IF THERE IS ONE SCRIBE WHO needs no introduction (with the possible exception of a certain notorious Adam), it is the one responsible for copying in whole or in part, gathering, and assembling the ‘miscellanies’ of London, BL, MSS Harley 2253, Harley 273, and Royal 12 C. xii.¹ The term ‘miscellany’ is destined to remain problematic, and one might be forgiven for thinking that the term is sometimes used when a manuscript does not lend itself to being otherwise described. Thinking of the great Continental French manuscripts which have been called miscellanies by scholars writing in English, miscellaneous is probably the last thing they are. In French, manuscripts such as Paris, BnF, MSS fonds français 837 and 19152, or Bern, Burgerbibliothek, MS 354, would be called recueils, that is books constructed by the purposeful gathering together of sometimes large numbers of texts.² Although fr. 837 seems to be mainly a collection of short verse texts, secular and sacred, both fr. 19152 and Bern 354 also contain full-length romances, and the designation (p.50) of them as ‘fabliaux manuscripts’ is somewhat misleading. The Paris manuscripts are both from north-eastern France and date from the end of the 13th or beginning of the 14th century, while the Bern codex may be as early as the second quarter of the 13th and has linguistic features which suggest a provenance in Burgundy. If one cannot go so far as to deny the existence of a Continental vernacular unplanned miscellany, it could be argued that as an apparent category, it is not conspicuous. It may be inconspicuous simply because of the large number of impressive structured recueils. It is worth noting that there has been little real theoretical discussion about the nature of the French recueil codex, although some major projects are beginning to get to grips with the question.³ In any case, until recently, the issue has not been problematised. I have argued that certain principles underlie the ordering and disposition of texts in these manuscripts, although the juxtaposition and sequencing of almost any number of disparate texts can, of course, generate meaning by way of thematic and stylistic similarities and contrasts when the texts are read as a group.⁴ There is no way of knowing to what degree we are dealing with what might be called ‘scribal or compilatorial intention’ and there is consequently no need to fall into the trap of scribal or compilatorial fallacy. Nevertheless, it is almost impossible to avoid asking what scribes and planners were thinking when ordering texts within a book, although the matter is not central to my arguments here. The texts and their disposition in the manuscripts are what they are and they interact as they interact, or (it could be argued) as we encourage them to. In this chapter, I consider the ramifications of multilingualism as they pertain to the books of the so-called ‘Harley scribe’ and other insular manuscripts, before comparing them to the work of the late 13th-century Italian copyist, Johannes Jacobi.

A consensus is still lacking on the meaning of the term ‘miscellany’, and comparisons could be made with discussions about the validity of the term ‘minstrel manuscript’ (or manuscrit de jongleur). The existence of this category had been more or less taken for granted before Andrew Taylor’s important article which essentially and convincingly debunked the notion of ‘the small, (p.51) plain, battered working texts that the minstrels are alleged to have carried with them on their travels’.⁵ That there are a number of modest, small-format manu scripts containing Old French epics (copied on both sides of the Channel) is a fact; whether there are grounds for categorising them according to a common hypothetical function is another matter altogether. Leaving aside the difficult case of the Oxford Chanson de Roland (BodL, MS Digby 23; England, s. 12^1/2), three examples among many are Göttingen, Niedersächsische Staatsund Universitätsbibliotek, MS Philol. 184.IV of the Chanson d’Aspremont (England, s. 12^ex.–s. 13^inc.); Paris, BnF, MS fr. 2493 of Raoul de Cambrai (northeastern France, s. 13^med.); and Tours, BM, MS 936 of Huon de Bordeaux (Picardy, s. 13^med.). There are also similar manuscripts containing romances and saints’ lives, for example Tours, BM, MS 942 of Chrétien’s Cligés (western France, s. 12^ex.–s. 13^inc.) or London, BL, MS Harley 270 of Becket’s life by Guernes de Pont-Sainte-Maxence (England, s. 12^ex.), but these rarely figure in discussions of the minstrel manuscript.⁶ The debate concerning both the minstrel manuscript and the miscellany surely derives from our stubborn scholarly obsession with categorising on the basis of modern criteria, and tends to ignore the elephant in the room, namely that each manuscript can usually be considered unique in one way or another.

