Articulating the author: Gower and the French Vernacular Codex.

ABSTRACT

Perceptions of authorship in the late medieval period are in a state of transition. This essay assesses ways in which ideas of authorship in French books provide a context for interpreting authorship in Gower's Confessio amantis. I consider author portraits, rubrication, and the use of the prologue in Le Roman de la Rose, Le Roman de Fauvel, and collections of works by Machaut and Froissart. Gower's use of Latin, far from being a sign of conservatism in any simple sense, emerges as a strikingly distinctive means of investigating the complex guises under which authorship was emerging in the books of vernacular writers.

As everyone knows, there are four John Gowers in Confessio amantis. As well as John Gower the historical author of the work, there is the Gower who is named as such in the Latin glosses: there is also the celebrated identification of Amans as Gower towards the end of the work. The fourth Gower--Gower as narrator--is harder to isolate, since in a sense he emerges only through that double shock that Amans is at once the author, and old. Through this disorienting revelation, the distinction between author and persona that had seemed so clear earlier in the work appears to fail. Confessio amantis causes the collapse of its distinction between auctor and amans to upset that between auctor and narrator. This is a powerfully unsettling manoeuvre, especially since it is performed both within and without the narrative frame of the work. It heavily qualifies our ability to regard author, narrator, and lover as either stable or distinct categories. (1)

Perceptions of authorship in the late medieval period are in a state of transition. We find considerable evidence of this both in Latin and in several European vernaculars. (2) Yet despite, or perhaps because of the dominance of Chaucer, the evidence for such perceptions in English writing remains to be fully explored. Here, as in the broader Latin and European context, it is not just a matter of considering an author's self-fashioning, but also the ways in which authorship is shaped by external factors located in the physical form of the book. Gower's Confessio amantis is a conspicuous instance of a work in which a debate about the author occurs on both fronts. I investigate this debate in Gower by means of the layout of copies of the Confessio, and that of books in a related, and highly influential vernacular. In both types of material, the relationship between authorship and the anthology is central: these are books in which authorial control is visible through decisions about how one kind of material connects with another. They are examples of compilations created by authors of their own work.

The naming of Amans, in what I have referred to as the inner frame of the Confessio, that is, where the device of the lover's confession is introduced, displayed, and concluded, marks the moment when the first-person narrator has ceased to be a lover. Teasingly, 'as it were halvinge a game', Venus asks for his name: ' "Ma dame," I seide, "John Gower" ' (viii, 2319--21). (3) This utterance closes the device. He is now 'John Gower' because his confession is over. To seal the process, Venus hands him a mirror. Through that marvellously macabre vision of his abruptly ageing face, the narrator can stop pretending to be a lover and reveal himself to be the author. Yet as John Burrow has subtly expounded, auctor and amans turn out to be not so much sharply distinguished as ironically united. What seems to be a transparent 'reidentification of narrator with author' through a rejection of the device of a persona becomes instead a recognition that the author and the persona are after all uncomfortably intertwined. The pose of lover is not so easily cast o. either by an author or a narrator of love poetry. (4)

The outer frame also has a profound role to play in our apprehension of Gower. By outer frame I mean two things: the Latin material (both the verse and the glosses) that physically as well as structurally frames the English text, and the various devices of mise en page--rubrics, headings, running titles, and various forms of paragraphing sign, and the pictures, initials, and other types of decoration--that further contribute to the articulation of the work. (5) The presence of Latin dislocates the notion of authorship in further directions. Firstly, there is the question of how verse and gloss relate. Although there is little reason to doubt that the Latin verse was composed by the historical Gower, scholars have been more cautious over making the same assumption about the glosses. The faint possibility (now usually disregarded) that the glosses bring a genuinely different voice into the work--at least into its outer frame--underlines the formal and structural difference of the two kinds of Latin. As well as a notion of authorship, then, the Latin causes us to ponder the formal boundaries of the work: whether Confessio amantis is a bilingual work to which a gloss has been added, or else a bilingual work that is constituted by frame, narrative, and gloss together. (6)

It is interesting that Gower is only named in Latin in the glosses. Because of their linguistic discontinuity with English, and their functional status as an oblique commentary on the work, while the glosses claim the title of auctor for Gower, they do so in a context that distances this claim from the other namings of Gower. The multiple articulation of Gower in Confessio amantis presents John Gower as Latin, English, auctor, commentator, narrator, and amans, with several of these voicings occurring simultaneously on any one page of the work.

In considering the boundaries of the Confessio, layout is crucial. The form of the work has a physical as well as abstract existence. The disposition of text on the page, its arrangement into sections, the use of headings, rubrication, and initials to create a hierarchy of meaning all contributes to, indeed helps to determine, the status of its constituent parts. For example, the reader might regard text as more authoritative if it occupied the central columns rather than appeared as marginal annotation. It is in this sense that mise en page becomes yet another means by which a concept of authorship is communicated; it is another factor that shapes the voice of authorship, and in the case of the Confessio, complicates our reception of it.

In all these areas--the academic tradition of auctoritas and of the commentaries and textual devices used to analyse and present it, compilatio and the principles of vernacular translation, and the interpretative value of various elements of the page--much important work has been published. (7) In the case of the latter, particular attention has been given to the author portraits, to patronal or dedication pictures, to speech markers, and to the programme of glossing and rubrication. (8) With so much primary observation to make and record, it is perhaps not surprising that discussion has remained largely local. Where an intellectual context has been sought, scholars have tended to go directly to Latin materials for comparison. The importance of Ovid and of the way Ovid was transmitted and interpreted in the medieval period is clearly of great relevance to Gower. (9) Yet the work of setting such scribal and intellectual practices within a broader vernacular secular context remains relatively unexplored. (10) It hardly needs stating that scribes working in England in the fourteenth and fifteenth centuries were not engaged in a narrowly insular enterprise, but were rather involved in literate circles that were continental in outlook and interests. This essay forms part of a larger project (or set of projects) to view English books from the perspective of the dominant manuscript culture of thirteenth, fourteenth, and fifteenth-century French secular writing. My purpose here is to assess some ways in which ideas of authorship in French books might provide a context for interpreting authorship in Confessio amantis. I will be considering author portraits, rubrication, the use of the prologue, and the interconnections between them.

