last two scenes where Royal and Additional both already rejected the ‘escuele’ in favour of the ciborium. It may be significant that in Rylands this is one of several scenes which are the work of a second painter, whose work does not reappear in Royal and Additional but may in some way be dependent on them, probably adopting a Grail-as-chalice motif 47 There is so far nothing conclusive about the original patron, a question which further work may help to solve. An analogous situation in Christian art for the inclusion of a female figure in images made for women is that of the presence or absence of the Virgin Mary at scenes showing the Ascension and Pentecost. She is present, for instance, in the manuscript made ca. 1190–1210 for Ingeborg of Denmark, repudiated wife of King Philip Augustus (see Deuchler, Der Ingeborgpsalter and the facsimile, Der Ingeborg Psalter, and absent in the twelfth-century Cluny Lectionary, Paris, BNF nal 2246, f. 79v (see Cahn, Romanesque Manuscripts, no. 56, ill. 125).

THE GRAIL IN MS RYLANDS FRENCH 1

519

from the context of the liturgical and devotional illustration which formed the bulk of his extant commissions.48 However, it is likely this painter was a collaborator of the first artist because his participation is limited to a few scenes in the middle of Rylands, and his work is followed in the manuscript by the work of the main artist. Had he taken over from the main painter, for instance at the latter’s death, one would expect to find his work at the end of the manuscript. So the change of Grail motif from bowl to chalice may not simply be a question of a motif preferred by an earlier as opposed to that chosen by a later artist. Lancelot Sleeps while a Wounded Knight is Cured by the Grail Because of his sin of adultery, Lancelot is denied entry to the Grail chapel and sleeps outside by a cross, while a wounded knight is cured by the Grail (Pauphilet, p. 59).49 This episode is a second instance where a Grail-scene is depicted (again by two different painters) in Royal and Rylands, but omitted altogether in Additional.50 The setting of this miraculous cure, in both text and pictures, is described and depicted in these two manuscripts with considerable attention to detail. On the left in Royal (f. 90, Fig. 23) 48 I have commented on this artist in Stones, ‘Short note’, and ‘Another short note’; his work is characterized by the use of a somewhat broader, more rectangular, format than the almost square miniatures of his collaborator, and by the use of thin fillet-like frames with a row of dots in white; he also uses a three-white-dot motif on draperies, and draws outlines with a firmer line than his collaborator. I have suggested he may be traced in manuscripts associated with Ghent; his work also appears, I believe, in the St Stephen miniature in the Diurnal of the cathedral of Châlons-sur-Marne, now Chaumont, BM MS 29 (containing Easter Tables from 1317, the likely date of the manuscript, to 1425), and on f. 1 of the Apocalypse associated with Renaud de Bar, Bishop of Metz (d. 1316), Dresden, Sächsische Landes-und-Universitätsbibliothek, MS Oc 50. [His work can also be traced in Cambridge, Trinity B.11.22, Oxford, Bodl. Douce 6 (Stones, Gothic Manuscripts, Part I, vol. 2, Cat. nos. III–74, 77), the fragmentary Antiphoner formerly in the Weigel, Martin Le Roy and other Collections (Christie’s 16 Nov. 2011, lot 13; two leaves are now at the Palais des BeauxArts de Lille, see Jan Fabre, Illuminations/Enluminures, Cat. no. 29), the Breviary of St Peter, Blandin, Ghent, London, BL Add. 29253 and the Franciscan Psalter in Copenhagen, KB 3384, 8o (Carlvant, ‘Collaboration’). Thus this painter seems either to have migrated south at a certain stage in his career, or to have accepted commissions from quite far afield. 49 This scene in BNF, fr 342, Bonn 526, and Rylands are illustrated in Stones, ‘Sacred and profane’, pp. 100–12, figs 4–6; for further discussion see Stones, ‘Seeing’, nn. 87–97. 50 Again, Additional aligns with a considerable number of manuscripts where this scene is also omitted: see ibid.

520

is the chapel from which the six-branched candlestick with lighted candles (five in the miniature, and only four of them lit), the silver Grail Table, and the Grail itself have emerged into the open air. The Grail table and the Grail itself have been painted in silver on top of the gold used for the background and the silver has not adhered well, making the Grail and its surroundings difficult to decipher. It may be that this painting was done at a later date, or perhaps even at two later dates, as the right-hand side of the Grail table and the candlestick show differences from the left side of the table and may possibly be a later attempt to repair the flaking silver. It seems that the Grail is a ciborium whose lid has a pointed top and a scalloped rim. The wounded knight (wearing a knotted headscarf ), touches his eye with his left hand, indicating either that he is cured, or that he is seeing the Grail — or both (the paint has also in part worn off his face, making interpretation difficult). In the background is his litter, a huge mesh affair covered with an orange canopy. Lancelot is shown sitting on the step of the chapel, in front of the Grail table and the cross on the stone steps, his head resting on his folded arms, his eyes closed, asleep. He is wearing ailettes argent, a lion gules.51 In Rylands (f. 195v, Fig. 24), the episode is presented at a different moment, and the setting is slightly different as well. There is no hint of the architecture that makes the Royal version appear rather cluttered, and which is also found in one of the earlier versions, BNF fr 342, where the scene, though it contains no depiction of the litter or the altar (or indeed the Grail), is spread over two columns.52 In Rylands, the wounded knight lies on his litter, whose blue canopy is raised for us to see him. The litter is drawn by two horses, shown still moving as they pass the candlestick with its six flaming candles on the right and arrive at the Grail table on the left. The table is made of speckled marble and is set on a base of the same material. Its top is painted a plain brownish colour and the Grail — shown as a silver chalice whose paint is flaking off — hovers above it against a reredos also painted in silver against the gold background. The effect is similar to that of a sarcophagus with an open lid, partly because there is a cross on the table that looks as if it is inside a hollow space rather than on top of the surface of the table. Lancelot reclines on the ground in front of the altar, his head on his right arm which is supported by his helmet. His unusual shield 51 For some further examples of the use of heraldry in these manuscripts, see Stones, ‘Les débuts de l’héraldique’, reprinted in these essays. 52 See Stones, ‘Sacred and profane’, fig. 4.

THE GRAIL IN MS RYLANDS FRENCH 1

521

projects out into the margin. His arms, argent, a canton gules, two hearts of the same, are repeated on his surcoat and found only in the miniatures painted by the second hand in Rylands. These arms may provide a clue as to an owner or patron — or perhaps someone who acquired the book as it was in production, since they do not appear in the work of the first artist — but the shield still needs identifying.53 Also interesting about the presentation here is the lack of ‘seeing’ on the part of the actors — not only Lancelot, asleep, but also the wounded knight, who faces away from the Grail and has yet to be cured. Only the viewer, and apparently one of the horses, whose eye is fixed on the Grail, have the privilege of seeing it. What is striking about the scenes of the curing of the wounded knight is the variety and difference in emphasis given in each of the versions. Neither the Rylands nor the Royal version is like any of the earlier or later illustrations, either in their treatment of the Grail or in the general presentation of the scene or the other component elements. The Grail is omitted altogether in BNF fr 342 and in the Avignon copy of 1319 (Florence, Laur. Ash. 121 [48]); in Bonn 526 it is shown as a chalice with a cross inside, hovering above the sleeping Lancelot, shown without the wounded knight.54 In the special version made in Lombardy ca. 1380 (BNF, fr 343, f. 18), the Grail is shown as a gold chalice with a round cup and a polygonal base, but unlike Rylands, it has a paten, also in gold, covering the top. Lancelot sits off to the left, separated from the healed knight by the litter, placed in the centre of the composition, while the knight, now cured, kneels on the right in the chapel where the Grail is depicted on the altar.55 In Paris, Arsenal, 3479 and BNF fr 117 (f. l, Fig. 25), both made in Paris ca. 1405, this episode is 53 The heart motif is rare in medieval heraldry: a famous example is that of Sir James Douglas, Lord of Douglas († 1330), who, at the request of the dying King Robert the Bruce of Scotland, undertook to convey the heart of the Bruce to the Holy Sepulchre in atonement for his having failed to fulfil a vow of pilgrimage. See Wagner, Historic heraldry of Britain, p. 50, no. 32, pl. VII, but the Douglas shield also has a chief azure, three mullets or, so is unlikely to be relevant here. Another heart shield occurs in the Breviary of Saint-Laurent, Liège, now Darmstadt 394, referred to above in the context of the illustration of the mass, where the shield gules, three hearts argent is borne by one of the three sleeping soldiers at the Resurrection; but again, that is not what is shown here. 54 Stones, ‘Sacred and profane’, fig. 5. 55 Reproduced in Loomis and Loomis, Arthurian Legends, fig. 333. See also the brief catalogue entry in Avril, Dix siècles, pp. 98–9, no. 84. I thank Marie-Thérèse Gousset for her generous assistance with this manuscript. For the text, see Bogdanow, Version post-Vulgate. See now Pastoureau and Gousset, Lancelot du Lac.