In an article first published in 1991 and revised for inclusion in the 2000 collection on Harley 2253, Theo Stemmler passes in review earlier scholarship on the notion and tries to distinguish between ‘miscellany’ and ‘anthology’, claiming that Harley 2253 is neither a miscellany nor a well-planned book, but that it can best be regarded as an anthology intended to be representative of certain types of text. Its main structuring features are the contrasts verse–prose and secular–religious, and although no sustaining organising principle is at work, groups of texts can be discerned, enabling us to gather insight into a compiler’s ‘psychology’.⁷ In her introduction to the same book, Susanna Fein takes a cautious approach: ‘The term “miscellany” is here used in the broadest sense (a collection of texts miscellaneous in genre, style, and languages, and not selected for a single discernible purpose). Its use is not meant to deny the sequenced grouping of items that is often apparent in the book.’⁸ (p.52) Ralph Hanna stresses the production of miscellany manuscripts in stages, concluding that their diverse nature may be due to ‘a combination of happen-stance acquisition and variously motivated selections’.⁹ Andrew Taylor similarly argues for the assembly of manuscripts such as Harley 978 subsequent to the acquisition of individual booklets. This well-known English manuscript from the 1260s, for example, contains music, medical texts, the Fables and Lais of Marie de France, Noctis sub silentio, Goliardic verse, and The Song of Lewes.¹⁰ At the beginning of an article on multilingual miscellanies, John Scahill states (perhaps unavoidably if not very helpfully) that ‘a miscellany has a cohesion of some kind’, while Tony Hunt, in a survey of what he terms ‘insular trilingual compilations’, had earlier observed that ‘they all, to varying degrees, exhibit some principles of arrangement’.¹¹

These varied (and considered) opinions should not be taken as suggesting that further deliberation is pointless, rather that anything other than a consensus in the broadest possible terms is unlikely and, it could even be argued, undesirable. Complex issues cannot be reduced to the scholarly equivalent of soundbites. It is in any case clear that conventions and procedures involved in the assembly of compilation manuscripts vary enormously from place to place, from language to language, and from period to period. More may be learned from the differences between manuscripts than from the similarities. Much has been written on the arrangement of texts within Harley 2253, and somewhat less on Harley 273 and Royal 12 C. xii.¹² All the possibilities of Harley 2253 have not yet been exhausted, and attention should still be paid (pace the matter of intentions) to Susanna Fein’s proposal that ‘To understand (p.53) the Harley scribe and his compilatorial purpose, we must continue to pursue refined studies of the many texts he purchased, borrowed, read, annotated, copied, altered, and set side by side with other texts in Harley 2253, Harley 273, and Royal 12 c. xii.’¹³ New compilations or miscellanies or recueils can be made from the entire corpus of texts contained in the three manuscripts and scholars may consider what light each may cast on the other.

There is enough material here for at least one book, particularly as some of the longer texts in the corpus are rarely studied in their own right, let alone in their manuscript context. The first section (fols 1–48) of Harley 2253 comes to mind (containing the Vitas patrum, biblical and hagiographical texts in French and not reproduced in Ker’s facsimile),¹⁴ Richard de Fournival’s Bestiaire d’amour, William of Waddington’s Manuel des péchés, the Purgatoire saint Patrice, Bozon’s Plainte d’amour (fols 70r–81r, 113r–190v, 191v–197v and 199r–203r of Harley 273), the Miroir de sainte Eglise, and Fouke le Fitz Warin (fols 17r–30r and 33r–61r of Royal 12 C. xii). The final section of Royal 12 C. xii (fols 77r–123r), with its prognostications and sortes, offers plenty of material which would tie in with many of the other works in the three books. The numerous exempla of the Manuel des péchés provide commentary on almost any of the narrative texts, while the Miroir de sainte Eglise offers a more austere view of the secular productions. Another clear possibility would be to read any number of the poems dealing with love (including the secular Middle English lyrics from Harley 2253) alongside Richard’s Bestiaire d’amour in Harley 273. The rules of friendship in Harley 273 (fol. 85r–v), Ami et Amile in Royal 12 C. xii (fols 69r–76r), and ‘Cyl qe vodra oÿr mes chans’ on fols 61v–62v of Harley 2253 might also form a mutually enlightening group. Generally speaking, some of the rarely read shorter texts would be worth re-examining in a broader, inter-codicological context. For example, the pilgrimage and crusade texts on fols 68v–70v and 129v–130v of Harley 2253 might be read together with the exemptions for Hospitallers of St John of Jerusalem on fols 2r–3r of Royal 12 C. xii. The first pilgrimage text in Harley 2253 is a description of routes and sites, leading into a second, another list of sites and the pardons relating to each; the crusade text is known as La terre des sarazins and is a translation of a Latin original, supposedly instructions from the Patriarch of Jerusalem to Crusaders. The text of the exemptions in Royal 12 C. xii also confirms the regional links of the group of manuscripts, specifically mentioning the Mortimers, the Bohuns, and the towns of Hereford (p.54) and Gloucester (there was a Hospitaller house at Dinmore, 7 miles north of Hereford).¹⁵