There are two main types of author picture in the Gower manuscripts: the portrait of Amans as either a young lover, kneeling at the foot of the priestly Genius, or, in the case of Oxford, Bodleian Library MS Bodley 902, an older man with a white beard. (11) This portrait tradition is part of an exploration of doubleness in relation to authorship that is fundamental to French secular narrative. The most celebrated example is that arch-text on love, Le Roman de la Rose. We commonly refer to it as if it were a single work, yet of course, it is double-authored, with the second text joined retrospectively to the first in a sleight of reproduction. The trick that Jean de Meun plays occurs in the mid-point of the conjoined text: in a speech by Amor, the Lover, Amant, is suddenly named as Guillaume de Lorris. Guillaume has pledged to Amor that he will begin a romance in which Amor's commandments will be written down. However, so Amor explains, Guillaume will die before finishing the poem. Fortunately, another writer will come to finish the romance. He will be called Johans Chopinel; he will be born at Meun, and he will continue Guillaume's work, more than forty years after his death. (12)

In this fascinating and peculiar passage, Jean abruptly, yet with carefully considered timing, makes the making of the romance his subject. He also turns the role of the author into his subject, in the sense that he makes the subject of the romance, L'Amant, become the author, Guillaume. Yet the act of doing so wittily casts into doubt the status of what we are now reading. If Guillaume will die, then whose work are we reading? In killing the author as soon as he has named him, Jean exposes the work that he has thus far pretended was single and continuous as broken-backed. By naming himself so late into his continuation, Jean shows how he has already successfully deceived the reader into thinking that the romance was the work of one author. From this moment on, we now think of the text in a different way, as well as wondering whether we should retrospectively change our false impression of what we have been reading up to now.

Gower's machinations in Confessio Amantis clearly recall the intricate web of confusion that Jean throws over his role as author. The naming of Amans as John Gower matches the naming of L'Amant as Guillaume de Lorris: in both poems, this is the first time the Lover has been identified. Yet whereas in the Rose, Jean is revealing a difference between two distinct authors, the Confessio explores this distinction within the tighter frame of a 'single' author. The doubt that grows in our minds as we watch Amans's face growing old is a kind of compressed version of that which we experience in the Rose: the doubleness involved in dealing with two authors becomes the doubleness involved in a single author choosing, with deceptive clarity, to lay bare his mechanism of pretence.

John Burrow has discussed the way in which the confession portraits in the Confessio manuscripts are caught up in a structure of misinformation. The majority, by representing Amans as young, could be said to contribute to the depth of the reader's surprise when he or she discovers that Amans is old. The tradition of author portraits in the Rose manuscripts provides an important context for this discussion. Drawing on the pioneering work of David Hult, more recent studies of the illumination of the Rose have confirmed that author portraits usually occur at one of two places: at the end of Guillaume's section of the text, or else at the mid-point where Jean names them both. (13) In the former, it is unclear whether the figure in the portrait (usually seated at a desk with a book open in front of him) is supposed to be Guillaume or Jean. Occasionally, scribal rubrics indicate that this is where Guillaume stops and Jean continues, but the portrait itself is not generally identified. As Hult points out, there is some ambiguity as to whether the figure represents an author or a scribe. In some, he is cast as a reader. (14)

This ambiguity in relation to the portrait is full of interest. First, as with Confessio amantis, though on a somewhat larger scale, the representation of the author-figure relatively early in the (combined) work contributes to the textual ruse. For most manuscripts, in showing only one figure, obscure the fact that one author is succeeding to another. Whether this happens as a result of confusion on the part of the illuminator (or person in charge of the overall design of the manuscript), or of his subtle involvement in Jean de Meun's scheme, it is often hard to say. In either case, the Rose shows authorship to be fraught with contradiction. Jean de Meun's declaration of himself as author is bold and assertive, yet it is also a means of displacing his authorship. He does not exist in his own right, but as a continuator of Guillaume's text: he does not even give himself a direct voice in his text, but, described in the third-person by Amors, he becomes half-fictionalized. (15) In the same way, the portraits at once reveal and veil his role as author. This is especially true of those that occur after Guillaume's section ends: on the one hand, one could argue that precisely because there is no reference in the text to a change of author, to show an author figure here at all indicates a raised consciousness of Jean de Meun's later claim. At the same time, the fact that they can be read as pictures of a scribe or reader fragments the author-figure in a different direction. He may not (for the most part) be shown to be two authors, but his role is none the less divided, in these cases into writer and reader.

A very small number go further and render the continuation as a visual event. In Paris, Bibliothe'que Nationale, fonds francais 1569, fol. 68v, a figure in red is handing a closed book over to another, who is waiting with his hands ready to receive it; in BL MS Stowe 947, fol. [30.sup.v], the two figures are shown side by side in the act of copying. One is working with a nearly finished page; the other has just started a blank leaf. Such examples indicate a high sense of control over the physical representation of the work: the scribes and illuminators are colluding with Jean's textual game to show how authorship is a matter of making a book, and not just an intellectual claim or abstract attribution.

Gui de Mori, a late-thirteenth-century clerical reader of the Rose, provides the most extraordinary example of such collusion. (16) He puts himself forward as a third person who, discovering two separate authored texts, decides to put them together. However, he does more than compile, he makes radical revisions, forging a new composite text on his own terms. In a passage comprising a combination of prose rubrics and verse interpolations, Gui explains how Jean's continuation came into his hands after he had finished copying Guillaume's poem. (17) Gui's role moves with astonishing fluidity between those of reader, scribe, and author. Inserting his own voice into the work at the juncture between Guillaume's poem and that composed by Jean, Gui positions himself as a third kind of author figure, somewhere between reader and scribe, mediating the work of the other two authors to a further audience of readers. He articulates this further by writing himself into Jean's obscure form of ironic prophecy: after Guillaume will come Jean, says Amors in Jean's Rose; and after Jean, says Amors in Gui's Rose, will come another ('uns autres vendra'). He will appear in 1290 and his name will be Gui de Mori, although he will not have the renown of Jean or Guillaume. (18) Taking his cue from Jean, Gui gives the humble copyist the author's own lines.