522

given unprecedented prominence. Lifted out of its narrative context, it is the last of four key events in the life of Lancelot that are placed in a four-part composite miniature at the very beginning of the set of three volumes that transmit the five-part cycle complete; and the Grail is shown as a ciborium. Jacques d’Armagnac’s manuscripts (BNF fr 116, f. 621; BNF fr 112, vol. II, f. 15v, Fig. 26) both include this curing scene in the body of the text, elaborating upon the treatment of the Grail, which in these books has been transformed into a highly ornamented and bejewelled ciborium.56 Conclusion Of the ten scenes where the Grail is depicted in one or more of these manuscripts, only two are present, and include the Grail, in all three copies: the Crucifixion, and the handing of the Grail by Josephé on his deathbed to Helein. Of these, the Crucifixion is in a sense inappropriate in all three, since what should be shown is the Entombment; but the wealth of other associations makes this an interesting iconographical choice, one that is emphatically reinforced by its use across all three copies. Further, Royal and Additional align by including the two thieves, while Additional stands out for including the angel and devil and the souls of the thieves. In the scene of Josephé’s journey, Additional and Royal align by placing Josephé on the right, and the inclusion of the Grail also distinguishes Amsterdam from Additional and Royal. Additional and Royal both abandon the Grail as ‘escuele’ in favour of the ciborium for the two episodes of Helein’s prayer and Helein receiving the Grail from Josephé. Amsterdam’s treatment of the Grail is consistent throughout the Estoire; the second hand in the Rylands section, the artist responsible for the two Queste illustrations that show the Grail, prefers the chalice. In no case are the reasons for the change of vessel apparent. In terms of the density of illustration, Additional is weak compared with the other copies as it has fewer scenes overall, omitting altogether five scenes which the other copies include (two Grail liturgy scenes, the curing of King Blackman, Manuscripts, p. 23; ‘Pictorial Synopsis’, p. 39. In other instances the relationship is different: Elspeth Kennedy’s work on the illustrations of the Lancelot, for instance, suggests that Additional is more fully illustrated, and its images more carefully structured, than those in Amsterdam. This is the case in the False Guinevere episode, see now Stones, ‘Illustration et stratégie illustrative’, reprinted in these essays. 56 57

THE GRAIL IN MS RYLANDS FRENCH 1

523

Arphasam, and both Queste scenes),57 and it also omits the Grail at the wounding of King Arphasam. Amsterdam/Rylands omits Helein’s prayer before the Grail, although it includes the (non-Grail) fishing scene which is in neither of the other copies, and it also omits the healing of King Arphasam. Royal includes more Grail episodes than either of the other copies: the cure of King Arphasam is only in Royal, and Royal also has the prefatory scene introducing the liturgy of the Estoire, if without the Grail; but there are scenes like Josephé’s journey and the wounding of King Arphasam (both omit the Grail) and the Grail appearance at Arthur’s court (the Grail is within the cloud and so not visible to Arthur) where Royal’s version is less appropriate in its treatment than the version in Amsterdam/ Rylands. Other differences between Amsterdam/Rylands and Royal in relation to particular scenes of the liturgy and the curing of the wounded knight concern the moment in the episode that has been selected for representation; the setting, particularly in relation to architecture (elaborate for the liturgy in Amsterdam, but not for the healing of the wounded knight in Rylands); the actions and placing of the actors; and the type of vessel selected to represent the Grail. However, each version of these scenes contains elements that are not present in the other, so that it cannot be asserted that a particular scene in one of these manuscripts is dependent upon the corresponding scene in the other manuscript. Nor can either scene in either manuscript be seen as intrinsically preferable. In contrast to the close stylistic affinities which link the three copies Amsterdam, Royal and Additional, their alignment in relation to the Grail scenes is remarkably complex. There are all kinds of variants in selection and detail so that no single pattern of superiority emerges, nor would it appear that the manuscripts stand in an overall dependency relationship to each other. They are not simply models and copies, any one of any other. Nor do they relate in a consistent way either to earlier or to later versions of the Estoire and Queste illustrative traditions. It is hoped that further studies of other episodes in the Estoire and Queste, as well as the other branches of the Lancelot-Grail, will in due course help to elucidate further the web of interrelations among these copies as well as the wealth of other pictorial and textual associations on which their artists drew.

XXIII Signs and symbols in the Estoire del saint Graal and the Queste del saint Graal 1

T

he Lancelot-Grail romance in French prose holds a special place among vernacular texts and manuscripts of the Middle Ages. Surviving in whole or in part in some two hundred copies,2 most of them illustrated, the stories of King Arthur and his knights, their chivalric exploits and their quest for spiritual fulfilment, were appreciated by audiences of a variety of social classes and ages. Indeed, these romances address the human condition on multiple levels, where chivalric values of valour and prowess, loyalty and fidelity in friendship and love, take their place alongside treachery and betrayal, cowardice, deceit and adultery. In the telling of the adventures, the reality of contemporary society is coupled with allegory and symbolism in the Quest for the Holy Grail and the events and elements that surround the search for spiritual purity that the quest embodies. Two branches of the five-part cycle focus particularly on the Grail and share many elements in common. Among these is the episode of Solomon’s 1 This article is based on work done in collaboration with the late Elspeth Kennedy as part of the Lancelot-Grail Project. I am deeply indebted to Elspeth for her contributions to the project in general and to this study (see http://ltl22.exp.sis.pitt.edu/lancelot/ [now http:// www.lancelot-project.pitt.edu] WhatisLancelotGrail.htm); Stones, ‘Teaching and Research on the Web’ and ead., ‘The Lancelot-Graal Project’ both reproduced in these essays. This essay was first published in Signs and Symbols (Harlaxton Medieval Studies XVIII), ed. John Cherry and Ann Payne, Donington, Lincs., 2009, pp. 150–67. 2 The most complete list is still in Woledge, Bibliographie and Supplément, although several manuscripts have come to light since 1975, most notably the Turin and Bologna Estoire fragments, Bologna AS b.l bis, n.9 and Turin BN L.VI.12, see Meuwese, ‘De omzwervingen van enkele boodschappers en een jongleur’. A cutting from the Lancelot, Oxford, Bodl. Ash. 828, f. 7 was sold at Christie’s in 2017, and acquired by Bodley. For the text editions see the list at the end of these essays. La quête du Saint-Graal, roman en prose du XIIIe siècle, ed. Bogdanow, tr. Berne, based on Berkeley, UCB 73, appeared too late to be taken into account here.

SIGNS AND SYMBOLS IN THE ESTOIRE AND QUESTE

525

enchanted ship and its miraculous contents, whose discovery foreshadows and anticipates the mysteries of the Grail itself, bringing peril and danger to the unworthy as well as salvation to the chosen. The ship is discovered in the Estoire del saint Graal by Nascien, one of the first converts of Joseph of Arimathea and his son Josephé, the first Christian bishop according to the story. In the Queste del saint Graal the ship appears again, and there it is found by the three chosen Grail knights, Perceval, Boort and Galahad, and its meaning and that of its contents is interpreted by Perceval’s (nameless) sister, a seer who will become a Grail martyr, an allegory of the redemptive powers of death and sacrifice, leading up to the achieving of the Grail Quest by the three knights. How the manuscripts illustrate these episodes and what can be made of their similarities and differences are the focus of this paper. We select in particular the three copies that form the core of the LancelotGrail Project, a collaborative effort to interpret these copies, which were produced in the same scribal and artistic environment in the early years of the fourteenth century in northern France; two earlier copies of the Estoire and two Parisian stylistically related copies of the second or third decade of the fourteenth century; and some images from other manuscripts are included for comparative purposes.3 These manuscripts are referred to by abbreviation of place or collection.4 Many other manuscripts could eventually be drawn 3 Add. = London, British Library, Additional MSS 10292–4; Amsterdam = olim Amsterdam, Bibliotheca Philosophica Hermetica MS 1; BNF, fr = Paris, Bibliothèque nationale de France, français; Douce = Oxford, Bodleian Library, MS Douce 215; Le Mans = Le Mans, Médiathèque Louis Aragon, MS 354; Rennes = Rennes, Bibliothèque municipale MS 255; Rylands = Manchester, The John Rylands University Library, MS French 1; Royal = London, British library, Royal MS 14 E.III; Yale = New Haven, Yale University, Beinecke Library, MS 229. A comparative table of the events discussed here is to be found on http:// vrcoll.fa.pitt.edu/stones-www/lancelot-project.html [now http://www.lancelot-project.pitt. edu]. 4 For the date and place of these manuscripts see Stones, ‘Another Short Note’. The comparative approach to Lancelot-Grail iconography in tabular form was pioneered by Blackman, The Manuscripts and Patronage of Jacques d’Armagnac, Appendix, and ead., ‘Pictorial Synopsis’. For the three early fourteenth-century manuscripts see also Meuwese, ‘Three Illustrated Prose Lancelots from the same Atelier’ and Stones, ‘The Grail in Rylands French 1’ reprinted in these essays. To the late Elspeth Kennedy is due the realization of the importance of the placing of miniatures in relation to the text rather than a consideration of the similarity of pictorial subjects alone. What emerges is that differing principles of picturetext structure were used by the designers: whether the pictures serve to introduce an episode (if placed at the beginning of an episode) or to accompany a textual description (if placed immediately above or adjacent to the relevant text passage). Different strategies were chosen even among the manuscripts made in the same atelier.

