CREATING MS C: AUTHOR, WORKSHOP, COURT

Machaut's set of complete works manuscripts forms a central pillar of our understanding of musical and generic developments and their courtly reception in fourteenth-century France. By applying the continuing scholarly advances made during the study of courtly practice and the professional Parisian book-trade to the earliest of these artefacts, this contribution reassesses the creation-history of the manuscript Paris, Bibliothèque nationale de France, fonds français 1586. The results tap into a number of enduring discussions within Machaut scholarship. These range from questions of patronage, to aspects of Machaut's authorial control and involvement in the production of his books, to the importance of order on the single-work level within a generic grouping, and to the practicalities of manuscript creation and intentionality. Finally, proposed adjustments to the dating of some compositions call for a review of existing notions of generic development and polyphonic composition in the early part of the century, thus resonating beyond Machaut's personal output.

to collect his own works into professionally executed presentation manuscripts. 1 Without these collections, our knowledge of his musical capabilities would be negligible and our understanding of wider chronology and stylistic development much poorer. 2 While some of his literary reputation would have remained, seminal works such as Le Livre dou Voir dit and La Prise d'Alexandre would not have come down to us. Without such works, we would have a poorer understanding of other authors' inspirations, creative practices and cultural influence. 3 The standing of artists such as the 'Master of the Remède de Fortune' and 'Master of the Bible of Jean de Sy' would undoubtedly have remained high due to their contributions to other books, but the appreciation of the links between them and perhaps even of the quality of their overall output may have been diminished. 4 From the medieval point of view as well, it seems clear that the physical book played an important part in establishing Machaut's works as cultural capital worthy of collection by both the highest aristocracy and the courtier class. 5 Thus, any adjustment in the dating of these sources or in the narratives that surround their creation has wide resonanance, both in modern scholarship and in the way we understand our creative past.
This article concentrates on one of these books, namely, Paris, Bibliothèque nationale de France, f. fr. 1586 (hereafter C). It discusses a missing element in current narratives surrounding this manuscript's creation, that is, the role, formation and use of exemplars. 6 The position of MS C as Machaut's first attempt at creating a 'complete works' manuscript is by now well established. 7 My interest here is in the circumstances and the process of bringing such a 'first' into being: the practicalities involved in carrying out such a project as well as the social reasoning behind it. Recent decades have seen a stepchange in our understanding of the kind of professional, Parisian book-production process with which this source is associated. 8 Still, there are multiple lacunae in Machaut's biography during the 1330s and 1340s and none of his exemplar materials, records or receipts survives. 9 Making assertions about practicalities and reasoning involves dealing with motives and with the myriad accidents and coincidences of daily life. This is inherently speculative. Still, we know the book was ordered and delivered, that exemplars existed, and that they were a central tool in the operation of the professional Parisian book trade. Ignoring their influence on the book-production process serves only to obscure the limiting factors involved in such undertakings, not to compensate for or resolve their loss. Thus, speculation based on known processes is more likely to yield interesting insights than speculation devoid of context. In this contribution, I will endeavour to suggest a reconstruction of these materials as they pertain to MS C and to assess the insights gained from looking at this level of production.
I begin with an examination of our characterisation of patronage, authorial control and the practices through which workshops engaged with external entities, especially in relation to the creations of exemplars and their use. Two commissioning models will be consideredone centred on the patron, the other on the authorwith the likelihood and implications of each assessed against our knowledge of Machaut's biography and against the general characteristics of MS C. Opting for the model of author as commissioner, the discussion moves to a close examination of the contents of this source and how they relate to decisions made during copying and illumination. Special attention is given to irregularities and difficulties apparent in this process, as these points are most likely to attest to the state of the exemplars from which scribes worked. I suggest that these anomalies can be explained as a result of the copying process itself, without recourse to external influences and coincidences. This shift, in turn, calls for adjustments in the narratives we construct around this manuscript's history. Finally, these various strands are brought together to form an alternative narrative for MS C's creation, assessing its implication for the ways we understand Machaut's social and creative practices, and with that, for wider generic development.
Were this narrative to be accepted, it would go beyond enhancing our concept of the process of creating manuscripts: it would call for 9 For a recent re-evaluation of biographical evidence, see E. E. Leach changes to our understanding of author-patron relationships as they manifest themselves through artefacts; it would reopen questions pertaining to authorial involvement, control and the meaning of order in manuscript collections; it would redate the emergence of some compositional practicesmost notably proposing that Machaut had begun setting rondeaux to polyphony and had used three-part textures in vernacular song composition before 1350and it would revisit the evaluation of differences between multiple copies of equivalent materials. While Machaut's case is unique, his central position in assessing wider practice in fourteenth-century France enables such outcomes to have broader reverberations. The key contribution that musical copying brings to this discussion should serve to bring musicological, codicological and literary analysis closer together. First, however, it is necessary to present the currently accepted narrative surrounding this source.

M S C A N D T H E P L A G U E : C U R R E N T H Y P O T H E S E S
It is hard for anyone looking at MS C not to marvel at its beauty and richness. 10 Still, a closer look suggests that something did not go to plan during its creation. Towards the end of the collection a set of generically-mixed lyrics set to music has been copied in what seems a hasty and rather ill-fitting manner. 11 Furthermore, a brief codicological inspection reveals that the first 120 folios of the manuscript follow a clear practice of deviating from the use of quaternions at the end of each narrative unit in order to match poetic and structural units (see Table 1). 12 The pattern changes in its remaining 105 folios, with the lyric poems exceeding the gathering structure seemingly pre-prepared to house them and the generic units of the musical section being copied through continuously with no reference to the manuscript's physical structure. The adjustments to gatherings XVIII and XXVI suggest that this was not due to an early and consistent decision to change copying procedure at this point.