Of the three manuscripts, two (Harley 2253 and Royal 12 C. xii) are trilingual, while the third (Harley 273) is bilingual French–Latin, mainly French with a few short texts in Latin. The two vernaculars dominate in the trilingual manuscripts, with French to the fore. There are only two substantial works in Middle English, namely the Metrical Chronicle in Royal 12 C. xii (fols 62r–68r) and King Horn in Harley 2253 (fols 83r–92v), although if the first 48 folios are included, Middle English accounts for about 25 per cent of the latter manuscript. Nevertheless, the textual corpus of the three manuscripts is predominantly insular, strongly suggesting that the literary milieu in which the Harley scribe moved and for whom he may have worked in the Ludlow area was to a large degree francophone.¹⁶ Alan Wilshere has argued in connection with Fouke le Fitz Warin in Royal 12 C. xii that French may not have been his native language, but his French is remarkably fluent, albeit showing the influence of West Midlands English: ‘The writer responsible for the extant FFW in prose was no doubt bilingual in everyday life, and may have spoken an “advanced” (creolised or pidgin) version of French, and pronounced it with a markedly English (probably West Midland) accent. The influence of Middle English may be discerned in his syntax, and in certain simplifications of verb and pronoun morphology.’¹⁷ The language of Fouke may be more ‘correct’, however, as the text is based on an earlier verse original, but it is believed that the Harley scribe ‘was himself the author of the prose remaniement which he copied’.¹⁸ The complex linguistic matters arising from situations such as this (p.55) constitute a salutary reminder of the importance of philological training and the desirability of collaboration between romanists and anglicists.

It has been argued, by myself among others, that the distribution of texts by language suggests that certain types of text were considered as best suited to expression in a particular idiom.¹⁹ The fabliau is the clearest example, with the Middle English Dame Sirith from Oxford, BodL, MS Digby 86 (fols 165r–168r), an early but unsuccessful bucking of the trend (see also The Fox and the Wolf on fols 138r–140r of the same manuscript).²⁰ If French is also the vernacular language of the long edifying works such as the Manuel des péchés and the Miroir de sainte église, it is not the exclusive medium for romance, for despite Fouke in Royal 12 C. xii, we also have the Middle English King Horn in Harley 2253, both in the Harley scribe’s hand. The Matter of England in a broad sense can, it seems, be appropriately expressed in both the native vernacular and that of the descendants of the Norman aristocracy. It is a critical commonplace (but nonetheless true for all that) to say that the expression in Middle English of the love lyrics in Harley 2253 is one of the book’s great achievements, for the love lyric in its various manifestations had long been principally the preserve of Continental French (with just a few chansons of insular provenance) and Occitan. If there are fuzzy notions of language-specific text-types, these notions can therefore also lead to attempts to undermine them (as in the case of the Harley lyrics or the two poems in Digby 86 mentioned above), but there are no rigid formulations such as those found in Ramon Vidal, Brunetto Latini, or Dante’s De vulgari eloquentia.²¹ By ‘undermining’ the notion language-specific text-types, I mean conscious attempts to give expression to a ‘genre’ or text-type in a language other than that in which it became prominent.

Although a number of regions in medieval western Europe used two or more vernaculars at various times, the trilingual manuscript containing Latin and two vernaculars seems to be a largely, albeit not exclusively, (p.56) English phenomenon.²² Manuscripts containing Latin and one vernacular, of course, are legion. There are few, if any, manuscripts, miscellanies or otherwise, which contain Old French, Middle Dutch, and Latin, or even just Old French and Middle Dutch. The vernacular used at courts in Brabant and Hainaut changed frequently, often with shifts in power, and although vernacular bilingualism was a practical necessity, it does not seem to have been reflected in book production.²³ The same goes for Old French and Occitan, where the bilingual border areas seem to have produced texts in one idiom with dialectal features from the other rather than manuscripts including both langue d’oc and langue d’oïl. In east-central France, western Switzerland, and the Aosta valley, a variety of dialects are considered ‘franco-provençal’ but there are no early bilingual manuscripts. There are texts in the langue d’oïl which quote the troubadours in the original langue d’oc, but that is a different phenomenon.²⁴ Tony Hunt, following on P. J. Frankis, lists some eight ‘insular trilingual compilations’, including Harley 2253 and Royal 12 C. xii, but excludes Harley 273 as merely bilingual.²⁵ He also considers other well-known compilations such as Oxford, Jesus College, MS 29; BL, MS Cotton Caligula A. ix; Cambridge, Trinity College, MS B.14.39; Oxford, BodL, Digby 86; BL, MS Egerton 613; BL, MS Additional 46919; and BL, MS Harley 913.