If the Rose manuscripts express a notion of authorship through the physical text, the MS fr. 146 version of Le Roman de Fauvel celebrates it with unusual extravagance. This manuscript, which has recently been the focus of detailed study from a wide range of disciplinary perspectives, explores the physicality of the book with a rare intensity and intricacy. (19) It stands out from the other surviving copies of Fauvel by the addition of a huge number of musical interpolations, supplemented by seventy-seven drawings and considerable textual amplification. The original roman (written c. 1310--14 in two books by the Parisian notary Gerve's du Bus) was reworked c. 1314--16 by Chaillou de Pesstain. (20) He made two kinds of change: in the first book adding musical pieces (mostly in Latin) but hardly altering the text, and part way through the second book, at the approximate mid-point of the whole work, adding a large number of further pieces (including the bulk of the French material) and around three thousand lines of extra narrative.

To condense what I have described elsewhere, the manuscript is a remarkable instance of a book that constitutes, rather than merely represents a work: the complex interpolation of music, text, and image forms a visual structure on each page that is crucial to the apprehension of the work. (21) For example, several musical pieces are silently placed on the page without being linked directly to the narrative with a textual cue. At such moments, layout takes over from narrative. If the work were being performed, one would have to assume from the layout either that the piece would have to be simultaneous with the narrative, or else that spontaneous (and unauthorized) decisions on when to interrupt the narrative would have to be made. (22)

The sophisticated thinking behind the production of MS fr. 146 shows how far the material form of the book could be conceived as an intellectual space that contributes in its own right to the meaning of the work it embodies. Both Jean de Meun and the Fauvel compilers are investigating the extent to which the idea of authorship is coterminous with the sense of a work as a whole, and hence, with a work's physical length. We can see this in the interesting adaptation made in Fauvel of Jean's handling of the trope of authorial succession. In the bulk of the Rose manuscripts, two alternative views of the extent of the work are given: that it is a short book by one author followed by a long book by another author, or that it is one long single work, fashioned as such by Jean, and provided with a vantage point in the middle that enables one to look over its full length from start to finish. (23) MS fr. 146 similarly provides two kinds of view of the extent of Fauvel. The original roman was divided into two books: this is retained in MS fr. 146, with an incipit and explicit on folio 11 and an author portrait. However, there is a rubric placed at the approximate mid-point of the roman as a whole which makes a different kind of attribution. It claims to mark the point at which Chaillou de Pesstain starts to make 'additions' to the 'livre': 'Ci s'ensivent les addicions que mesire Chaillou de Pesstain ha mises en ce livre, oultre les choses dessus dites qui sont en chant.' Moreover, it is immediately preceded by eight lines newly set into the narrative, in which Gerve's du Bus is riddlingly named as the author of the first roman:

   <g>clerc le roy francois, deRues,
   Aus paroles qu'il a conceues
   En ce livret qu'il a trouve
   Ha bien et clerement prouve
   Son vif engin, son mouvement;
   Car il parle trop proprement:
   Ou livret ne querez ja menconge,
   Dieux le gart! Amen. (ii, 41--48) (24)

In a manoeuvre seemingly copied from Jean's, Chaillou, like Gui de Mori, takes the opportunity, at the mid-point of the work, simultaneously to assert his own presence and pay homage to the author he is displacing. (25) This 'double authorial signature' (as I have called it elsewhere) signals not so much a trick this time, as a new compositional procedure. (26) Chaillou, from this point on, creates a generic change in the work by turning a satiric allegory into a love narrative with inset verse. Much the same (in reverse) could be said of the naming of Amans as Gower in Book viii of the Confessio: from this point on, Gower, having flirted with the figure of the love poet, reverts to satire. (27)

There is not space here to give full due to the compositional and visual ingenuity of the revised Fauvel: however, what can perhaps be granted from the above discussion is that MS fr. 146 presents a highly developed sense of a vernacular book as the product of an author. This, of course, is a radically new concept in the context of the scribal culture of production characteristic of the thirteenth century and earlier. The accepted story is that before Machaut, the first surviving 'single author codex' is the late-thirteenth-century Adam de la Halle manuscript (BN f. fr. 25566), with some recognition of the role played by chansonnier compilers in creating author based collections of songs. (28) Yet this account leaves out the MS fr. 146 version of Fauvel. (29) It is necessary to emphasize how fully in Fauvel, several decades before Machaut's works were first produced as author-collections, the author-figure was transformed into an agent of control over the material form of the book.

To talk of Jean de Meun and Chaillou de Pesstain as 'authors' needs some further comment. We have learnt, through the work of Alastair Minnis on scholastic theories of authorship, and Rita Copeland on the theory and practice of translation, how profoundly mobile notions of authorship were in the fourteenth century. There is always a difficulty in grasping how far Latin terminology and intellectual assumptions permeate vernacular writing, especially when vernacular writers, notably Chaucer, appear to spend much energy creating a distance between themselves and auctoritas. Translatio studii, as several scholars in French, and more recently Copeland have argued, turns out to be a key means by which 'academic modes of reading' are transformed 'into modes of discursive production'. (30) It is not just the terms auctor and auctoritas that are finding a vernacular meaning, but others such as compilator, collector, editor, scriptor, commentator, actor, assertor (Minnis, pp. 94--112), and persona. (31) Many of these terms are given vernacular equivalents in copies of the Rose and other thirteenth and fourteenth-century romans and dits. Both Jean de Meun and Chaillou are key figures in a process of developing vernacular authorship because they are instrumental in making a connection between the author and the book. Their work, moreover, with its ingenious apprehension of the text as existing in the material space of successive pages in a book, brilliantly opens up the notion of authorship to reveal it as a multitude of constituent elements. Authorship, in their hands and their scribes', is not merely a single, unifying conceptual framework, but a means of articulating the process by which texts make meaning. The process of translation results not merely in a transference of authority from Latin to vernacular, but the realization that authority is itself susceptible to being construed in a variety of ways. (32)