I will discuss these characteristics in detail below. For the moment, it will suffice to say that they have been taken to represent evidence of a break in this manuscript's copying process: one which requires external, patron-focused narration. The accepted context for these  King John II of France), who, it is claimed, may have had a hand in the original commissioning. 13 As a royal commission, the irregularities in the manuscript's structure are explained as a result of Bonne's untimely death. This separates a CI layer (fols. 1-186, and perhaps also fols. 211-25), designed to appeal specifically to Bonne and copied before September 1349, from a much smaller CII layer (fols. 187-210) added later, in time for a completion date before 1356. This divide assumes a neat terminus ante quem and terminus post quem for the contents of the two sections. For example, as all the rondeaux set to music in this source appear in CII, it is currently assumed Machaut first began interacting musically with this form in the early 1350s. 14 The apparent lack of organisation in the supposed post-Bonne additions is taken to evince Machaut's lack of involvement in the final stage of the work. Perhaps surprisingly, no specific patron-based explanation has been forthcoming for the decision to complete the manuscript in the early 1350s, or for the gap between it and Bonne's death. There is no consensus as to who funded MS C's completion, the exact state it was in when work ceased, or the origins of the added material. 15 13 This narrative was first suggested in U. Günther, 'Der musikalische Stilwandel der französischen Liedkunst in der zweiten Hälfte des 14. Jahrhunderts' (doctoral thesis, University of Hamburg, 1957), and has been further developed by Earp, Avril, Wimsatt, Kibler and others. A useful summary of the changing approaches to this source and of the circumstantial basis of much of the evidence on which the current reconstruction is based can be found in Leach At this point, it is worth stating that considerable lacunae exist in Machaut's biography, and that patronage can range from longstanding and consistent provision of professional services to a single, indirect point of contact. Other than his close association with John of Luxembourg up to perhaps the early 1330s and more settled residency in Reims from the mid-1350s onwards, we have no archival evidence for Machaut establishing exclusive, long-standing relationships with any known patron. 16 He certainly dedicated works to individual patrons and had dealings with others; the recurrence of such evidence suggests closer relationships, but this remains speculative. For example, the dedication of his Jugement du roi du Navarre (c. 1349) and Confort d'ami (c. 1357) to Charles of Navarre, together with the survival of a document referring to a gift of a horse from Charles to Machaut in 1361, testify to a long-standing relationship. Still, in itself, such evidence does not necessitate a decade of close proximity and continuous contact between the two men. Works can be dedicated from afar and may indicate initial contact rather than an established relationship. The situation with Bonne is even more precarious as Machaut only refers to her specifically in his Prise d'Alexandrie, written some twenty years after her death, and no archival documentation linking them survives. Thus, scholars who identify her as the dedicatee of the Remède de Fortune, of the supposed first conceptualisation of the Navarre, and as the moving force behind Machaut's monophonic virelais, consider her to be a major patron of the utmost importance while relying on problematic evidence. 17 Simultaneously, scholars with other interests can claim that 'the traces of Machaut's association with [Bonne] are so slender and insubstantial that no such hypothesis appears very credible, at least on the evidence currently available'. 18 In what follows, I do suggest that Machaut's circumstances in the mid-1340s would have led him to consider Bonne as an obvious potential patron. This, however, does not imply such a relationship necessarily materialised. While some of my concluding remarks suggest that official acceptance of service may have occurred before Bonne's death, this is not central to my analysis for a simple reason: neither author nor patron exists independently of a practical context. The story, therefore, cannot only be about them.
As we have seen, the accepted narrative surrounding MS C is primarily patron-focused, at times also incorporating the author's wishes, but for the most part avoiding the practicalities of the bookproducing workshop. The considerable amount of excellent codicological analysis based on it has thus been aimed at dating the copying work, or at identifying in the final product traces of Machaut's or of his patrons' momentary wishes and needs. 19 Middlemen (and perhaps women) and the preparatory work their involvement necessitated of the commissioner rarely feature. Close interrogation of this stance raises a number of other inconsistencies. Assessing those requires a closer codicological look at the manuscript, but not before some clearer definitions are offered for the author-patron relationship and for the way each of them interacted with book-producers and their representatives. This is where we turn next.
It is becoming increasingly clear that some of the modern shorthand terms describing the process of manuscript creation are rather problematic in practice. John Lowden, for example, exposes the fluidity behind terms such as 'patron', 'workshop' and 'iconographer', even in cases where a manuscript's commissioning history is relatively well documented. 20 We know that a number of the artists engaged in the illumination of MS C cooperated elsewhere, and, as a group, they have been associated with service to the royal household in Paris. 21 Still, my use of the term 'workshop' is both wider and less institutional than these facts may suggest. I take it to refer to the non-binding or necessarily geographically and temporally concurrent assemblage of professionals charged with executing any element of a manuscript as distinct from author, patron or owner. For convenience's sake (and following the likely procedure in actual circumstance), it is imagined that the interaction with this vague grouping was funnelled through a representative subcontractor associated with the term libraire. 22 The level of execution, materials used, division of labour and quality of miniatures in MS C suggest a highly professionalised workshop. 23 Choosing such an entity to create one's manuscript had distinctive advantages, but also some specific practical implications. As is well documented, manuscript creation was a long and arduous process, often requiring a number of years from start to finish. 24 Indeed, the rationale behind organising professional workshops such as the one that produced MS Cwith their established procurement arrangements and stable, efficient and specialised workforce working simultaneouslywas to reduce the production period with minimal loss of quality. 25 Such multi-actor arrangements required a contractual basis and their smooth operation relied on clarity of plans and materials. A workshop could thus legitimately place certain demands on any prospective commissioner of manuscripts.