Jesus College 29, pt II (Worcestershire) and Cotton Caligula A. ix (Herefordshire) both date from the second half of the 13th century and have a number of texts in common, such as the Middle English Owl and the Nightingale and the French works of Chardri (the Petit Plet, Les set dormanz, and a version of the Barlaam and Josaphat legend). Both also contain short religious works in Middle English, but if the Jesus manuscript seems to be more oriented to the didactic, the Cotton manuscript contains a version of Layamon’s Brut and an Anglo-French chronicle which lend it something of a nationalistic air.²⁶ Trinity College B.14.39 (Worcester [?], c. 1260) is largely homiletic in nature, with more than half of its texts in Latin, and some (p.57) macaronic pieces. It is best known for its Middle English religious lyrics but also contains extracts from Robert of Greatham’s Evangile des domnées, the Passion des jongleurs, and a long extract from Wace’s life of St Nicholas.²⁷ Digby 86 has already been mentioned above and is sufficiently well known not to require further analysis here.²⁸ Egerton 613 (south-west Midlands, s. 13^4/4) is a composite volume, including religious texts in French (verse and prose), and two versions of the Middle English verse sermon, Conduct of Life.²⁹ BL, Additional 46919 (Herefordshire, c. 1340) is the ‘commonplace book’ of William Herbert, Franciscan Friar of Hereford, and supervised by him, although written by over a dozen different hands. Its four principal sections contain 40 didactic and religious works in French (fols 2r–36v and 38r–153v), notes, excerpts, and outlines in Latin (fols 154r–203v), and 19 Middle English lyrics by Herbert himself (fols 205r–211v); Herbert annotates Nicole Bozon’s sermons and Walter de Bibbesworth’s Tretiz.³⁰ Harley 913 is the only trilingual manuscript known to have been mainly copied and assembled in Ireland, probably at the Franciscan monastery of St Saviour’s, Waterford, c. 1335. It contains 17 items in Hiberno-English, 31 in Latin, and three in Hiberno-French. The contents include preaching materials for the Franciscans in verse and prose, and satirical and political pieces. Its most substantial French text is an occasional poem usually known as The Walling of New Ross.³¹

The other area where multilingual compilations might be expected is Italy, where the Tuscan of Dante and Boccaccio was frequently found in the company of both the langue d’oc and the langue d’oïl. It is clear, of course, that the use of French in this region of medieval Francophonia is of a different order to that of England. In England (or rather, the British Isles, to be more precise), insular French (also in its Cambro-, Hiberno-, and Scoto-variants) was a practical language in everyday use in certain social contexts whereas in Italy, (p.58) its use, like that of Occitan, was largely attributable to literary fashion. It was, however, also a language of administration and commerce. I leave out of consideration here the complex question of Franco-Veneto (earlier referred to as Franco-Italian).³² If we are dealing with colonialism here, it is a form of cultural colonialism in which the colonists were welcome and invited guests. The vast majority of manuscripts in the langue d’oïl produced in Italy from the middle of the 13th century onwards are of longer works and consequently self-contained. They are predominantly of Arthurian prose romance, the romans antiques (usually Benoît de Sainte-Maure’s Le roman de Troie), chansons de geste, and universal histories such as L’histoire ancienne jusqu’à César and Li fait des romains.³³

Although copies were made in Italy of the Lancelot-Graal (Vulgate) cycle, numbers of this are relatively small in comparison with those of the Post-Vulgate, the prose Tristan and its derivatives, such as Guiron le courtois and Meliadus. Pisa was a major Italian centre of production of manuscripts in the langue d’oïl, and there is one Arthurian manuscript copied there (Paris, BnF, MS fr. 12599, s. 13^4/4) that is curious for two principal reasons. Firstly, it is an anthology of extracts from Guiron le courtois, the prose Tristan, the Post-Vulgate Queste del saint Graal, and the Suite du Merlin, and secondly, because fols 17r–38v (part of Guiron) are copied in Italian by the scribe of fols 1r–106r, 222r–230r, and 501r–511r. There is no break in the text during the change to Italian and the return to French. This scribe names himself as ‘Oddo’ on fols 63r and 71r. It is not clear why Oddo switched from French to Italian and back again, but the important point here is that he did, without apparent difficulty, and without the need for comment. He would also, of course, have been competent in Latin as well.³⁴