In the last part of this discussion of French contexts for the idea of authorship in Confessio amantis, I want to turn briefly to Gower's more immediate predecessors, Guillaume de Machaut, Jean Froissart, and the author of the Tresor amoureux. Much has been written about Machaut's construction of himself as a poe'te, and his role in the making of his own books of poetry and music. (33) There are twelve surviving manuscripts that contain a more or less complete collection of Machaut's works, seven with music, two with texts only and three that include works by other authors: several Machaut works are also copied out individually in numerous further compilations. (34) I am going to concentrate on two (interconnected) aspects of his books, the use of rubrication, and the Prologue to his works. Programmes of rubrication in French secular vernacular manuscripts are widespread: they range from speaker markers and song headings to extensive glosses and interpolations, some in Latin. A high proportion of Rose manuscripts give speech-markers for L'Amant and L'Auteur, as well as other characters such as Bel Acueil (for example, BN f. fr. 1569, fol. 23v) and Reson (BN f. fr. 1569, fol. 44r and 46r), (35) and this practice is taken up in the Machaut and Froissart manuscripts. As I have noted elsewhere, it is particularly interesting to find a distinction between 'guillaume' and 'l'acteur', indicating a separation between the named persona of the poet and the voice of the writer. (36)

The closeness of the use of speaker markers in Machaut to the rendition of dialogue in a dramatic work such as Le Jeu de Robin et de Marion shows that one important function of them is to point up the characteristically dramatic quality of Machaut's personification. The use of multiple personas within a first-person text is not merely a powerfully conceptualized procedure, but a means of giving utterance to distinct voices. The names (Franchise, L'Amant, La Dame, Doutance, and so on) can occur in the margins or else (more commonly) in the body of the text columns: in either position they are more than glosses, they are performative signs. Their dramatic character does not, however, cancel out their nature as types of rubric. At the risk of stating the obvious, it is important to remember that they have a double role: in identifying the speaker, they also serve as an interpretative framework allowing the text to be viewed as a set of distinguishable playing parts.

I want to emphasize this double character of the rubrics in turning to the layout of Machaut's Prologue. The Prologue, which announces Machaut as the author of his whole Luvre, is probably the most emphatic instance in Machaut's writings of his articulation of himself as author. This work takes the form of four paired ballades, followed by one hundred and eighty-four lines of narrative octosyllabic couplets. In the first pair of ballades, Nature commands Machaut to make 'nouviaus dis amoureus'. For advice and encouragement, she offers him three of her children: 'Scens, Retorique et Musique'. The second ballade takes the form of the poet's reply in which he duly accepts her charge. In the next pair, Amours offers three more children: Dous Penser, Plaisance, and Esperance as 'matere' for his compositions. He responds again accordingly, with gratitude and humility. The Prologue occurs complete in three late manuscripts (in chronological order, A (MS fr. 1584), F (MS fr. 22545) and Pm (Pierpont Morgan Library MS M.396)), and is preceded in A with an original index that is itself prefaced by the oft-quoted rubric: 'Vesci l'ordenance que G. de Machau wet qu'il ait en son livre' [Here is the arrangement that Guillaume de Machaut wants there to be in his book]. It seems, then, that Machaut newly composed the Prologue as part of a much larger project to present his writings as a whole, authored collection: indeed, as Lawrence Earp has remarked, 'in Machaut's last years, it was the organization and presentation of his life's works--rather than the composition of new works--that was his main preoccupation as an artist' (p. 73). (37)

As well as the rubric just cited, the Prologue is heavily laced with rubrics introducing each ballade. For example, the first is introduced as follows: Comment Nature, voulant orendroit plus que onques mes reveler et faire essaucier les biens et honneurs qui sont en Amours, vient a Guillaume de Machaut et li ordene et encharge a faire sur ce nouveaux dis amoureux, et li baille pour li conseillier et aidier a ce faire trois de ses enfans, c'est a savoir Sens, Retorique et Musique. Et li dit par ceste maniere. (38)

Such detailed rubrics have so marked a presence on the page that they form an extra prose layer within the mixed generic structure of the Prologue. Their function is multiple: for as well as being glosses or summarizing headings for each ballade, as a result of the topic of each ballade they are also a description of the organizing principles of the work, and thus work in tandem with the index. Even more than this, they also act as instructions to the illuminator, since each ballade is magnificently illustrated, followed by an author portrait at the start of the narrative section. (39) Finally, in the way that they name the author, 'Guillaume de Machaut', they function as highly elaborate speaker markers. They thus participate in a rhetoric of performance: they put the author's voice into the place of that of the compiler or scribe. This does not reduce the status of the author, but rather raises the status of the compiler. Moreover, the fact that each rubric is also an instruction to the illuminator means that it is not just commenting on the page, but making the page come into being.

Machaut's Prologue evidently acts as a defining moment for medieval vernacular authorship: a remarkable aspect of it is the weight Machaut places on rubrication as a means of defining his role. This deserves emphasis because often the activity of compiling is taken to contrast with that of authoring or authorizing. Much has been made, for example, of Jean de Meun's or of Chaucer's appropriation of the vocabulary of the compiler as a means of evading the responsibility of authorship. (40) Yet for Machaut, the opposite appears to hold: as Gui de Mori did through the Rose, he proclaims himself author through the very medium and language of the scribal compiler, and, though there is no space to demonstrate it, of the copyist. The Prologue to Boccaccio's Decameron provides an illuminating comparison. (41) Boccaccio defends his choice of stories at first by blaming their authorship on the ladies who told them, but then, in a sly manoeuvre, effectively puts himself forward as the 'inventor' after all:

But even if one could assume that I was the inventor (lo'nventore) as well as the scribe (lo scrittore) of these stories (which was not the case), I still insist that I would not feel ashamed if some fell short of perfection, for there is no craftsman other than God whose work is whole and faultless in every respect. (42)

A very similar claim, with the same kind of slippery bravado, is made in the early thirteenth century by Jean Renart in the Prologue to his own Le Roman de la Rose. It works in reverse, for Jean brings in songs by other trouve'res yet argues that they fit so well into his romance that people will assume he composed them. He is a compiler claiming, through his compiling, to be indistinguishable from an author.