526

upon for further comparison, some of which are mentioned in the notes, and eventually we hope to achieve fuller pictorial coverage of these and other manuscripts on our web site. 1.1. Estoire del saint Graal In the Estoire del saint Graal the episode of Nascien on Solomon’s ship is preceded and prepared for by the maritime adventures of King Mordrain, another of Joseph and Josephé’s early converts, who is miraculously transported to an island. There he is visited by Christ, who names himself ‘Tout-enTout’, and by the devil. They arrive in turn, in boats of different colours, and Tout-en-Tout assures Mordrain he will be saved by the sign of the cross. The narrative then turns to Nascien, transported from Calafer’s prison to the Turning Island from where he encounters Solomon’s ship, boards it and examines its contents, with dire results. Between the adventures of Mordrain and Nascien comes the abduction of Nascien’s son Celidoine, and at the end of the sequence all three heroes are reunited. After other adventures, the seacrossing of Joseph, Josephé and their followers to England, bearing the Holy Grail, concludes the maritime sequence. The manuscripts vary considerably as to which aspects of these events are selected for illustration and where in the text the pictures, or markers in the form of champie initials (gold letters against a party-coloured pink and blue background), are placed. Among the manuscripts compared here, Royal and Amsterdam are closest to each other in their positioning of eight markers for the sequence as a whole (not counting the Hippocrates story and the Tomb commission made by Nascien’s wife Flegentine which interrupt the sea-adventure episodes), eight champie initials as favoured in Amsterdam and eight miniatures in Royal; Rennes gives fewest illustrations, five historiated initials, four of them corresponding in position to Le Mans and Add. 10292, and one without parallel. Le Mans has five champies and two historiated initials, all corresponding in position to a miniature in the Additional manuscript. Add. 10292 has more markers than the other manuscripts: thirteen miniatures and two champie initials, allowing for a much more expansive pictorial treatment than in any of the other manuscripts. This is characteristic of the approach to pictures taken in general in Add. MSS 10292-4 which, so far as we know, is the copy that is the most densely illustrated of all the Lancelot-Grail manuscripts, boasting a total of 748 illustrations for the entire five-part cycle. But it must also be

SIGNS AND SYMBOLS IN THE ESTOIRE AND QUESTE

527

noted that, even in the sequence considered here, Add. 10292 omits markers at points where Royal and Amsterdam give a miniature or a champie, so the notion of what constitutes a ‘complete’ picture cycle is variable. Finally, the two Parisian manuscripts show considerable difference from each other and from the rest of the group, despite their textual similarity: BNF, fr 9123 has nine miniatures, three without placement parallels in other copies, six corresponding to a miniature elsewhere, while BNF, fr 105 has ten miniatures, six without parallels for placement and four corresponding in position to other copies. These differences are most likely indications that several different selection strategies were at work, even among copies made, as these were, in the same artistic environment or workshop, and that observation is further substantiated when the subjects of the illustrations are compared. However, certain scenes are marked in all copies: the opening of the episode (a little later in BNF, fr 105); Celidoine led away; Nascien at the Turning Island (a little later in Rennes); Nascien on Solomon’s Ship, the central precursor to the rediscovery of the Ship in the Queste. 1.2. King Mordrain on the Island The island locus is marked at the beginning of the sequence in Estoire by a champie initial in Le Mans and Amsterdam, and by an illustration in Rennes, Add. 10292, Royal, BNF, fr 9123, and a little after the beginning in BNF, fr 105. Mordrain’s miraculous journey — in his own splendidly designed bed, a unique rendering so far as I know — is the focus in Rennes (Plate 1), while his desolation at finding himself alone on the deserted island is what Add. and Royal have chosen to emphasize by showing him sitting on the ground by a cave on the rock, head on hand. In fr 9123 and fr 105 Touten-Tout’s appearance in his white boat, followed by the devil in the form of a horned woman in a black boat are immediately introduced, whereas the other manuscripts place these appearances later in the text. Add. 10292 gives three illustrations, framing Tout-en-Tout and his silver boat with its white sail and red cross (Plate 2)5 by a double depiction of the devil’s black boat which appears before and after the white boat, thus emphasizing pictorially Mordrain’s temptation by the devil, whose boat’s second appearance has the 5

‘Cele neif estoit petite mais ele estoit a meruellles bele car li mas estoit ausi blans comme flor de lis et desus en haut avoit vne crois vermeille...’ (Add. 10292, f. 24). The colour of the sail is not actually mentioned, unless it was assumed to be the same as the colour of the mast.

528

shields of Mordrain and Nascien on its mast, as the text describes. In the text, the white boat also makes a second appearance that is not illustrated. Rennes and Royal each give a single illustration, choosing Tout-en-Tout and omitting the devil, thus focussing on Mordrain’s eventual triumph over the forces of evil. Rennes shows Tout-en-Tout in his boat with a white sail (but no cross), positioning the miniature later than Royal but earlier than Tout-en-Tout’s textual appearance. Royal positions its miniature at the very beginning of the boat sequence, emphasizing still more the eventual outcome where Mordrain resists the devil and triumphs thanks to the sign of the cross on the sail of Tout-en-Tout’s boat. Most remarkably, Tout-en-Tout’s boat in Royal is shown with a red sail (Plate 3) — an important pictorial variant compared with what is depicted in Rennes and Add. 10292. It depends on a textual variant in Royal, where the colour of the cross has been omitted and the sail itself is described as red, ‘Chele nef estoit petite et toute dargent et si estoit li mas dor et li voiles estoit ausi tous viermaus ... Et quant il vit el voile le signe de la sainte crois si fu auques asseures’ (Royal f. 36) — and that variant is what the artist has followed, a rare instance in this cluster of manuscripts, proving that whoever specified the subjects of the miniatures (designer, artist, patron?) paid very close attention to what the particular variant in Royal actually says. The picture was not simply copied from an illustrated model. Such examples are relatively rare in general across the illustrative tradition of the Estoire del saint Graal, and a full study of what all the other illustrated copies do at this point in the text, and what their textual variants are, is beyond the scope of this article.6 However, BNF, fr 105 and 9123, both illustrated by an artist I call the Sub-Fauvel Master, are two further cases where another textual variant has in part been followed in a picture.7 Both 6 The Mordrain episode is not always illustrated: there is no picture in this section in BNF, fr 95, for instance, although the related Queste, Yale 229, gives copious illustrations to the Solomon’s Ship episode, as I show below. 7 BNF, fr 9123 is a collaborative effort divided between the Sub-Fauvel Master and another Parisian artist known as the Maubeuge Master on the basis of his work in the Grandes chroniques de France, Paris, BNF, fr 10132, sold in 1318 by the Parisian libraire Thomas de Maubeuge. See Stones, ‘The Artistic Context of le Roman de Fauvel’; Rouse and Rouse, Manuscripts and their Makers, I pp. 185–87, 213–15, 248, 373 n. 11, 380 n. 89, 381 n. 103, 391 n. 99, 392 n. 122, Appendices 7F, 7M, 8D, 9A, ills. 105, 107, 108. Unusually, the text of BNF, fr 9123 is prefaced by a numbered list of rubrics which correlate with the rubrics accompanying the miniatures. There is no such list in BNF, fr 105. Rouse and Rouse, I, p. 106, note the different treatments of Vortiger’s election to the throne; and there are many

SIGNS AND SYMBOLS IN THE ESTOIRE AND QUESTE

529

manuscripts give the same textual variant (not the same variant as in Royal): BNF, fr 105 on f. 47v says ‘cele nef estoit petite toute dargent et si estoit li mas dor et le voile et le voile sic estoit ausi blanc comme noiz negiee et si auoit el mi lieu de la nef vne grant crois toute uermeille...’ and the rubric preceding the miniature describes the boat and its contents in similar terms, ‘Comment li rois Mordrains vit venir dune roche de mer ou il estoit vn moult bel homme en vne nef tout dargent et auoit enmi lieu de cele nef vne crois tout vermeille.’ In the picture, the boat is painted silver, its sail white, its mast gold, and a red cross stands in the boat towards the prow. In BNF, fr 9123, the text on f. 40 is substantially the same as in BNF, fr 105 (or vice-versa, since it unclear which, if either, came first or indeed whether one depends on the other): ‘Celle nef etsoit moult petite toute dargent et li mez si estoit dor et li voiles estoit aussi blans comme noif negiee et si auoit el milieu de la nef vne grant crois toute vermeille...’ (f. 40). There are two illustrations for this sequence, on ff. 38 and 4lv, whose rubrics, respectively, say ‘Comment li roys mordrains vit venir parmi la mer vne nef dargent a gouuernal dor .i. homme dedens et ou mileu de la nef auoit vne crois vermeille Et comment elle ariua a la roche ou li roys estoit’ (f. 38, numbered xviij, Plate 4), and ‘Comme li roys estoit estoit sic en orisons a nostre seigneur et comment la bene nef dargent et lome dedens reuint a lui’ (f. 41v, numbered xix). But the illustrations both show Tout-en-Tout handing a large gold cross to Mordrain, an action corresponding neither to text, rubric, nor to any other picture of this episode so far as I know, and unexplained. Both manuscripts continue the sequence by depicting the devil and his boat, but differently: whereas BNF, fr 105 on f. 53 gives three devils in a boat with shields on the side approaching Mordrain, the devil in BNF, fr 9123 on f. 43v appears in the guise of a horned woman. Furthermore the pictures are differently placed in the text, and their treatment unrelated to what is in Add., the only one of the Add.-Royal-Amsterdam group to include the devil’s boat. At the least, these examples suggest something of the complex and changing nature of text-picture relationships for this episode, to the point that even the same (or closely related) artists were quite capable of producing a different picture if a text variant required it — or even if one did not.

more examples both of similarity and of difference that remain to be explored. See now ‘L’Estoire del saint Graal dans la version adaptée par Guillaume de la Pierre pour Jean-Louis de Savoie, évêque de Genève’, reprinted in these essays.