The Harley scribe’s trilingual production is trumped by another Italian scribe, one Johannes Jacobi, whose output is known to have included at least four manuscripts: Paris, BnF, MSS fr. 12571 and 24376, Chantilly, Musée Condé, MS 470, and Lyon, BM, MS 739. There is absolutely no question about the common provenance of these four manuscripts: the identity of the hand itself, which dates from about 1280, is perfectly clear, and many details of the mise en page are identical. Most of the texts copied by this scribe date from the last quarter of the 12th century. Johannes names himself on the (p.59) verso of the final, unnumbered, folio of the Chantilly manuscript: ‘Per me Johannem Jacobi’. Linguistic evidence may point to Verona as his place of origin and work. Three of these manuscripts are single-item monolingual codices: fr. 12571 is of Jaufre in Occitan; fr. 24376 contains Florimont by Aimon de Varennes in Old French; and Chantilly 470, the Chanson d’Aspremont in Old French. Lyon 739, however, is trilingual Latin–Old French–Italian, containing some biblical texts and a prayer in Latin; poems on the Passion and the Virgin in French; the Concepcion Nostre Dame by Wace; La prière de Théophilus (II Prière 37) by Gautier de Coinci; Les XV signes du jugement dernier; and a poem in praise of the Holy Ghost in Italian. There are marginal notes in Italian in another hand in fr. 24376, and instructions for the illuminator which sometimes double as (inappropriate) rubrics for the unexecuted miniatures in Chantilly 470: many begin ‘Fa qui como …’ (‘Do/Show here how …’), ‘E qui fa como …’ (‘And here do/show how …’), or ‘E qui como …’ (‘And here how …’). The language of the instructions points to Treviso, and that of the rubrics, to Ferrara.³⁵

A number of conclusions may be cautiously drawn from this group of manuscripts. As regards the nature of texts they transmit, Florimont (two other Italian copies) as an Alexander-text falls into the broad category of romans antiques frequently copied in Italy, and the Chanson d’Aspremont (five other Italian copies) has a serious history there. The three Italian copies of Jaufre, of which much of the humour depends on its parodic treatment of Chrétien de Troyes, suggests a greater knowledge of the verse tradition of Arthurian romance in Italy than is generally supposed and a broader knowledge of Occitan than that derived from the troubadour lyric.³⁶ The presence in this corpus of the religious poems comes as no surprise, but they are serious testimony to the circulation of the texts in question, which are not usually associated with Italy. In their own way both the Harley scribe and Johannes Jacobi reflect the narrower interests of their time and place as well as a corpus of literature in the langue d’oïl that has a universal appeal and transcends specific contexts. Any notion of the West Midlands/Welsh Marches area as culturally isolated necks of the woods is belied by some of the ‘classic’ French (p.60) texts included in both Harley 2253 and the somewhat earlier Digby 86.³⁷ The same is clearly true of northern Italy in the latter part of the 13th century and first decades of the 14th.

These two scribes have produced a body of work in which their non-native languages predominate. Like the Harley scribe, Johannes Jacobi seems to have moved effortlessly from one language to another, and seems to have been competent in all, although his Old French and Occitan show varying degrees of Italianisms,³⁸ just as the Harley scribe’s French may reveal the influence of his own dialect of Middle English. Detailed linguistic analysis may confirm initial suspicions that the Concepcion Nostre Dame and the XV signes du jugement dernier were copied from Norman or insular exemplars. There is precedent for the circulation in Italy of Norman or insular manuscripts in the early Milan, Biblioteca Ambrosiana, MS D.55 Sup. of Le roman de Troie (Venice, s. 13^inc.), in which some features of the exemplar are visible through the veneer of Italianisms.³⁹

Comparison of the vernacular manuscript production in the different regions of medieval Francophonia is a fascinating and fruitful one: from Ireland to the Welsh Marches, from the Lowlands to the Midi, from northern Italy to Cyprus and the Levant, comparative codicology may reveal much about the role of literature in the various manifestations of the langue d’oïl, and even form the basis for a new history of literature in Old Frenches. Such a history will also contribute to the history of multilingualism and transgress traditional critical and philological domains. It may even need to include consideration of manuscript ‘miscellanies’.