In short, vernacular authors working in a French tradition are seeking authority, not by denying their right to claim it, but by exploring the many different forms of articulating such authority through, and by means of, the construct of the book. Actor and auctor, compiler and assertor, scriptor and recitor, these are all roles that begin to coalesce through the ingenious and imaginative inventiveness of the vernacular poet. We can see this in two more examples: the rubrics that introduce Froissart's two manuscripts of his collected love poetry, and the anonymous Tresor amoureux. Following directly in Machaut's path, Froissart's manuscripts (BN f. fr. 830 and 831) open and close with rubrics describing how the collection has been ordered: Vous deves scavoir que dedens ce livre sont contenu pluisour dittie et traitie amourous et de moralite, les quels sire Jehans Froissars, prestres, [...] a fais, dittes et ordonnes [...] Et vous ensagnera ceste table comment il sont escript ou dit livre par ordenance. (43)

Like Machaut, Froissart makes the rubric, the standard medium for communication between scribe or compiler and reader, a place to announce his own authorship. In fact, he takes Machaut's practice one step further by announcing the list of contents as his own work too. A Brussels copy (1410--15) of the Tresor amoureux (Bibl. Royale MS 11140), the first few folios of which are richly illuminated, is another close imitation of Machaut's Prologue. (44) Once again, the miniatures are often accompanied by lengthy rubrics that act both as instructions to the artist and glosses for the reader. (45) Here, through image and text, the work makes its own ordering its own subject: a large section of the work is taken up with a speech by Amours instructing the author on the way he should compose and write his book. (46) In the Tresor, the explanatory compiler's rubric, in the form in which Machaut used it, has now fully entered the work: it no longer introduces the work to come, it constitutes the work's narrative.

In returning, finally, to Gower, it might seem as if we have travelled some distance from Confessio amantis. Indeed, it has recently been argued that evidence of authorial status in English manuscripts before the fifteenth century is slight. Author collections, such as they are, 'reveal only intermittently any coherent sense of the identity of the author and even more intermittently any inclination to signal that in distinctive ways'. (47) Yet even in this context of apology for the English scene, Gower is recognized to be something of an exception. Some early interest in a collected Gower can be found, for example, in Oxford, Bodleian Library MS Fairfax 3, where the Confessio is put together with the English poem 'In Praise of Peace', the Traitie, and some Latin pieces. (48) And of course Confessio amantis itself (as many have pointed out), as a large compilation, functions in its own terms as a collected Gower.

In what way can the structure and layout (and the structure of the layout) of the Confessio be compared to the French books we have been considering? I would like to make various concluding observations. In terms of rubrication, particularly speaker markers, it seems clear that Confessio manuscripts, like those of Chaucer's Troilus and Criseyde and other fifteenth century English collections, are very much part of a developing tradition in French of giving dramatic voice to the different elements of a first-person narrative (whether rendered in lyric or narrative form). (49) None the less, it is striking how thoroughly the Confessio manuscripts present their structure through a Latin, rather than an English framework. One or two copies, as Syan Echard has discussed, translate the glosses into English, but these are the exception that proves the rule. (50) Unlike Machaut and Froissart, and other fourteenth-century French contemporaries, Gower does not explore the power of the rubric to create a growing recognition of the vernacular author in the vernacular.

It is Jean de Meun and the compilers of the MS fr. 146 version of Fauvel that provide the closer comparison in this respect. Yet even here, the similarities are not as extensive as one might imagine. The large-scale glossing of the Rose includes many examples of Latin annotation whereby the Rose is linked to a 'learned textual network' and indeed 'begins to take on the appearance of a vernacular florilegium of the auctores'. (51) None the less, these highly Latinate Roses are the product of the work's readers: Gower's decision to add a Latin layer to his own authorial compilation ranks as rare and distinctive even in the broader context of European vernacular writing. (52) The 169 interpolations in Fauvel include a large number of Latin pieces, including some newly composed Latin proses: (53) again, they provide a mixed comparison because many are drawn from other (usually anonymous) authors and composers.

One way in which the Fauvel interpolations can be brought to bear on Gower's Confessio concerns their teasingly floating existence on the page. While some are cued into the main roman narrative, many others are not. As Michael Camille has convincingly argued, this is part of a larger scheme within the work to invert the relation between centre and margins: in the fundamentally perverted world of the corrupt horse, Fauvel, attention is transferred from the centre to the edges of the page, and many tricks are played in and through the text, images, and music to make this satiric point. (54) It would be hard, and no doubt misguided, to make a case for a similarly deliberative mise en page in the Confessio: at the same time, it provides an intriguing context for the variable practice of the Confessio scribes in placing the Latin sometimes in the margins, and sometimes in the central columns. At the very least, Fauvel gives clear grounds for thinking that it was possible to attach some kind of polemical distinction to the positioning of material on a page, and hence that it may well be worth paying greater attention to variations in scribal practice of this kind.

The naming of John Gower, with which I began, is a case in point. Looking across the Confessio, apart from the admission of his name in English to Venus (viii, 2321), Gower is named directly only in a Latin gloss in the Prologue (22, 34*):

Hic in principio declarat qualiter in anno Regis Ricardi secundi sexto decimo Iohannes Gower presentem libellum composuit et finaliter compleuit, quem strenuissimo domino suo domino Henrico de Lancastria tunc derbeie Comiti cum omni reuerencia specialiter destinauit.

Hic declarat in primis qualiter ob reuerenciam serenissimi principis domini sui Regis Anglie Ricardi secundi totus suus humilis Iohannes Gower, licet graui infirmitate a diu multipliciter fatigatus, huius opusculi labores suscipere non recusauit, set tanquam fauum ex variis floribus recollectum, presentem libellum ex variis cronicis, historiis, poetarum philosophorumque dictis, quatenus sibi infirmitas permisit, studiosissime compilauit.

Like Machaut and Froissart before him, Gower makes use, not only of the Prologue, but also of the form of the explanatory rubric, to announce himself as author. Moreover, he takes the opportunity, especially in the MS Bodley 294 version, to make this claim by making larger reference to his collected works. 'Iohannes Gower' is defined not merely through this one work, but as a soubriquet that draws together a whole oeuvre.

In several manuscripts this gloss finds a response in a sequence of Latin statements, in verse and prose, (usually) placed after the final 'Explicit' verses. Here, Gower is named three times in a final flourish, in which his principal works are named, catalogued, and described. Gower, the author, thus gains articulation through a wide variety of locations on the page: the authorial gloss--sometimes presented marginally, sometimes in the central text column--the English narrative, and finally, in a sequence of additions to the main work that act as a retrospective commentary, even index. (55) All these utterances combine to give powerful expression to the authorial Gower. As in many French books, rubrics are used to signal authorial control; at the same time, the mechanism of annotation and commentary is partly internalized, a feature of the work which is implicitly recognized in those manuscripts that integrate the glosses into the central text. Gower appears to be experimenting with different locations for authorship. On the analogy of the Fauvel manuscript, there seems to be a desire to investigate the possibilities of meaning in these various sites. As a result partly of Gower's own decisions to convey his text in a variety of voices, rhetorical strategies, and languages, and partly of the decisions of those who planned the layout of the manuscripts, his presence as an author has many forms. In the light of French comparison, we can understand their plurality as part of a larger late-medieval cultural process in which all those involved in producing books--authors, scribes, compilers, and illuminators--are examining how authorship might be received and expressed. An important element in this process concerns the different kinds of weight carried by English, Latin, gloss, rubric, prologue, or colophon.