530

1. Rennes, BM 255, L’Estoire del saint Graal, f. 30v, King Mordrain transported in his bed (photo: Lancelot-Grail Project)

2. London, BL Add. 10292, L’Estoire del saint Graal, f. 25, King Mordrain and Tout-en-Tout by the enchanted ship with white sail and red cross (photo: British Library)

SIGNS AND SYMBOLS IN THE ESTOIRE AND QUESTE

531

3. London, BL Royal 14 E.III, L’Estoire del saint Graal, f. 36, King Mordrain and Tout-enTout by the enchanted ship with red sail (photo: British Library)

4. Paris, BnF fr 9123, L’Estoire del saint Graal, f. 38, King Mordrain sees Tout-en-Tout arrive by boat, carrying a cross (photo: Bibliothèque nationale de France)

532

5. London, BL Add. 10292, L’Estoire del saint Graal, f. 31, Nascien on Solomon’s enchanted ship sees the bed, crown, and sword, and three spindles (photo: British Library)

6. Le Mans, MM 354, L’Estoire del saint Graal, f. 110, Nascien on Solomon’s enchanted ship sees the bed; a soldier terminal; a cleric terminal reading (photo: Lancelot-Grail Project)

SIGNS AND SYMBOLS IN THE ESTOIRE AND QUESTE

533

7. Rennes, BM 255, L’Estoire del saint Graal, f. 46, Nascien on Solomon’s enchanted ship sees the bed and spindles; followers watch from the shore (photo: Lancelot-Grail Project)

8. London, BL Add. 10292, L’Estoire del saint Graal, f. 31v, Fall of Adam and Eve: serpent with female head, Adam clutches his throat, Eve holds a branch (photo: British Library)

534

9. Paris, BnF fr 105, L’Estoire del saint Graal, f. 65v, Cain kills Abel with a spade, beneath the red tree (photo: Bibliothèque nationale de France)

10. Paris, BnF fr 9123, L’Estoire del saint Graal, f. 53v, Cain kills Abel with a spade beneath the green tree; Eve holds a green tree (photo: Bibliothèque nationale de France)

SIGNS AND SYMBOLS IN THE ESTOIRE AND QUESTE

535

11. Paris, BnF fr 19162, L’Estoire del saint Graal, f. 65, The Tree changing from white to green; Eve churns butter (photo: Bibliothèque nationale de France)

12a. Berkeley, University of California, Bancroft Library 106, L’Estoire del saint Graal, f. 145v, Cain kills Abel with a spade beneath the tree; three trees, white, green, red (photo: University of California, Berkeley)

12b. Berkeley, University of California, Bancroft Library 106, L’Estoire del saint Graal, f. 145v, M initial, three figures before a three-branch green tree and plank (photo: University of California, Berkeley)

536

13. Hildesheim, St Godehard Treasury, St Albans Psalter, p. 215, Psalm 118: 33, two women before two men; three trees (photo: Hildesheim, St Godehard)

14. London, BL Add. 10292, L’Estoire del saint Graal, f. 34, Solomon and his Queen order the construction of the Ship (photo: British Library)

SIGNS AND SYMBOLS IN THE ESTOIRE AND QUESTE

537

15. London, BL Add. 10292, L’Estoire del saint Graal, f. 35v, Nascien swims ashore as Solomon’s ship collapses (photo: British Library)

16. London, BL Royal 14 E.III, La Queste del saint Graal, f. 125v, The Grail Knights and Perceval’s sister approach Solomon’s ship with spindles, sword, crown, and bed (photo: British Library)

538

17. New Haven, Yale University, Beinecke Library 229, La Queste del saint Graal, f. 253, Adam and Eve eat apples; Sacrifice of Cain and Abel (Arthurian Romances. General Collection, Beinecke Rare Book and Manuscript Library, Yale University)

18. London, BL Royal 14 E.III, La Queste del saint Graal, f. 125v, God tells Adam and Eve to make love beneath the white and green Tree (photo: British Library)

SIGNS AND SYMBOLS IN THE ESTOIRE AND QUESTE

539

19. Oxford, Bodleian Library, MS Douce 215, La Queste del saint Graal, f. 35, The Three Grail Knights on Solomon’s Ship read and hear the explanation (photo: Bodleian Library)

20. London, BL Royal 14 E.III, La Queste del saint Graal, f. 130v, The Three Grail Knights and Perceval’s sister on Solomon’s Ship read and hear the explanation (photo: British Library)

540 21. New Haven, Yale University, Beinecke Library 229, La Queste del saint Graal, f. 257v, Perceval’s sister hands David’s sword to Galaad on Solomon’s Ship (Arthurian Romances. General Collection, Beinecke Rare Book and Manuscript Library, Yale University)

22. London, BL Add. 10294, La Queste del saint Graal, f. 46v, Perceval’s sister receives communion on her deathbed (photo: British Library)

SIGNS AND SYMBOLS IN THE ESTOIRE AND QUESTE

541

23. London, BL Add. 10294, La Queste del saint Graal, f. 47v, Lancelot reads the letter identifying the dead damsel as Perceval’s sister (photo: British Library)

24. New Haven, Yale University, Beinecke Library 229, La Queste del saint Graal, f. 262v, Lancelot reads a scroll identifying the dead damsel as Perceval’s sister; he brings her and the boat to a hermit on shore (Arthurian Romances. General Collection, Beinecke Rare Book and Manuscript Library, Yale University)

542

1.3. Nascien and Solomon’s Ship Next comes Nascien’s discovery of Solomon’s Ship. In all seven manuscripts the sequence begins with Nascien in Calafer’s prison, marked in Amsterdam by a champie initial and in the others by an illustration, where Rennes (in part), Add. and Royal, and BNF, fr 105, show Nascien released from prison by a hand seizing him (in the manner of Habakkuk) by the hair. Le Mans shows a slightly earlier moment, when he is thrust into the prison, while Rennes couples the release with Nascien and his son Celidoine both in prison before Nascien’s release; and BNF, fr 9123 shows Nascien in a cloud and flames descending upon Calafer. A second illustration (a champie in Le Mans and Amsterdam) is given in all seven manuscripts, this time with a focus on Nascien’s wife Flegentine, including their son Celidoine again in Rennes and Additional 10292; and BNF, fr 105 has an additional scene focussing on Celidoine. The Turning Island is marked next in all seven manuscripts (again by a champie initial in Le Mans and Amsterdam), at a slightly later place in Rennes where Nascien approaches the island on horseback, while in Add. and Royal he lies there dreaming of birds.8 At this point, Add. again gives a sequence of illustrations that are without parallel in the other copies, framing the sequence with Nascien on Solomon’s Ship at the beginning and end, and marking the history of the three spindles with three explanatory miniatures and two champie initials in between. The first miniature depicts Nascien on the Ship, finding (with hands raised in surprise) its contents of bed with crown and sword, and the three spindles of different colours, white, green and red (Plate 5). Their meaning is critical not only to the story in Estoire but also to the reappearance of the ship in the Queste, where it marks an important stage in the Grail Knights’ understanding of the allegory of the Quest for the Holy Grail in which they are participants. Yet only a few of the illustrations depict the contents of the Ship and the events surrounding Nascien’s discovery. In the end, Nascien’s curiosity leads to disaster as the ship collapses and he is forced to swim to shore. For the Ship was not intended for him, but for the pure knight yet to come — who in the Queste del saint Graal would achieve the Grail Quest 8

What makes the island turn and where does the motif come from? No hints are given in the text, and there is no mention of this question in Szkilnik, L’archipel du Graal. It surely reflects an aspect of scientific knowledge such as the principles of magnetism, and remains to be explored. See Séguy, Le Livre-monde, 67–71, citing the Cosmographia of Bernard Silvestris.