Initial negotiations would have involved the libraire assessing the timescale and price of the work envisaged. For this, he or she required an overview of the contents of the project and preferably its order. Nontextual elements would be particularly important. Historiation and illumination required the greatest investment of time and resources: illuminating a ten-line-high miniature clearly involved more expensive materials and would have taken longer to execute than copying ten lines of text. Illumination also meant disassembling copied gatherings and allowing for paint to dry before reassembling them. 26 Music created particular challenges in terms of layout and required specialist scribes. 27 Other essential parameters included the size of the 22 For a demonstration of both the fluidity between direct employment and the use of a 'project manager' even within a single commission, as well as for the use of a libraire by commissioners residing in Paris full-time, see R. H. Rouse, 'Pierre le Portier and the Makers of the Antiphonals of Saint-Jacques', in Croenen and Ainsworth (eds.), Patrons, Authors and Workshops, pp. 47-68. 23 The time-frame of this production period is discussed further below. For a partial division of labour, see Table 1  intended manuscript, its writing block and the quality of its materials and visual programme. 28 In practice, the more unusual the copied materials (in terms of contents, specific programming and arrangement), the greater the importance of clear and well-ordered exemplars becomes. Late adjustments to the plan were possible and exemplars could be sent in tranches, but this would have cost and time implications, making it harder for the libraire to plan ahead or to orchestrate simultaneous copying and illumination. It could also potentially spoil the appearance of the manuscript or lead to some materials being copied a second time. 29 All in all, this was best avoided, though it undoubtedly occurred at times. Due to the need to make sure that the workshop obtained all the materials it needed and knew how to reproduce them correctly, a commissioner's involvement was most intense at the beginning of the process. The multiple technical challenges offered by MS C made such involvement essential, though authors commissioning first presentation copies of their own work necessarily constitute the most extreme cases of such involvement. 30 Once decisions were made and exemplars supplied, the commissioner's role diminished. Their opinion may have been sought whenever problems arose, or they may have wished to interfere with the process in order to expand, delay or hasten it along, 31 but aside from funding payments they were no longer essential to the workshop's activities and the two entities did not have to maintain direct, regular contact. 28 Recent analysis has suggested that, perhaps uniquely to C's version of the Remède de Fortune, Machaut may have placed very specific demands in respect to these elements. In discussing the notion of 'commissioner', it is worthwhile separating it from both 'patron' and 'author'. It is clear that in some circumstances it is possible to conflate two or three of these terms. Still, each carries a distinctive role within the creation process. With the 'author' role I associate the responsibility for providing the materials to be reproduced: the contents of the book. Clearly, the actual author of the works contained in any manuscript can be entirely absent. 32 Still, within the professional workshop's set-up, it is not the scribe's responsibility to choose and order the works to be copied. He or she requires an external 'author', be that a literal one, a compiling superior in the project's hierarchy, or an external provider of content. The 'patron' denotes the perceived recipient of the finished object, and, most likely, its eventual funder. Clearly, some presentation manuscripts have been created 'in expectation' of patronage before a specific recipient was chosen, or they have been created in such a way as to appeal to multiple potential recipients. 33 In such cases a degree of financial risk is taken by an intermediary, identified as the 'commissioner'. In this sense, the 'commissioner' is the person on the ground who is involved in negotiating with the libraire, specifying decisions and delivering the exemplar materials. It is clear that a commissioner can also act (in whole or in part) as both author and patron, be associated with one or the other function, or exist independently as entrepreneur or agent. When entrepreneurialism is concerned, the commissioner can also double as a workshop professional. Clearly, lines can easily be blurred. As far as the workshop is concerned, for example, the creation of a book containing an original work destined for the patron who commissioned its composition translates into an author-commission: it is he or she who will be responsible for providing the materials to be copied, decide on the format of the book and come up with its visual programme, even if these actions are undertaken under time or financial constraints imposed by the copied work's patron. Patronising a work of literature is not the same thing as patronising its manuscript copy. 34  When Charles V commissioned and paid for a translation, he acted as the work's commissioner in relation to its author. Commissioned authors, though, maintain their positions as the commissioners of the books in which their works were presented to their patron. As such, they can justifiably expect to be recompensed a second time upon producing the object of the book. The resulting conflations between patron and circumstances, the difference between the patron-commissioner and author-commissioner model relates mostly to whether decisions are dictated by a manuscript's eventual recipient, or whether they more cautiously rely on a pre-emption of the elements patrons are likely to find appealing. But how does this relate to MS C?
The patron-commissioner model as it is currently narrated raises many practical questions regarding MS C's process of creation: Why was it commissioned? Why then? If indeed a break occurred and Machaut had no involvement with its completion, why would he have allowed a project in which he seems to have invested so much slip out of his control? From where and how did the late materials arrive at the workshop? Was the final decision concerning the manuscript's contents and ordering more likely to have been taken by a particularly involved (though not the original) patron, or by the workshop producing the source? If Machaut designated an original order, was it followed? If not, why? If it was completed as a memorial to Bonne by her widower or orphan (the Dauphin Charles has been suggested), 35 why were the works in this source arranged so as to privilege her father? What triggered the decision to complete the source in the first place? What would have caused a patron to demand the presumed speed in which the work was completed (a characteristic used to explain the oddities of presentation in the CII additions)? After all, the need for the manuscript to be finished by 1356 so that John II could take it to England during his captivity and expose Chaucer to its contents could not have been foreseen. 36 The author-commissioning model also raises a number of questions. Somesuch as 'why?' and 'why then?'are very similar. Others, such as affordability, remain in the realm of patronage. Beyond this, however, they begin to diverge. As continuous authorial involvement and direct contact with the workshop can be taken for granted, the irregularities and inconsistencies within the manuscript have to be explained as a result of the creation process itself, rather than any external intervention.
author-commissioner relationships makes much of the discussion of Machaut there irrelevant to this study. 35 See Earp, Introductory Study, pp. 30-1. 36 See n. 15 above.