Gower emerges from this study of French precedent as remarkably, perhaps unexpectedly, original. As Alastair Minnis notes, this word, originalia, late in the twelfth century comes to designate 'the whole work of an auctor'. (56) Gower's use of Latin, far from being a sign of conservatism in any simple sense, seems rather a strikingly distinctive means of investigating the complex guises under which authorship was emerging in the books of vernacular writers. It is possible to understand it, in other words, not merely as a means of affirming his auctoritas, of lending gravity and cultural seriousness to his writing, but rather as a voice in a much larger dialogue, embracing vernacular as well as Latin, in which authorship is newly figured.

I am grateful to Derek Pearsall, Sian Echard and Nicolette Zeeman for kindly reading and commenting on a draft version of this essay. I am further grateful to Derek Pearsall for most generously sharing information on the Gower manuscripts from the catalogue currently in preparation by himself, Kate Harris, and A. S. G. Edwards, drawing on research by the late Jeremy Griffiths. Sian Echard, too, has been generous in sharing her work in advance of publication, in particular, her forthcoming essay 'Last Words: Latin at the End of the Confessio Amantis'.

(1) As a further example of the complexity of these identifications, Genius also 'functions as a projection of the "author", or as a voice of the "intentio auctoris" ' (Rita Copeland, Rhetoric, Hermeneutics, and Translation in the Middle Ages: Academic Traditions and Vernacular Texts (Cambridge: Cambridge University Press, 1991), p. 205).

(2) A. J. Minnis, Medieval Theory of Authorship: Scholastic Literary Attitudes in the Later Middle Ages (London: Scolar Press, 1984), and Magister Amoris: The Roman de la Rose and Vernacular Hermeneutics (Oxford: Oxford University Press, 2001); Copeland, Rhetoric, Hermeneutics, and Translation; David F. Hult, Self-Fulfilling Prophecies: Readership and Authority in the First Roman de la Rose (Cambridge: Cambridge University Press, 1986); Sylvia Huot, From Song to Book: The Poetics of Writing in Old French Lyric and Lyrical Narrative Poetry (Ithaca and London: Cornell University Press, 1987).

(3) The EnglishWorks of John Gower, ed. by G. C. Macaulay, 2 vols, EETS (1900).

(4) This last point takes the argument further than Burrow.

(5) Of course, several notions of frame apply to Confessio amantis: for example, the device of the lover's confession is in turn framed by the prologue and epilogue. On the Latin material, see Derek Pearsall's pioneering article, 'Gower's Latin in the Confessio Amantis', in Latin and Vernacular: Studies in Late-Medieval Texts and Manuscripts, ed. by A. J. Minnis (Cambridge: Cambridge University Press, 1989), pp. 13--25; The Latin Verses in the Confessio Amantis: An Annotated Translation, ed. by Sian Echard and Claire Fanger (East Lansing: Colleagues Press; Woodbridge: Boydell and Brewer, 1991); Echard, 'With Carmen's Help: Latin Authorities in the Confessio Amantis', Studies in Philology, 95 (1998), 1--40.

(6) It is worth noting that the verse also has headings, further obscuring the distinction between text and gloss. As Derek Pearsall has pointed out to me (in a private communication), one might even find a fifth Gower here: 'very smart, occasionally obscene, a would-be high Latin rhetorician'.

(7) In particular, Minnis, Medieval Theory of Authorship, and Copeland, Rhetoric, Hermeneutics, and Translation.

(8) Jeremy Griffiths, 'Confessio Amantis: The Poem and its Pictures', in Gower's Confessio Amantis: Responses and Reassessments, ed. by A. J. Minnis (Woodbridge: Boydell and Brewer, 1983), pp. 163--78; R. F. Yeager, 'English, Latin, and the Text as "Other": The Page as Sign in the Work of John Gower', Text, 3 (1987), 251--67; Dhira B. Mahoney, 'Courtly Presentation and Authorial Self-fashioning: Frontispiece Miniatures in Late Medieval French and English Manuscripts', Medievalia, 21 (1996), 97--160; Sian Echard, 'Glossing Gower: In Latin, in English, and in absentia: The Case of Bodleian Ashmole 35', Re-Visioning Gower, ed. by R. F. Yeager (Asheville: Pegasus Press, 1998), pp. 237--56, and 'Dialogues and Monologues: Manuscript Representations of the Conversation of the Confessio Amantis', in Middle English Poetry: Texts and Tradition. Essays in Honour of Derek Pearsall, ed. by A. J. Minnis (York: York Medieval Press, 2001), pp. 57--75.

(9) See WinthropWetherbee, who makes this point while also setting Gower in the context of Boethius, 'Latin Structure and Vernacular Space: Gower, Chaucer and the Boethian Tradition', in Chaucer and Gower: DiVerence, Mutuality, Exchange, ed. by R. F. Yeager (Victoria, BC: English Literary Studies Monograph Series, 1991), pp. 7--35 (p. 27).

(10) Minnis, Magister amoris, has begun the task of addressing 'vernacular hemerneutics' more directly by way of Jean de Meun's Roman de la Rose.

(11) See Jeremy Griffiths, 'Confessio Amantis: The Poem and its Pictures', and J. A. Burrow, 'The Portrayal of Amans in Confessio Amantis', in Gower's Confessio Amantis: Responses and Reassessments, ed. by A. J. Minnis (Cambridge: Brewer, 1983), pp. 5--24 (p. 12, n. 7). According to Griffiths, of the twenty-seven manuscripts with illuminations, just two (Princeton, Robert H. Taylor Collection and Philadelphia, Rosenbach Foundation, MS 1083/29) have author portraits as such, in the initial to Prol. 1. However, there are a further eighteen manuscripts with Confessor (and Amans) pictures (including four with blank spaces).