SIGNS AND SYMBOLS IN THE ESTOIRE AND QUESTE

543

along with his companions. Additional 10292 is again alone in depicting the untoward consequence of Nascien’s curiosity (Plate 15). Earlier in the text, Add. depicts an episode where another transgression of Nascien is punished: he presumes to look into the Grail, and is blinded as a result. What is depicted in Additional 10292 (f. 20v, and also in Royal, f. 30v) is not actually the blinding of Nascien, nor his subsequent curing, but rather the punishment meted out to Josephé for not having prevented it — Josephé is pierced in the groin with a lance aimed by an angel, a punishment endured for the second time. Josephé had earlier failed to baptise everyone and was lanced in the groin for that.9 Josephé is subsequently cured, as was Nascien.10 What is not shown in any of the copies considered here is a later episode where Nascien and Mordrain are again on Solomon’s ship and Nascien presumes to draw the sword, whereupon it breaks and he is inflicted with a wound in the shoulder by another sword, in punishment. King Mordrain reunites the broken sword and returns it to its sheath. The other six copies all show an aspect of Nascien’s initial discovery and some of the contents of the ship, the subject of Add.’s first miniature in the sequence (except for Amsterdam which has a champie initial). In Additional 10292 (Plate 5), all the elements are present in both pictures: the bed with crown (painted in silver or tin), sword and sheath, and the three spindles, two at the front of the bed and one at the back, and the ship has an inscription on the side, reading ‘Saches iou ne sui foy non et creanche.’ The treatment of the ship in Royal is similar — bed, crown, sword and spindles (all on the near side of the ship) are also present, and there is an inscription on the side of the ship, less easy to read, ‘iou ne ai ie my ... iou sui se fe ... non.’ These manuscripts are the ones which most faithfully depict the symbolic contents of the ship, whose meaning is elucidated in the text but only partially explained in the pictures, and nowhere, to my knowledge, as explicitly depicted as in Add. 10292. In Royal, Nascien, somewhat incongruously holding his gloves, contemplates the boat from the shore to which he has just swum — and the ship, unlike Add. 10292’s final version, is still intact. Rennes and Le Mans each give a less detailed picture of the ship: in Rennes, Nascien looks at the bed with its three spindles which form three comer-posts but the crown and sword are not shown, and companions watch from the shore (Plate 7); in Le Mans, Nascien is alone on the ship 9

At S I, p. 77/Pon p. 159; no illustration either in Add. or in Royal. The second lancing is at S 1 p. 80.17/Pon p. 165.

544

(apart from a knight terminal and a cleric terminal above and below) and the bed is present but the crown, sword, and spindles are absent (Plate 6). BNF, fr 105 and fr 9123 are both cursory in their treatment of the ship. In fr 105 the boat is shown twice, the first time with the bed on which is the sword and its decorated sheath and pommel on f. 60, but without the crown and spindle, but shows only the bed on the boat on f. 70; fr 9123 gives a return of Tout-en-Tout and omits bed, crown, sword and spindles. 1.4. The Three Spindles The intermediary images in Add. 10292, and BNF, fr 105 and fr 9123, which are similar in content but different in placing, depict highlights of the story of the spindles, grown from the branch taken by Eve at the Explusion from Paradise, and the story is repeated in the Queste, with several differences in emphasis, as I show below. The miniatures in Add. begin with the Fall, where the serpent with a female head is coiled around the tree; Adam clutches his throat, having eaten the apple, while Eve holds in her hand a leafy branch, possibly (the image is indistinct) still with an apple attached (Plate 8). In the text this branch, planted outside Paradise, grows into a tree whose wood is white. Depicted in Royal is the sequel, where God commands Adam and Eve to make love beneath the tree, whereupon its wood turns from white to green. Beneath the green tree Cain murders Abel and then the tree turns to red, and offshoots of the tree at each stage retain their original colour, but Add. 10292’s trees are all shown as green. The murder in Add. 10292 is achieved with the jawbone of an ass,11 whereas in BNF, fr 105 and fr 9123, 10 The curing of Nascien (but not his wounding) is depicted in Bonn, UB, 526, f. 16v. There are obvious parallels with cures effected by relics, such as the near-contemporary case of the future Louis VIII, son of Philippe II and Isabelle, who was cured of dysentery by the touch of relics brought from Saint-Denis: a nail, a thorn, and an arm of St Simeon. On the same day Philippe, abroad and suffering from the same illness, was also cured. This is reported in the historical portion of Yves de Saint-Denis’ work on the Life and Miracles of St Denis, found in Paris BNF, lat 5286, ff. 202v–203 and lat 13836, ff. 96–97v, and printed by Duchesne in Historiae Francorum scriptores coaetanei, V, pp. 288–89. These references were kindly drawn to my attention by E.A.R. Brown. 11 Several scholars have investigated the origins of the jawbone motif. See Schapiro, ‘Cain’s Jawbone’. The oldest instance cited by Schapiro is in Aelfric’s Heptateuch, where the text is silent about the instrument but the picture shows a jawbone; and the earliest literary reference Schapiro brings to bear on the subject is the Anglo-Saxon prose Solomon and Saturn, where the jawbone is called a cinbân (citing Kemble, The Dialogue of Solomon and Saturn p. 187), which Schapiro claims would have sounded similar to ‘Cain bana’ (where ‘bana’ means

SIGNS AND SYMBOLS IN THE ESTOIRE AND QUESTE

545

the instrument is a spade, and the deed is shown taking place beneath a red tree (Plate 9); in BNF, fr 9123 the miniature also includes Eve holding the green tree (Plate 10). Some other manuscripts also focus on the changes in the tree: thus in Paris, BNF, fr 19162, the miniature on f. 65 shows the tree in the process of changing colour from white to green, while on the border Eve stands working a butter-churn (Plate 11); in Berkeley, UCB, 106, a two-part miniature shows Cain killing Abel with a spade, with the three trees painted red white and green below (Plate 12a); and Paris, BNF, fr 95, f. 49v has a historiated initial containing three trees, the middle one with a double blue trunk, the outer two with red trunks, but all with green leaves.12 One fifteenth-century unillustrated manuscript, Paris, BNF, fr 1426, left a large space on f. 158v preceding the passage starting ‘Moult dura longuemen ceste arbre...’ on f. 159, and a note reads ‘Cy doit estre larbre de vie paint en /quatre [sic] coulleurs.’ And a drawing of three trees is also found in an otherwise unillustrated English thirteenth-century copy of Estoire, London, BL, Add. MS 32125 (f. 205v), at the same place in the text. murderer). Henderson, however, points out that differences in pronounciation and other linguistic inconsistencies would have made such a confusion unlikely (Henderson, ‘Cain’s Jaw-Bone’). Henderson agrees with Schapiro that the Aelfric representation is the earliest, and further suggests that the motif may have transferred from Solomon, citing Emerson, ‘Legends of Cain’, p. 859, referencing Judges 15: 15; or, just possibly, to have been borrowed from a plough coulter as on tenth-century Irish crosses. Barb, drawing upon a neglected article by Scheiber, ‘Kajin’s Abel-’ldozati fiistjérol szóló legenda életrajza’, cites ancient Egyptian evidence for the use of flints inserted into the animal’s jawbone — and later, into wood — to form a kind of sickle (Barb, ‘Cain’s Murder-Weapon’). For the transition from ‘primitive jawbone tool’ to not only the toothed sickel and the serrated saw but a fearful weapon and status symbol with a venerable history, Barb further cites Makkai, ‘Early Near Eastern and South East European Gods’. The same configuration, with a backturned Abel, and the same instrument, are also found in the marginal scene in a devotional book illustrated by the Add./ Royal/Amsterdam artist, the Hours of Liège (not Maastricht) Use, London, BL Stowe 17. 12 In BNF, fr 19162, Berkeley, UCB 106 and BNF, fr 95 these images are placed at S I p. 130. 6, Pon vol. II, pp. 279, 445, where Add. gives a champie initial; the place is unmarked in the other four manuscripts. BNF, fr 19162 and its twin, Paris, BNF, fr 24594 are stylistically related to manuscripts associated with Cambrai, Saint-Omer and Thérouanne: Bonn, UB 526 (written in Amiens in 1286, giving a miniature of Solomon’s Ship with only the bier and the spindles) and Paris, BNF, fr 110, two missals of Cambrai, MM 153 and 154; the Guillaume d’Orange cycle manuscript written in 1295, Boulogne-sur-Mer, BM 192; the Vincent of Beauvais, Speculum historiale, Boulogne, BM, 131; and the partial bible, SaintOmer, BM, 5 and many other related manuscripts. BNF, fr 95 is particularly close stylistically to the Psalter-Hours of Thérouanne use, Paris, BNF, lat 1076 and Marseille, BM 111 so