Answering both sets of questions requires a degree of speculation and while logical, well-founded explanations offer consolation to scholars, experience teaches us that this seldom happens. As we are dealing with creative decisions and luxury objects, some emotive and ephemeral decision-taking is to be expected. We have seen that in practice, the two models may not be as far removed from one another as one might first expect. It is hard to imagine Machaut not acting as the commissioner, regardless of whether he fulfilled that role at the behest of a specific patron's order or undertook the project of his own accord, expecting to find a patron at a later point. Still, in what follows I attempt to demonstrate that a shift away from the patron and towards practicalities defuses certain questions and helps answer others, placing us on a more secure methodological and logical footing when reconstructing this manuscript's early history.
I would argue that the contextual and physical characteristics of MS C do not support a mere agent role for its content's author. As his first collection, Machaut must have been the source for some if not all of the materials to be included, probably forcing upon him some intensive preparation work beyond simply handing over his personal archive. 37 While single works undoubtedly circulated independently, collecting and collating as many of them as possible (keeping in mind that they were written for different patrons at different times and places) would probably not result in what retrospectively seems a complete collection. 38 Furthermore, MS C contains newly codified generic structures, the latest in music notational techniques, an intimate relationship between both text and music and between these and the illumination programme. 39 All this would have required an 37 On the shape of these materials and the preparation work required for them to function as successful exemplars, see below. intensive initial contact between libraire and commissioner and, preferably, the provision of particularly good exemplars. MS C does not present any obvious signs of specific, aristocratic commission: it sports no dedication, presentation portrait, motto or coat of arms in its presentation. All the same, in other contexts such characteristics have been taken to suggest royal intimacy rather than distance. 40 Our understanding of the patronage of this source relies on a combination of the careers of the artists who worked on it and the arrangement of the works it contains. Even when set up, the poetpatron relationship seems expectative rather than pre-ordained: the book does not show signs of being externally commissioned by a patron, but rather of an expectation it would find a willing audience ready to celebrate (and compensate) the poet for his efforts once presented with the finished artefact.
In case the funding of such a project should be deemed beyond the means of a cleric in Machaut's situation, one should note that an author-commissioner could have ended his or her involvement with a manuscript before any prohibitively expensive illumination took place. In order to ensure the correct ruling and layout for the text and musiclet alone estimate the price of the project as a wholethe illumination programme had to be agreed early on in the production process, probably as part of the initial negotiations. Instructions to the artists could have been kept by the libraire for a long while, separating the copying from the illumination bill, the latter being potentially settled by a later patron. 41 Even without this proviso, Machaut's inability to afford such an investment is not beyond challenge. Our best price comparison is the manuscript Paris, Bibliothèque nationale de France, f. fr. 167, a copy of the Bible moralisée made for King John II of France in the early 1350s and illuminated by some of the same artists who worked on MS C. 42 The cost of its illuminations amounted pp. 260-73; on musical and underlay innovations in this source as well as its wider context, see to some 320 livres at a rate of about 1 livre per working day: well beyond Machaut's means. This task, however, involved executing 5,112 illuminations as compared to MS C's 107. The pigments, gold leaf and occasional larger illuminations of MS C surely made it more expensive per image than the grisaille of the Bible moralisée, and both cases would have involved many additional costs such as buying parchment, ink, copying and binding. Nonetheless, a proportionate reduction in the expense of illumination suggests that such an undertaking may have been within the scope of a cleric enjoying a stable annual income totalling 100 livres. 43 Even if we assume that each of MS C's illuminations were three times dearer than their Bible moralisée counterparts, the total price of this element of the production would still amount to around 20 livres. A substantial investment, but not unmanageable.