(12) Le Roman de la Rose, ed. by F. Lecoy, CFMA, 3 vols (Paris, 1965--70), ll. 10465--650. Guillaume is named and described at ll. 10496.., Jean at ll. 10535..

(13) Hult, Self-Fulfilling Prophecies, pp. 74--93. On the use of author portraits to mark the transition between authors, see Lori Walters, 'Appendix: Author Portraits and Textual Demarcation in Manuscripts of the Romance of the Rose', in Rethinking the Romance of the Rose: Text, Image, Reception, ed. by Kevin Brownlee and Sylvia Huot (Philadelphia: University of Pennsylvania Press, 1992), pp. 359--73. From her study of ninety-one pre-1400 manuscripts (there are around a hundred and forty altogether surviving), Walters finds author portraits in fifty-nine, that is 65 per cent, positioned either at the end of Guillaume's Rose or at the mid-point of the conjoined Rose. Many of those manuscripts that have no portraits none the less contain rubrication that makes reference to authorship.

(14) See the discussion of Gui de Mori below.

(15) Chaucer uses a similar tactic in the Prologue to the Man of Law's Tale, where Chaucer as a writer is discussed in familiar terms by the fictional Man of Law.

(16) Gui's work was first discussed in two articles by Ernest Langlois, 'Gui de Mori et Le Roman de la Rose', Bibliothe'que de l'Ecole des Chartes, 68 (1907), 249--71 and Marc-Rene Jung, 'Gui de Mori et Guillaume de Lorris', Vox Romanica, 27 (1968), 106--37, respectively. More recently, Hult, Self-Fulfilling Prophecies, pp. 34--55 gives a full and subtle account of Gui's authorial appropriations; on Gui as a reader, see also Huot, The Romance of the Rose and its Medieval Readers: Interpretation, Reception, Manuscript Transmission (Cambridge: Cambridge University Press, 1993), pp. 85--129.

(17) Ten manuscripts survive that contain material from Gui's version; unfortunately, of the two oldest and most complete copies, one has been lost. The reconstructed interpolated passage in question is given by Hult, Self-Fulfilling Prophecies, pp. 36--39.

(18) Hult, Self-Fulfilling Prophecies, p. 49.

(19) Fauvel Studies: Allegory, Chronicle, Music and Image in Paris, Bibliothe'que Nationale MS francais 146, ed. by Margaret Bent and Andrew Wathey (Oxford: Oxford University Press, 1998).

(20) Some debate exists about whether Gerve's wrote both books: according to one school of thought, the first book is anonymous. See Bent and Wathey, Fauvel Studies, pp. 12--13.

(21) Ardis Butterfield, 'The Refrain and the Transformation of Genre in Le Roman de Fauvel ', in Fauvel Studies, pp. 105--59 (pp. 105--06).

(22) It is possible, of course, that the work as a whole was not intended for performance.

(23) See Hult's interesting analysis, which includes discussion of the anonymous continuation of Guillaume's poem that precedes Jean's work in a small number of copies (Self-Fulfilling Prophecies, pp. 89--92).

(24) The riddle is most clearly spelled out in the shorter ('uninterpolated') version of the roman (Le Roman de Fauvel, ed. by A. Langfors, SATF (Paris, 1914--19)), where the author is named at the end of Book ii as follows:

   Ge rues doi .v. boi .v. esse
   Le nom et sournom confesse
   De celui qui a fet cest livre:
   Diex de cez pechiez le delivre (ll. 3277--80)

As Bent and Wathey explain, 'the words doi, boi, and esse in the first line are the spelt-out letter names of D, B, and S; the line may thus be read as 'Ge rues d.v.B.v.s' i.e. Gerve's du Bus, who was a clerk in the French royal chancery' (Fauvel Studies, p. 2). This colophon is replaced in the longer version by the eight lines given above. In the latter, the first initial is left blank, preceded by a miniscule g, correcting the d of derues. Langlois suggested that .I. ('Un') was intended to fill the blank, allowing the line to read 'Un clerc le roy francois, Gerues'. Bent and Wathey suggest that this may have been done as a means of further disguising the author. For further discussion and references, see Fauvel Studies, pp. 1--2, and n. 6).

(25) On Gui's riddling references to his name, and the different use he makes of verse and prose to convey his identity, see Hult, pp. 47--53.

(26) Butterfield, 'The Refrain and the Transformation of Genre in Le Roman de Fauvel ', p.109.

(27) Though, as I argue in 'French Culture and the Ricardian Court', in Essays on Ricardian Literature in Honour of J. A. Burrow, ed. by A. J. Minnis, Charlotte C. Morse, and Thorlac Turville-Petre (Oxford: Clarendon Press, 1997), pp. 82--121, this does not mark an ultimate rejection of love poetry per se.

(28) See, for example, Huot, From Song to Book, pp. 64--74. Huot argues (p. 72) that the mid-point of MS fr. 25566 is also significant (this turns out to be the Jeu du Pelerin, where Adam is named). While this suggestion is full of interest, some other examples of her discussion of mid-points in manuscript compilations are less convincing.

(29) Gautier de Coinci is a further figure of considerable (and neglected) importance in the construction of vernacular authorship. I discussed the formative role of Gautier in an unpublished paper, 'Imagining the Author in French and English Books', delivered at the 'Imagining the Book' Conference, Queen's University, Belfast, April 2001.

(30) Copeland, Rhetoric, Hermeneutics, and Translation, p. 150.

(31) St Bonaventure uses many of these terms in a famous passage from the Prologue to his commentary on the first book of Peter Lombard's Sentences. It is cited in Latin (with translation) in Hult, Self-Fulfilling Prophecies, p. 61, n. 99, and in translation only, Minnis, Medieval Theory of Authorship, p. 94.

(32) It does not seem to me to be surprising, then, that the ascription of authority to the vernacular in the later medieval period appears as an obscure process, rather than a gradual glorious flowering (pace Minnis, Magister amoris, pp. 317--18).

(33) See, in particular, Brownlee, Poetic Identity in Guillaume de Machaut (Madison: University of Wisconsin Press, 1984); Huot, From Song to Book, Chapters 8 and 9.