546

No direct reference is made to the Crucifixion either in text or picture, although the presence of Nascien as the discoverer of the ship anticipates the achieving of the Grail Quest by his descendant Galaad, while the descent of Christ from the lineage of Solomon through the Virgin Mary, the new Eve, is hinted at in text and picture, but not elaborated upon. Galaad is in effect a type of Christ, though the link is never made explicit. Similarly the link between Solomon and the wood of the trees grown from Eve’s branch may also allude indirectly to the legends about the Wood of the Cross of Christ, grown from seeds obtained from Paradise by Seth which produced three shoots of different trees, cedar, cypress, and pine (no mention of colours) that were discovered by Moses. The three shoots then grew together to form a single tree, which Solomon tried to use for building the Temple (but the planks were always too long or too short) and which eventually served to make the Cross of Christ. Nothing is directly said in the text of the Estoire to link the three spindles and their colours to these Cross legends, however. Conversely, so far as I know, there is no reference in any of the Cross legends to the coloured trees of Estoire and Queste nor to the spindles made from them, and the Cross legends are unillustrated apart from the Middle High German Lutwin version, Vienna, ÖNB 2980,13 and the border illustrations in the Hours of Catherine of Cleves, New York, Morgan M.917/945 (reproduced on Corsair). emanates from the same diocese in the decade of the 1290s; and it is closely related to Yale 229, whose Queste del saint Graal I consider below. See Stones, ‘The Illustrations in BN fr 95 and Yale 229’, reprinted in these essays. Berkeley, UCB, 106 is the odd-one out: in a personal communication F. Avril has suggested it is part of a cluster of manuscripts, mostly bibles, associated with Jumièges and other Norman monasteries in the 1270s and 1280s. Given the possibility of monastic production and/or patronage it is interesting to note that in UCB 106 the Estoire is preceded by the Vies des pères and other material (including Gautier de Coinci’s Theophilus Miracle), illustrated by the same artist(s), although the Estoire is written in two columns whereas the Vies des pères is in three, and the two parts need not always have been together. For further analysis see Gehrke, Saints and Scribes. 13 The early literature on the Cross legends is summarized in Napier, History of the Holy Rood-Tree and in Seymour, The Cross in Tradition, History and Art. More recent studies include Shields, ‘Le bois de la croix’; Prangsma-Hajenius, La Légende du Bois de la Croix; Baert, A Heritage of Holy Wood and now ead., ‘Adam, Seth and Jerusalem’. To the best of my knowledge, the only manuscript to couple Cross-Legend material with the Estoire is Paris, BNF, fr 95 which opens with Estoire, Merlin, and Suite vulgate, followed by Sept Sages and Pénitence Adam. For the latter see Quinn, The Penitence of Adam. However, Quinn observes that BNF, fr 95 transmits the Class II version of Pénitence Adam in which the Class III interpolation of a rood-tree legend is absent. Quinn refers to the Lutwin version, ÖNB 2980, on p. 51.

SIGNS AND SYMBOLS IN THE ESTOIRE AND QUESTE

547

There is one exceptional depiction of three trees coloured white, green, and red, which may allude to the Cross legends and also, because of the colours, to anticipate the trees of the Estoire and Queste. These trees are found in the historiated initial to Psalm 118, 33 in the early twelfth-century Psalter of Christina of Markyate (the St Albans Psalter, Hildesheim, Pfarrkirche St Godehard, p. 313, Plate 13).14 They are placed in a rectangular panel at the bottom of the initial, by a stream outside a structure within which, on the left, a woman holds a leafy branch and a small round object, and faces a man holding a bird on his wrist; a second pair of figures are another man holding a small round gold object grasping a woman (the same woman?) by the hand, as she gestures at him, raising the fingers of her left hand.15 Geddes interprets this as a depiction of Christina subjected to the temptations of cupidity and lust, based on the commentaries of Ambrose and Augustine.16 The trees, with their Paradise-Salvation connotations, would add another level of meaning, suggesting that Christina would withstand these temptations.17 The question remains as to what sources the St Albans image — and the Estoire and Queste — are drawing upon. The likely date of the St Albans Psalter in the second decade of the twelfth century is a sure indication that some, at least, of the Cross legends, were known in England and at St Albans. One Estoire manuscript has a hint of what might be an allusion to the Cross legends: the Cain-Abel-Trees miniature in UCB106 is followed by an initial M (Plate 12b) containing a three-part green tree behind which may be seen a vertical brown plank, and three figures in front above a wavy grey motif in the foreground (a boat? water?). Might the vertical plank be an allusion to the Cross and its legends? and might the three figures refer forwards to the three Grail Knights of the Queste? If so, the reference is discrete at best, and the manuscript in its present state does not include a Queste del saint Graal. 14 I thank Jane Geddes for drawing my attention to this image and for her views on its significance. See Geddes, The St Albans Psalter, Plate 82. 15 Interpretations of the round objects are given in Haney, The St. Albans Psalter, p. 592 n. 800, citing Goldschmidt, Der Albanipsalter, p. 122: an egg; Pächt, Wormald, Dodwell, The St Albans, p. 251: an apple (pl. 77); for Haney, the gold object is a gold coin. She further notes that the trees and water are lacking in the Stuttgart and Utrecht Psalters. For the flowering branch and round objects as Marian attributes see Heslop, ‘The Romanesque Seal of Worcester Cathedral’. 16 Geddes (see above, n. 14), p. 99. 17 Geddes (see above, n. 14), p. 105 n. 15.

548

The colour triad of white, green and red no doubt resonated for medieval audiences with additional symbolism not made explicit either in text or rubric in the Estoire, nor indeed in the Queste. Abundant exegetical literature linked these colours with the precious stones of Exodus and the Apocalypse, the elements, the cardinal virtues, and the colours of liturgical vestments.18 White, red, and green were even associated with the French monarchy, as by the fourteenth century, threads in these colours were used to attach seals to royal charters.19 1.5. Solomon, his Queen, and the Making and Fitting of the Ship What is shown next in Add. 10292 is the making of the ship by carpenters in the presence of King Solomon and his Queen (not named in the text) (Plate 14). It is she, not Solomon, who chooses the objects for the boat and explains for whom they are destined — so she is presented as a parallel for Perceval’s sister in the Queste. The spindles and the crown, sword and sheath had already been depicted as seen by Nascien and would be shown again in Add. 10292’s concluding miniature (Plate 15). Much is made in the text of the symbolism of these objects — the spindles made from the wood of the three trees associated with Adam and Eve and Cain and Abel, while the crown and sword are David’s and had been kept in the temple. Solomon was to fit out the sword with a new pommel of precious jewels, and a new sheath;20 Solomon’s Queen would take 18 Useful references are in Kirschbaum, Lexikon der christlichen Ikonographie under ‘Farbensymbolik’, cols 7–10. For instance, Bede, Rupert of Deutz, Honorius Augustodunensis all link the four elements to the colours purpureus, hyacinthus, coccus and byssus. 19 Many examples were on view at the Archives Nationales in Spring 2007.1 thank Hervé Pinoteau for drawing this important feature to my attention. See Pinoteau, ‘Deux triades de couleurs’. 20 ‘El temple que vous avez fait en lonor ihesu crist est lespee le roi david uostre peire le plus merueilleuse qui onques fust forgie et la plus trenchant que onques fust baillie par main de chevalier, prendes la si en ostes le poin et lenheudure et quant uous aures la lemele mis a vne part nous qui connusons les forces des herbes et les uertus des pieres et la matere de toutes coses terrienes faites .j. poin de pieres precieuses si soutilment quil nait apres vous regart domme terrien qui puist connoistre lune de lautre ains quide che chascuns qui le verra que ce soit vne misme piere. apres i faites vne enheudure si merueilleuse que nule ne soit si uertueuse ne si riche, et apres i laites le feure si merueilleuse en son endroit comme lespee sera endroit soi. et quant vous aures toutes ces coses faites si i meterai les renges teiles comme il me plaira....’ (Add. f. 34v: S I p. 133. 19–31). The mention of precious stones suggests knowledge of early bejewelled swords of the Sutton-Hoo or Childeric types, although of course not those particular swords which were unknown in the Middle Ages. See Bruce-Mitford, The Sutton Hoo Ship Burial, inv. 95, noting that a sheath with straps (‘renges’) is also reconstructed; and

SIGNS AND SYMBOLS IN THE ESTOIRE AND QUESTE

549

care of the ‘renges’ or straps, anticipating the need for better ones to be made by Perceval’s sister in the Queste del saint Graal. It is surprising is that only Add. 10292, BNF, fr 105, and fr 9123 have a miniature depicting Solomon; and his wife accompanies him only in Add. A mixed pattern of relationships between picture and text, and among pictures, emerges from the illustration of these maritime adventures in Estoire. Whereas the structure and placing of the miniatures has been most carefully thought out in Add. 10292, the manuscript which also has the largest number of illustrations, it is also clear that the planners/illustrators of Royal, BNF fr 105 and fr 9123 were scrupulous, on occasion, in the attention they paid to the words of the text in those copies. That they chose to omit the contents of Solomon’s Ship is surprising, perhaps an indication that symbolic objects other than the cross itself — handed out without textual justification to Mordrain — were of little interest. Indeed, the absence of illustrations of these episodes in copies that were otherwise illustrated suggest that patron interest and patron wealth were also important factors in determining how, and how much, these texts would be illustrated. II.1. La Queste del saint Graal Of the seven manuscripts examined above for their treatment of maritime adventures in Estoire, only the three Lancelot-Grail Project manuscripts also contain the Queste del saint Graal. I include in addition the Yale manuscript that may have formed part of a Lancelot-Grail cycle together with Paris, BNF, fr 95. Although the maritime adventures form only a small part of the Estoire picture cycle in BNF, fr 9521 — a surprising omission — the corresponding sequence in the Queste del saint Graal is fully illustrated in Yale 229 and offers interesting points of comparison with the three Lancelot-Grail Project manuscripts, particularly as it was made in the same region of eastern Artois for Childeric, see Childeric-Clovis (exhibition catalogue) and Bruce-Mitford, ‘A comparison between the Sutton Hoo burial and Childeric’s treasure’. Unlike Excalibur, the four other swords in the Lancelot-Graal romance: David’s sword, the sword drawn from the anvil by Arthur in the Merlin, the one drawn from the stone by Galaad, and King Pelles’ broken sword mended by Galaad in the Queste — are not given a name. In his final speech to Excalibur at the end of the Mort Artu, Arthur acknowledges the supremacy of David’s sword, ‘Ha! Escalibor, boine espee, la millor que len seust el monde, fors seulement cele as estraingnes renges, ore perdras tu ton maistre et ton droit seignor...’ (S VI, p. 379). 21 See above at n. 12.