Let us now consider Machaut's situation at the point of commissioning his first collected-works manuscript. The dating of the illuminations, the dits it contains and the time it would have taken to create such a manuscript combine to suggest that the decision to commit the necessary time and capital was probably taken in the second half of the 1340s. 44 Machaut would have been in his mid-forties. He was clearly an established musical and literary figure with links to the ruling dynasties of France and Bohemia, and enjoyed a secure income from two prestigious benefices. 45 This was, nevertheless, a period of instability in Machaut's artistic career, following the loss in 1346 of his most important, long-standing and intimate patron, John of Luxembourg, King of Bohemia. Furthermore, up to this point considerable portions of his adult life had been spent travelling with John's court. As a result, it may well be that many of his works did not circulate in his home country and that his personal relationships with the wider French nobility were perhaps not as strong as he would have liked them to be. Roger Bowers has suggested that regardless of Machaut's whereabouts in the decade before John's death, he remained officially affiliated with the King of Bohemia until 'upon the dissolution of his [John of Luxembourg's] household Machaut would have needed to find alternative work with some other employer. Certainly it appears that at this juncture he did not resort to the role of a canon residentiary of Reims.' 46 He may well have moved away from John's inner circle a good number of years earlier, and wasaccording to Lawrence Earp but contrary to Bowers's opinionperhaps more or less permanently attached to Bonne's court by this point. 47 If the Remède was indeed written for her, a link of some kind must have existed: it is not too fanciful a notion considering Machaut's services to her father and the links between the Luxembourgs and the Valois. They already belonged to an established network, in which cultural artefacts played a part. Nevertheless, we have seen that the surviving documents from this period link Machaut with John's and no other specific patron's business. At least officially, he seems to have remained 'John's man' until the king's death, even if from a distance. If this was the case, he would have needed to attach himself to a new court in a more official capacity at this point in his career, be that with Bonne and John of Normandy, as Earp suggests, or with Charles of Navarre, as Bowers proposes. The associations of single works cannot be seen as decisive evidence here, as they could have been written while in John of Luxembourg's service. Still, as the proposed dedicatees of his earlier large-scale narratives point towards Bonne (Remède) and her eldest son or court (l'Alérion), this would have been a reasonable first port of call when looking to establish a new, official allegiance. I tend to support this association from the point of view of Machaut's likely prospects at a given point of time and set of circumstances, regardless of whether it materialised. Seeing the book as a means to acquire patronage makes sense of the object, regardless of the result once finished, and its lack of stated affiliation echoes this function. Yet many questions remain, not least the possibility that the same function relates to a different point in Machaut's career. 48 Why should Machaut have chosen this precise moment to take upon himself the complicated and costly project of collecting his complete works in a sumptuous presentation collection? There are, of course, any number of possible explanations. For example, Earp suggested as one such reason a heightened sense of mortality in the face of the ravages of the Black Death. 49 To my eyes, however, the most likely option would be the potential use of such an artefact 46 Bowers, 'Guillaume de Machaut', p. 10. 47  in cementing a new patronage relationship. A certain amount of hubris can perhaps be pinned to a decision to create a complete-works manuscript rather than a presentation copy of a single work, especially as Machaut would have been hard-pressed to find a model for this kind of commission. 50 Furthermore, the Remède de Fortune, with its musical interpolations and didactic structure, would have been ideal for such a purpose. 51 As Froissart's reading of his Méliador to the Count of Foix makes clear, works such as the Remède could have been used in this capacity regardless of their original dedicatee. 52 Still, one can assume Bonne already had access to a good number of Machaut's works, including the Remède. 53 If she was first on the list of potential new patrons, the Remède alone would not do as an introduction to his abilities. More generally, his previous professional position in Bohemia, away from the Francophone heartlands, may have caused him to believe that a fuller résumé of the extent and versatility of his abilities would help his search for a new patron, especially if Bonne were not to prove forthcoming. 54 Thinking along these lines, it would have been a pity not to be able to demonstrate both the impressive bulk and the variety of his work. After all, any dedicatee interested in cultural production would have appreciated his specific achievements in narrativewriting inspired by the Roman de la rose, as well as the special case of the Remède and his cutting-edge preoccupation with both lyrical 50 On the development of author-controlled, complete-works books in the thirteenth and fourteenth centuries, see Huot, From Song to Book, pp. 211-42. 51  and musical fixed forms. In this scenario, speed would have been important but not the only objective. After all, Machaut already had links with powerful potential new patrons as well as an independent, stable income. More importantly, such a decision would have required him to have organised his works into at least rough, initial exemplars. Had he not done so, meaningful negotiations with a libraire would have been impossible. Apart from preparing his exemplars, Machaut would also have had to raise funds to get the manuscript project started. As the choice of workshop attests, quality was paramount. Up to this point, my imagined contextualisation tallies well with the accepted narrative of the creation of MS C. The two differ only in the book's potential to establish a patronage relationship rather than to celebrate one and in drawing attention to Machaut's need to create exemplars and negotiate with the workshop as first steps in the creation process. While the affiliation with an author-commissioner model rather than a patron-led process weakens the notion that Machaut was not involved in the source's completion, it will only become compelling as an alternative narrative if the specifics of the manuscript's structure can be accommodated at least as convincingly without external input.

T H E D I T S : B E T W E E N O R D E R A N D G A T H E R I N G S T R U C T U R E
The narrative dits that open MS C are each carefully copied onto separate sets of gatherings. 55 This arrangement was taken to indicate Machaut's wish to be able to change their order until the very last moment, with the final decision having been designed to appeal to Bonne of Luxembourg, or revised (perhaps by her son Charles) after her death. 56 While the physical analysis is beyond question, its implications are not. The most basic problem is that of intentionality: does the fact that the copying technique allowed for last-minute changes mean Machaut intended to exploit this flexibility, or did he have an order in mind regardless of the gathering structure, it being the result of workshop practicalities? If order was an ongoing artistic preoccupation, 57 he could have specified it while deciding on the number, position and contents of MS C's illuminations, perhaps in 55 See Earp, 'Scribal Practice', pp. 132-4, 371-2, reconstructed as Table 1 above. 56  the form of an overall list. 58 If he did not specify an order at the beginning of the process of creation, could it be that he invested so much in the visual presentation of this source but did not care about the order in which items were to be placed? Did he need extra time? How would more time enable him to improve the order? Was it because he was not sure which order would most appeal to Bonne, or because he hadn't yet settled on the book's final recipient? As the implicit or suggested patrons for all five dits to appear in this collection are Bonne's immediate family, 59 could he have gone wrong? Would it have been possible to order this group of works in such a way as to avoid the appearance that it was made for her or for a member of her immediate family? If Machaut was not sure who this manuscript was for, should we not try to come up with an explanation other than the link with Bonne as to why he commissioned it in the first place?