(34) Earlier descriptions are now superseded by Lawrence Earp, Guillaume de Machaut: A Guide to Research (New York and London: Garland, 1995), pp. 73--128. The preponderance of 'complete oeuvres' compilations is an innovation: no French author before Machaut had so large-scale and carefully orchestrated a production of his poetic works, and hence of himself as a poet, although, as I have noted above, Adam de la Halle and especially Gautier de Coinci are important precedents. My comments here are drawn from eight of the 'complete' manuscripts.

(35) Huot reports finding 'Aucteur-Amant' rubrication in fifty-three out of seventy-two manuscripts: there are around a hundred and forty altogether surviving today from before 1400 (From Song to Book, p. 91, n. 18, and Appendix A).

(36) For example, Paris, Bibliothe'que Nationale MSS fr. 1587 (Le Jugement dou Roy de Behaingne and Le Jugement dou Roy de Navarre), fr. 9221 and fr. 22545 (Navarre). See Ardis Butterfield, 'Mise en page in the Troilus Manuscripts: Chaucer and French Manuscript Culture', Huntington Library Quarterly, 58 (1995), 49--80; repr. in Reading from the Margins: Textual Studies, Chaucer, and Medieval Literature, ed. by Seth Lerer (San Marino, CA: Huntington Library, 1996), pp. 49--80.

(37) For a summary of the arguments concerning Machaut's personal supervision of the copying of his works, see Earp, 'Scribal Practice, Manuscript Production and the Transmission of Music in Late Medieval France: The Manuscripts of Guillaume de Machaut' (unpublished doctoral dissertation, Princeton University, 1983).

(38) 'How Nature, wishing more than ever before to make known and exalted the goods and honors pertaining to Love, comes to Guillaume de Machaut, ordering and charging him herein to compose new poems about love, entrusting to him, for counsel and aid in this enterprise, three of her children, to be specific Meaning, Rhetoric, and Music. And she speaks to him thus.' Cited in Earp, Guide, p. 145 (with translation taken from Guillaume de Machaut: The Fountain of Love (La Fonteinne Amoureuse) and Two Other Love Vision Poems, ed. and trans. by R. Barton Palmer (New York and London: Garland, 1993), p. 3).

(39) See Earp, Guide, pp. 145--47. On the iconography of the portraits, see Francoise Ferrand, 'Les Portraits de Guillaume de Machaut a' l'entree du Prologue a' ses oeuvres, signes iconiques de la nouvelle fonction de l'artiste, en France, a' la fin du XIVe sie'cle', in Le Portrait, ed. by Joseph-Marc Bailbe (Rouen: Publications de l'Universite de Rouen, 1987), pp. 11--20.

(40) See Minnis, Authorship, p. 209.

(41) For a very interesting argument connecting the authorial use of prose rubrics in Boccaccio's Teseida with Chaucer's narrative divisions in Troilus and Criseyde, see Phillipa Hardman, 'Chaucer's Articulation of the Narrative in Troilus: The Manuscript Evidence', Chaucer Review, 30 (1995), 1--23.

(42) Cited in Minnis, Authorship, p. 204.

(43) 'You ought to know that in this book many dits and treatises of love and morals are contained, which Sire Jean Froissart, priest [. . .] has composed, rhymed and arranged in order [. . .] And this table will teach you in what order they are written in this book.' Le Paradis d'Amour; L'Orloge Amoureus, ed. by Peter F. Dembowski, TLF (Geneva: Droz, 1986), p. 2 (my translation).

(44) C. Gaspar and F. Lyna, Les Principaux manuscrits a' peintures de la Bibliothe'que Royale de Beligique, 2 vols (Brussels: La Bibliothe'que, 1937; repr. 1984), i, 453--54, Notice 191.

(45) For further discussion of this kind of rubric, see J. J. G. Alexander, Medieval Illuminators and their Methods of Work (New Haven, CT and London: Yale University Press, 1992), Chapter 3, also pp. 117--18.

(46) For further discussion, see Butterfield, 'Mise en page', pp. 69--70. One of the marginal rubrics takes up this instruction as the heading for a picture: 'Comment amour deuise a l'auteur la maniere de faire son livre' (fol. 13).

(47) A. S. G. Edwards, 'Fifteenth-Century Middle English Verse Author Collections', in The English Medieval Book, ed. by A. S. G. Edwards, Vincent Gillespie, and Ralph Hanna (London: British Library, 2000), p. 109.

(48) There are forty-nine complete or near complete Confessio manuscripts (these include six not known to Macaulay). Of these, a large majority (thirty-one) contain satellite material such as the Traitie (this occurs in eight), and varying numbers of Latin pieces and prose rubrics. There are seven manuscripts like MS Fairfax 3 that form more substantial compilations by virtue of including six or more such additional pieces. I am indebted to Derek Pearsall for this information.

(49) On speech-markers in the Confessio manuscripts, see Echard, 'Dialogues and Monologues'; on the French tradition and its connections with English manuscripts, see Butterfield, 'Mise en page', pp. 67--68, 71--77. Space does not permit me to expand on the comparison between the layout of the Confessio manuscripts and those of French examples: in general, the Gower manuscripts strike me as more academic in layout than those of Machaut and Froissart. For example, speaker markers are mostly keyed in from the margins, rather than set on a separate line within the text column as they tend to be in the Machaut manuscripts. CUL MS Dd.8.19, however, does follow the latter practice (markers begin from fol. 10r).

(50) Echard, 'Glossing Gower'.

(51) Huot, The Romance of the Rose, pp. 48--75 (p. 55).

(52) For examples, see Minnis, Magister amoris, pp. 312--19, referring, inter alia, to Dante's Convivio, Boccaccio's self-commentary on his Teseida, Evrart de Conty's (French) commentary on the Eschez amoureux, and the Castilian and Latin commentaries that accompany various late-medieval collections of Castilian and Spanish poetry.

(53) Prose in this context means a text drawn more or less closely from biblical or liturgical sources and set to chant.

(54) Michael Camille, 'Hybridity, Monstrosity and Bestiality in the Roman de Fauvel ', in Fauvel Studies, pp. 161--74.

(55) Two manuscript and two early printed copies of the Confessio are supplied with tables of contents. See Echard, 'Pre-texts: Tables of Contents and the Reading of John Gower's Confessio amantis', Medium AEvum, 66 (1997), 270--87.

(56) Minnis, Authorship, p. 156.

ARDIS BUTTERFIELD

University College, London