550

or western Flanders a decade or so earlier. II.2. Solomon’s Ship The events surrounding the discovery of Solomon’s Ship in the Queste link back to Solomon’s Ship in the Estoire and also point forward to the end of the Queste as they anticipate the achieving of the Grail Quest by the three chosen Grail Knights, Perceval, Boort and Galaad, aided by Perceval’s sister (nameless, as was Solomon’s wife in the Estoire). Galaad is to be the eventual winner of the Grail, and it is for him that the Ship was prepared in the Estoire. Like Solomon’s Queen who tells Solomon what items are to be put in his Ship, Perceval’s sister is a seer who knows the meaning of the Ship and its contents which she explains to the knights even before they find the confirmatory letter of explanation in the Ship itself. Two of the four manuscripts frame the Ship sequence with miniatures emphasising this revelation, most explicitly in a single miniature in Royal where the hand gestures and poses make clear that Perceval’s sister is telling the knights about the significance of the Ship (Plate 16), and in two small miniatures in Yale where they are shown first at sea and then entering the Ship. In Add. 10294, Royal and Douce, attention is paid to which knight is which, as they are distinguished by heraldry, painted on surcoats and occasionally ailettes, so that in all three Perceval is shown in an orange surcoat while in Add. 10294 and Royal, Galaad bears the arms argent [white] a cross gules, referring back to the shield that he wins at the beginning of the Queste,22 and which in turn was King Mordrain’s shield, on which Josephé inscribed the cross in his own blood in Estoire.23 Douce, curiously, gives Galaad the arms normally reserved for Gauvain: argent [white] a canton gules, a surprising case of heraldic mis-identity. Though they vary across the three manuscripts, these are important instances of the beginnings of heraldry used for the knights, whereas in Yale, made a decade or so before, less attention is paid to using consistent colours for the shields or surcoats of the knights.24 Then follows in all four copies the Adam and Eve sequence, a match for what is shown in Estoire, where Add. 10292 marked the story with

22

S VI pp. 22–24; Pau pp. 29–32. S I p. 284.21; Pon p. 555–874.1. 24 See Stones, ‘Les débuts de l’héraldique,’ reprinted in these esssays; and for later Arthurian heraldry, Pastoureau, Armorial des chevaliers de la Table ronde. 23

SIGNS AND SYMBOLS IN THE ESTOIRE AND QUESTE

551

two champies and two miniatures, of the Fall and the Murder of Abel, the latter shown also in BNF, fr 105 and 9123 but not illustrated in the other manuscripts. This time all four manuscripts depict the Fall, where Royal concentrates on God’s reproach after Adam and Eve have eaten the apple, while the other three show the apple-eating taking place, encouraged in Add. and Douce by a serpent with a female head as in Add. 10292’s Estoire miniature. Yale has a second part to its Fall miniature, the Sacrifice of Cain and Abel, with details of flames rising and falling (Plate 17), reminiscent of the Sacrifice scene in the Psalter of St Louis, Paris, BNF, lat 10525. Add. 10294 and Yale both also give the murder of Abel, but at different places in the text, with Cain using different instruments, a short straight knife in Yale and a curved one in Add. 10294 — neither picks up the jawbone motif used by Add. 10292 for the Estoire murder. Douce’s miniature is an addition, showing the Expulsion, on a page containing a champie initial, while Royal gives a second miniature where God addresses Adam and Eve, this time explaining to them the meaning of the change in the colour of the Tree, from white to green, and the change is also shown in the miniature where the tree has a green crown and white trunk (Plate 18). All four manuscripts show the Grail Knights and Solomon’s Ship, but with different emphases. Indeed this is one of the most frequently represented episodes in Queste illustration in general, and there are many variants as to who and what are shown.25 Perceval’s sister is omitted in Douce’s miniature (Plate 19) but she is present in the other copies, explaining meanings again in Add. 10294 and particularly in Royal, where Boort turns back to listen to her (Plate 20). The giving and hearing of explanations, from the seeress and from God, is a focus particularly favoured in Royal. In Yale, Perceval’s sister is given a more active role as she is shown on the shore, handing to Galaad David’s sword with its decorated scabbard and pommel and guard of gold — a subject rarely illustrated, so far as I know (Plate 21). But the ‘estranges renges’ anticipated by Solomon’s wife in the Estoire, and made, according to the text of the Queste, by Perceval’s sister from her own golden hair, are not shown in Yale nor in the other miniatures.

25 It is found, for instance, in Brussels, BR 9627–8, BNF, fr 339, 342, 344, Florence, Laur Ash. 121, Cologny-Geneva, Bodmer 147, Dijon, BM 527, Oxford, Bodl., Rawl. Q. b. 6, Paris, Ars. 3480, BNF, fr 112, 116, 122, 123, 1423, and no doubt others. The variants are too numerous to discuss here.

552

The reading of the letter, found in a decorated purse hanging on Perceval’s arm in Add.10294 and Royal, is the main action in Add. 10294, Royal and Douce, and it also figures in Yale where the letter is shown as a scroll (badly rubbed) read to Boort on the boat by Perceval, complementing the sword-giving in the upper part of the miniature and underlining the importance of written proof. Whereas the sword was prominent in Royal’s opening Ship miniature, it is not shown again at this point in Royal, nor in the other copies considered here, although other manuscripts do include it. The crown, equally prominent in Royal’s opening miniature, is also absent.26 Here emphasis is given in Add. 10294, Royal and Douce to the bed with its multi-coloured cover (no colours are specified in the text), and to the three spindles, white, red and green, prominently displayed on the front plane of the boat, as they were in the Estoire miniatures (Plates 19, 20). The Ship in Yale also has the bed, covered in grey with a green pillow, on a frame painted white with gold finials — but the three spindles and the crown are absent (Plate 21). Other details about the Ship also vary — in Royal there is a sail, absent in the other versions, while Douce’s boat has a tiller. II.2. Perceval’s Sister The lower part of Yale’s miniature shows the four protagonists setting sail towards the next episode, where Perceval’s sister will donate her blood to cure the Leprous Damsel and will die from the bloodletting — and it is notable that none of these manuscripts illustrates the bloodletting incident, only the events that result from it.27 Add. 10294 is the only manuscript to actually depict the death of Perceval’s sister (Plate 22), with an emphasis on her last communion, and Add. 10294 precedes this with a presage heralding divine intervention in the form of the White Stag accompanied by four lions, emblematic of Christ, seen in the forest by Perceval and Galaad. Scenes of destruction precede and follow Perceval’s sister’s death — in all four manuscripts the Grail Knights see the Leprous Damsel’s castle destroyed, in atonement for the death of Perceval’s sister, while leading up to her death are variously placed scenes of combat in which the Grail Knights kill a large 26

The crown is rarely represented at this point. As I have shown elsewhere, the bloodletting is illustrated in Bonn, UB 526 and Florence, Laur. Ash. 121 (Stones, ‘The Illustrations of BN fr 95 and Yale 229’, reproduced in these essays). It is a surprisingly rare subject in Queste illustration. 27

SIGNS AND SYMBOLS IN THE ESTOIRE AND QUESTE

553

number of antagonists. In Royal the focus is once again on explanation, by a priest carrying a ciborium, of these deaths and an expiation of them through the Sacrament. Douce marks the place with a champie initial. All four manuscripts depict Lancelot’s discovery of Perceval’s sister’s body on a boat that links back to Solomon’s enchanted Ship (if only with a champie initial in Douce, and an added miniature of Lancelot hearing a voice). This time the bed has become a bier, draped in Add. 10294 with the same multicoloured cover as in Solomon’s Ship, and again the pictorial emphasis is on explanatory reading as Lancelot finds the letter left by Perceval which explains who the dead damsel is (Plate 23). Once again the written word is an important witness to the truth. Yale alone continues the pictorial narrative, where Lancelot brings the body to a chapel and is greeted there by a hermit (Plate 24). II. 3. The Final Scenes Another death follows shortly afterwards, that of King Mordrain, a character whose story is mainly played out in the Estoire, who dies in the arms of Galaad, physically and psychologically linking the two branches, the two stories, and the two heroes, and marking the succession of Galaad the Grail Winner whose rôle has long been anticipated, as far back as the Estoire. After this come two symbolic acts, the joining of King Pelles’ broken sword by Galaad, an incident paralelled in the Second Continuation of the Perceval of Chrétien de Troyes (where of course Perceval, not Galaad, is the Grail Winner). The illustration in Paris, BNF, fr 1453, f. 218, the only one among the corpus of surviving Perceval illustrations, shows a valet bringing the two pieces of the sword to the table before Perceval and the Maimed King.28 Add. 10294 (f. 50b) is the only manuscript of the Queste manuscripts considered here to illustrate this event, while Yale’s illustrative programme culminates in the Grail Liturgy, shown, surprisingly, in a small rather than a large miniature, but one that is remarkable for its depiction — or rather its concealment — of the Grail, hidden from view within a shrine-like tabernacle placed on the altar.29 It is notable that the other manuscripts omit an illustration of 28

407.