Before attempting to answer some of these questions, another relevant physical characteristic should be taken into account. When looking at the five dits MS C contains, the Behaingnewhich opens the manuscriptis the least sumptuously presented of the lot. Its image-per-folio ratio is the lowest and the illuminations were not executed by the best artists available. 60 More often, it is the beginning of a collection that is given the most visual attention, as this would be the first locus of engagement between it and the prospective 58 For an interpretation of the famous 'index' of A along these lines, see U. patron-reader. 61 If this work was given the least amount of visual attention, why was it placed first? Why would Machaut decide on an overall illumination programme that attracts attention away from it rather than towards it? The 'calling card' attitude presented above may offer an explanation. Placing the Behaingne first can be read as creating familiarity, this being the most famous and widely distributed of Machaut's dits. By playing on existing associations, the reader is drawn in, raising the likelihood he or she will go on to examine the other works on offer.
The enhanced presentation of the other works then draws the reader further in. The Remède, placed directly after the Behaingne, is the most effective in this regard. It looks more like a manuscript opener: it presents the highest level of artistic ability and execution in the whole collection, occupies the most space of all the dits present, and has the extra interest of the integrated music. 62 After being drawn in, one cannot fail to be impressed. 63 The concentration on Machaut's needs rather than those of a commissioning patron goes some way to explain the lack of specificity here: the association of single works with patrons concern the literary content, not the object of the book.  63 While the technique used is diametrically opposed, our knowledge of the early circulation of lais suggests that they may have been the most familiar of Machaut's musical works at the time of the creation of C. The visual attention given them here (see below) can thus be seen as an orientation tool, helping the reader to find more familiar works at the back of the book before delving into the rest of his output. 64 On this duality, see discussion of Froissart and his Meliador in n. 52 above. patron without their direct commissioning or prior acceptance of the project. 65 Table 1 demonstrates that not only were dits copied on separable gathering-structures, but that each such physical and literary unit was illuminated by a single artist. A common explanation for this is that it allowed the artists greater control in creating and executing the programme of illumination. 66 I contend that such assertions are only relevant in terms of the consistency of the visual presentation within each work. After all, the planning of the programme would have taken place between the commissioner and libraire long before the scribes or artists got involved. Scribes, in turn, fixed the location and size of the illuminations through the amount of space left empty for their execution. Finally, illuminations were executed following either sketches, verbal descriptions or reference to external lists of images or instructions. 67 Furthermore, due to the practicalities of drying, artists regularly dismantled gatherings and illuminated loose bifolia. 68 Put together, this process would have obscured the text's narrative flow and resulted in non-consecutive illumination, making it unlikely that an artist had much interaction with the text illustrated. Evidence from other Machaut sources raises further questions about this link. MS E often matches work and gathering structure, even though all but its first two illuminations were executed by a single artist. 69 The through-copied Vg, on the other hand, includes gatherings from a single work that were divided between a number of artists. 70 While the latter practice may not be ideal, it seems that as long as no marked stylistic changes got in the way of enjoying an illumination programme as a whole, commissioners, workshops and, by implication, readers accepted multi-artist collaboration within single narratives.
Another technical element that can, at times, be misconstrued is the meaning of gathering signatures. This is particularly pertinent for MS C as evidence remains of two different signature systems (see Table 1 above). Signatures are often associated with the very last element of production, that is, a manuscript's binding. It is of course true that a binder would require instructions for ordering gatherings (without engaging with their contents), as well as a mechanism through which to ensure each gathering was internally consistent. 71 Still, the need to keep track of the materials and to be able to take apart and rearrange sets of gatherings would have been required much earlier in the process. As already mentioned, artists illuminating one bifolio at a time would have had to do just that and would have required a system for reassembling and putting gatherings back in order once the paint had dried. The second set of signatures appearing on gatherings IX to XIII, for example, does not necessarily indicate that the Alérion and Vergier were at one point intended to begin the collection, only to be relegated to the end of the dits section. It more likely refers to the internal structure of the bunch of gatherings sent to the 'Master of the Coronation Book of Charles V' for illumination, perhaps even added by him/her before dismembering them.
Why, then, copy dits into independent structures? I contend that the best reason for doing so would be that this represents the state of the exemplars used. I do not suggest that the presentation in MS C is a faithful reproduction of the shape and size of the exemplars provided. Indeed, I would find it rather surprising for Machaut's personal copy to have included more than annotated suggestions for an illumination programme, thus saving much space. Furthermore, there is no easy way of reconstructing the size of his copies or how many lines and columns per page they contained. There is also no reason to believe that the various ditswritten over a protracted period of timewere kept in the same format. I do not believe that Machaut would have had any practical alternative to keeping his personal copy of each work in some kind of rationalised and unified physical structure. 72 As rhymed verse results in the length and number of lines for each work being set, 71 In this regard, it should be noted that the fact that C contains gatherings of various lengths and sizes could have been rather confusing for a binder, unless an external list detailing the size of each gathering was also available. it would only require a straightforward calculation to transit from one layout to the other and to plan gatherings accordingly. Indeed, a change in layout would make it more practical for one scribe to copy an entire dit, as it would be harder to coordinate dividing up the copying work if old and new layouts did not match. As the Remède was to receive special treatment by scribe, flourisher and artist, it made sense to be able to maintain its structural integrity (and by implication, that of the other dits). This would have facilitated passing it around between the various practitioners, setting it apart from the manuscript's other materials. This practice would have been unavoidable if we imagine Machaut sending his works to the workshop in dribs and drabs rather than as a complete unit, but this does not seem likely. 73 Without imagining a long build-up to the commissioning of MS C, a situation ruled out by viewing it as a reaction to John of Luxembourg's death, preparing new exemplars for the five existing dits would have been time misspent. Viewing the commission as a more spontaneous or reactive act leaves no obvious, practical alternatives for Machaut other than to have sent the narrative work to the workshop in their existing, self-contained structures. This position is further supported by the mirroring of MS C's readings in the surviving single-work circulation, often deemed to be 'early'. 74 Using such materials would have been a helpful starting point in negotiating the details of the commission, and would have enabled work to begin immediately while the more problematic exemplars for the shorter works were being prepared. The workshop, in turn, would have had no practical or technical reason not to follow this arrangement. 75 After all, it made life easier, both in terms of the division of labour and in maintaining uniformity of visual presentation. While the ability to tweak the final manuscript order later on may have been an added bonus, I suggest this would not have been the driving force behind opting for the gathering structure visible in this manuscript. 73 Unless one sees the extra attention given to the Remède as a sign that it was dealt with first, with the slight lowering of presentational expectation being the result of a decision that maintaining the same production values was unaffordable or too time-consuming.