See Les Manuscrits de Chrétien de Troyes, eds. Busby, Nixon, Stones, Walters, II, Plate

29 For further discussion see Stones, ‘The Illustrations’; and ead., ‘Seeing the Grail’, both reproduced in these essays.

554

the liturgy, even those where the liturgy in Estoire is illustrated with several miniatures, not just one.30 In Add. 10294, Royal and Rylands, the final miniature depicts the Grail Knights carrying the Grail and the Grail Table, aided by a cripple (who will of course be cured by the Grail). II.4. Conclusions The illustrations in these romances were not mass-produced from models, nor were the illustrative programmes standard. The stories and their interpretations were multi-valent and patrons, producers and illustrators were well aware of the multiple pictorial possiblities that the stories presented. They capitalized on the many options the texts offered to create products whose illustrations were most likely dictated by a variety of factors, most no longer ascertainable with certainty today, but varied and variable as no two copies, even when produced by the same or closely related craftsmen, are identical in their choice, placement, and treatment of the illustrations. Of the manuscripts considered here, the least expansive are Le Mans and Amsterdam/Douce/Rylands, which so often give a champie where the related manuscripts have a miniature. These manuscripts also simplify, so that the contents of Solomon’s Ship found by Nascien are reduced to just the bed in Le Mans, and Perceval’s sister is omitted in the important scene of the three Grail Knights on Solomon’s Ship in Queste in Douce. These omissions suggest that these copies are reductions of more expansive versions, not early simple versions.31I still believe Rennes to be the earliest surviving illustrated manuscript,32 yet its picture-cycle is meagre, with numerous champies. However, it does include unusual details, notably King Mordrain transported in his bed, but this version was not taken up elsewhere, so far as I know. Yale also gives idiosyncratic detail, notably in relation to the Grail Knights and Perceval’s sister on Solomon’s Ship: and again, so far as I know, its treatment was not repeated. Most thoughtful, in

30

It is notable that BNF, fr 95, most likely the pendant of Yale 229, also illustrates Josephé’s liturgy in the Estoire as do Royal and Amsterdam but not Add. See Stones, ibid. and ead., ‘The Grail in Rylands French 1 and its Sister Manuscripts’, reprinted in these essays. 31 For the opposite view in relation to Amst. in Estoire, see Meuwese, ‘Three Illustrated prose Lancelot Manuscripts’, cited above in n. 4. 32 Stones, ‘The Earliest Illustrated Prose Lancelot Manuscript?’, reprinted in these essays

SIGNS AND SYMBOLS IN THE ESTOIRE AND QUESTE

555

their different ways, are Royal and Add. 10292 and 10294, all taking care to frame important episodes with a ‘before’ miniature and an ‘after’ miniature, in Royal especially in relation to the Grail Knights and Perceval’s sister on Solomon’s Ship and in Add. in relation to the corresponding sequence in Estoire depicting Nascien’s adventures on the Ship. Yet neither manuscript duplicates the choices that the other (and its makers) has made, and their strategies at times are quite different. So Add. 10292-4 give more miniatures overall, with several appearances of Tout-en-Tout’s boat and that of the devil in Estoire, and include the Adam and Eve sequence both in Estoire and also in Queste. Royal’s strategy focuses on images of contemplation, reflection and listening, to God in Estoire and to Perceval’s sister in Queste, and both pay careful attention to the details of what is in each text for the treatment of Estoire’s sail. The strategies in BNF, fr 105 and 9123 are different from each other, and difficult to fathom. Again, attention was paid at times to what the text says yet at other times the content of the miniatures seems almost random, a strange mixture of deliberate thought and careless inattention. These examples show that there was a great deal of flexibility in what was done, and how: exactly why these particular choices were made we shall probably never know, but the variety and difference in pictorial emphasis these manuscripts represent are indications of how eagerly the texts were embraced by medieval makers and owners alike.

XXIV The Illustrations of the Queste del saint Graal in Yale 229 and Other Queste Manuscripts

T

he illustrations in the Queste del saint Graal in Yale 229 share the same formal characteristics as the rest of the Lancelot-Grail romance in Yale 229 and its sister manuscript, Paris, BNF fr 95. Both contain mostly singlecolumn miniatures, for the most part divided into two registers, allowing for elements of sequential narrative to be presented concurrently, a format matched only in Cologny-Genève, Bodmer 147 — or smaller miniatures occupying half a text column, or, just occasionally, historiated initials.1 A distinctive and unusual feature of the illustrated pages in these volumes is the attention given to border decoration, where birds, animals, hybrids, and human figures abound, sometimes complementing the subjects of the main illustrations, at other times satirizing it, or standing in no apparent relationship to the primary narrative.2 No other copy of the Lancelot-Grail romance devotes such attention to this minor decoration. The purpose of the present enquiry is to situate the iconography of the Queste in a broader context, by comparing Yale 229, as I did for Yale’s Mort This essay was first published in The Queste del saint Graal in Yale 229, ed. E. Willingham, Turnhout, 2011, pp. 301–43. 1 Yale 229 is the Beinecke Rare Book and Manuscript Library MS 229 [Yale]. BNF fr 95 is Paris, Bibliothèque Nationale de France, MS français 95 [BNF fr 95]. For Cologny-Genève, Fondation Martin Bodmer, MS 147, of uncertain provenance, see Vielliard, Manuscrits français, pp. 46—60. The manuscript is also available online; see Virtuelle Bibliothek in the Works Cited and directly at . 2 For more on the general layout, see Stones, ‘Illustrations of the MortArtu in Yale 229’; ead., ‘Mise en page’, and others. Reference is made to Yale 229 and its marginalia in Wirth, Les marges à drôleries, pp. 64—65, 135, 147, 157, 166, 224, 253, 274, 357, 363. See also Hunt’s Illuminating the Borders, pp. 6, 10,16, 19, 26, 38, 40, 46, 48, 64, 68–69, 71, 78, 79–88, 97, 110, 124, 126, 132, 151, 174, 191 n. 5, tables 10–11.

THE ILLUSTRATIONS IN YALE 229 AND OTHERS

557

Artu, with the three core manuscripts of the Lancelot-Grail Project, made a decade or so later than Yale in a neighbouring region of northern France or Flanders. These manuscripts are hereafter referred to as follows: London, British Library, MS Additional 10294 (Add.) London, British Library, MS Royal 14 E.III (Royal) olim Amsterdam, Bibliotheca Philosophica Hermetica, MS 1 (Amst.)4 Oxford, Bodleian Library, MS Douce 215 (Douce) Manchester, The John Rylands University Library, MS French 1 (Ryl.) The Queste is transmitted in two versions, like the Mort Artu, and Yale 229 transmits the long version, whereas the three manuscripts with which I compare it all transmit the short version. The differences, however, are slight, so that there are many points of comparison to be made in terms of how the pictures are distributed and treated. Once again, as is the case with the Mort Artu, there are more illustrations in Yale than in the other three manuscripts: Yale has a total of 38 illustrations, Royal contains 31 and Add. and Ryl./Douce have 25 each (though Ryl./ Douce is incomplete and no doubt had illustrations in the sections that are now missing). Some missing illustrations can be reconstructed on the basis of cuts on adjacent pages, as noted in the Table that follows this essay). Whereas the three Lancelot-Grail Project manuscripts all adopt the format of singlecolumn miniatures and lack border illustration altogether in their Queste section,5 Yale is characterised by a variety of types of illustration, ranging from single-column miniatures in a single register (three examples), the same in two registers (fifteen examples), small (half a text column) miniatures (seventeen examples), and small (less than half a text column) historiated initials (three examples), all accompanied by curvilinear borders supporting a large variety of birds, animals, hybrids, and people — often playing out a scene which sometimes continues the narrative of the major illustration, but at other times stands in no apparent relation to it. See The Lancelot-Graal Project at