T H E P R O B L E M S B E G I N : C O P Y I N G T H E L O A N G E
The structure of Machaut's personal collection of shorter lyrics is harder to gauge. While the imagined act of composition depicted in the miniature that opens this section of MS C shows the author surrounded by various unordered scrolls of parchment (fol. 121 r ), the non-musical lyrics (usually referred to as the Loange des Dames following the rubric introducing them in MS F-G) may already have been kept in a booklet. After all, writing down the shorter works required little space, making a one-per-sheet storage technique rather chaotic and involving scraps of various sizes. Unlike copying music, poetry requires no special design, ruling or preparation, and can be copied in succession without creating layout difficulties. The large number of poems -198 in MS C, over twice the number of musical workswould have called for some kind of organisation. As their composition appears more or less continuous, a gradual addition to a pre-prepared set of gatherings seems a reasonable assumption. 76 As a personal archive, these materials may still have presented the poetry in a suboptimal arrangement. If, for example, efficiency and economy of space were prime considerations, single-column ruling of as large a writing block as possible may have been employed. This would imply that the poetry was copied in a prose-like fashion rather than being lineated. This technique was familiar to Machaut: the residual texts of his ballades and rondeaux employ it consistently in all his later manuscripts. In any event, in preparing materials for the workshop, the creation of some kind of structure was likely, as a large bundle of individual works on single sheets would have been badly received and would have increased the potential for error or omission. As with the music discussed below, the process of achieving ordered exemplars could have taken place before the official commission, allowing Machaut to have a better overview of his materials and thus negotiate the planning and execution of the source more efficiently. Alternatively, the ability to send the dits straight away in their pre-existing structure may have given Machaut time to create new exemplars for the shorter works at this point. This second option is perhaps more appealing, as it more easily accommodates the structural anomalies that begin to appear from this point on.
As already hinted, current understanding views this part of the manuscript as the point where the problems began. As with the dits, 76 This is not to say that Machaut did not first write lyrics down in some other format, only that it would have made sense for him to preserve them by having a copy added to a larger physical collection immediately. a separate, irregular gathering-structure was also prepared for the collection of lyrical poetry (gatherings XVI to XVIII), though unusually, the copying of this section was divided between the two text scribes. When it came to copying, this section overshot the prepared space by nearly two folia. 77 The expanded gathering XVIII (a quinternion) suggests a clear expectation existed for the space needed to copy the lyrics, most probably reliant on the form in which they arrived at the workshop. The division of labour may at first glance suggest that this section was copied by the two scribes simultaneously, but the copying of Hé! gentils cuers, me convient il morir (Lo37) over the seam between gatherings XVII and XVIII negates this, as well as the possibility that gathering XVIII was a late insertion. 78 It is possible to consider Il m'est avis qu'il n'est dons de Nature (Lo188)which covers the seam between gatherings XVIII and XIXand the lyrics that follow it as later additions, but as the scribal hand changes with the structural break and not the lyrical unit, this seems less likely. In all, I contend that the most likely scenario for this 'overshoot' was a mismatch between the layouts used in the exemplar and the manuscript copy. If Machaut's personal, un-lineated copy of the Loange was used as an exemplar, the libraire's difficulty in assessing the number of lines required for this section becomes more understandable. The addition of an opening illumination and scores of rubrics labelling genres may also have interfered with the workshop's planning. In practice, this was compounded by the difficulties of Scribe A in particular in deciphering lyrical structure: he often split poetic lines in two, strengthening the notion that the lineation in the exemplar was suboptimal. 79 As we have seen, the rationale for professionalism was the saving of time. Professional scribes would 77 Earp, 'Scribal Practice', p. 137. Earp's suggestion that this was due to more poems becoming available than was planned for only works if the discrepancy occurred between the commissioning of the source and the arrival of the first set of exemplars. Otherwise, the copying of the entire music section would have to have been delayed considerably. It does not explain why the gathering structure was not adjusted to accommodate this, as grafting a bifolio to the pre-prepared structure would have been straightforward whether the addition was noticed before or towards the end of the copying of this section. 78 Numbering of Loange lyrics as well as of musical works follows Earp, Guide to Research, pp. xvii-xviii, 243-55. 79 The twenty-one poetic lines of Helas! pour quoy n'est bonne Amour si dure (Lo146), for example, are distributed over thirty-nine manuscript lines on fol. 141 r . Interestingly, such difficulties are not evenly spread, but are largely restricted to fols. 140-3, in the middle of gathering XVIII. It should be noted here that the many guide-letters appearing in Vg's copy of the musical ballades can perhaps be linked to the production stage rather than regarded as later additions. If so, they are more likely related to the placement of text in relation to music than to purely poetic lineation. The phenomenon (and later-reader interpretation) is discussed in Earp, Introductory Study, pp. 24-5.
not have expected to have to engage deeply with the author's intentions or have to interpret the materials. Indeed, there was no guarantee they would have had the poetic knowledge to do so even if they had the will and time at their disposal. In total, these issues added 151 lines to the space required by this section, accounting for more than half the 'overshoot'. 80 Had they been avoided, a single grafted folio would have sufficed to accommodate the estimation error. I find this hypothesis appealing as it combines a likely technique for lyrical archiving unrelated to the practicalities of manuscript-creation with a useful tool for explaining the structural anomalies in MS C as resulting from difficulties encountered by various workshop members. At the point of this structural seam, a scribal decision had to be made. It was not obvious to decide to start a new quaternion as it would have been abundantly clear that the remaining Loange materials would not suffice to fill it. A different option would have been to copy the remaining materials on single sheets or on a bifolio, grafting them into the structure once the remaining Loange lyrics were copied. While this would have enabled the workshop to maintain a correlation between genre and structure, it was perhaps decided that grafting further materials onto the already extended gathering XVIII would make it too bulky and result in a visible anomaly once the various gatherings were sown together. 81 Furthermore, considering the estimation error involving the lyrics, the workshop may have been justified in worrying about the copying of the innovative and technically more challenging music section. Committing to grafting here may have forced multiple subsequent graftings at a later stage. This would not only make the production process more complicated, but could potentially diminish the overall look of the final product and thus dent the workshop's professional reputation. The decision to copy through thus seems reasonable on a number of levels, even though it sacrifices the ease with which division of labour can be organised. As a result, only one change of scribe occurs beyond this point. Its location at a generic point of transition in the middle of gathering XIX attests to consecutive, rather than simultaneous, copying here. The decision itself and change of scribal hand at this point suggests it was at least given some thought and was probably taken by a superior. This cumulative lack of confidence may have 80 The entire 'overshoot' is of 282 lines of text, leaving 131 linesthat is, the equivalent of just under three and a half columns or about six balladesunaccounted for. This represents a 2.4% estimation error in a section containing some 5,450 lines. 81 Such grafting does occur later on in the manuscript (see below), but there the process is applied to a reduced gathering. While a substantial collection of nearly a hundred works would have called for some organising principle, I suggest this was more likely a separation into generically based bundles than gathered booklets. Such an internal division would have made the music easier to handle than an equivalent single stack of poems, the visual presentation of each genrereliant on the internal repetitions unique to each onesimplifying differentiation between groups. Nevertheless, using a single pre-ruled layout for each genre would have been inefficient, indeed wasteful. Within each generic group, different lengths and complexities of both poetic and musical setting could require vastly different amounts of space for the various compositional elements. For example, the decision to set a ballade in two, three or four voices would have meant significant implications for the arrangement of the page, and in order to accommodate the latter, a reproduced layout would not be efficiently used in the former. Variability of intra-generic poetic structure would result in different amounts of space being required for each piece, voice and textual residuum, with these differences potentially amplified by changes in the setting technique. Thus, identically structured texts may have required completely different layouts when presented musically, according to the number of voices chosen and the length and frequency of melismas within them. Without the ability to reproduce a ruling pattern or to predict the layout, the appeal of using pre-prepared musical gatherings is greatly reduced.
As with the Loange, the workshop was likely to have placed at least minimal demands on the exemplar materials. This was perhaps even more important for the music, as current knowledge suggests that no precedent for the collection of Ars nova secular polyphony was available at the time. 83 In what follows I suggest that Machaut's ordering technique was to find a consistent process for organising old works into quaternion units for each genre, postponing contending with works that could not be fitted into such structures (either due to unusual layout, or simply because a quaternion was full) to a later date.
The first two generic groups to appear after the Loangesets of virelais and ballades copied continuously without a correlation of genre and gathering structureare structurally interesting nonetheless. As attested by the space they occupy in the manuscript (fols. 157 v -165 r ), the sixteen ballades of CI would fit perfectly onto a quaternion, one on each of the sixteen sides this structure affords. 84 The twenty-three monophonic virelais that precede them required a little more space in the MS C copy (fols. 148 v -157 r ), but considering a likely less expansive copying technique for their many residual strophes, I contend they too represent exactly a quaternion's worth of exemplar materials. 85 Indeed, the inclusion of twenty-three rather than twenty-four virelais here supports the quaternion hypothesis. While including twenty-four songs would have resulted in a mathematically pleasing ration of three songs per folio, translating this into a gathering structure would have resulted in a virelai being stretched over every single page turn. Copying only twenty-three works allowed for a single work to be copied onto the first recto and the last verso of the gathering, and three songs to be copied on each of the seven full openings this structure offers. This eliminates the page-turning problem. 86 All the virelais that appear in MS C are monophonic, and while the number and length of lines varies, their generally syllabic style 83 The importance of this element is discussed extensively in Leach, 'Machaut's First Single-Author Compilation', pp. 260-70, with the technical novelties demanded here also being summarised in Earp, 'Interpreting the Deluxe', pp. 232-4. For an assessment of the problems that arise when exemplars are not well integratedeven on a small scale of seven piecessee the copying history of the music of the Remède discussed in Bullock, 'The Musical Readings', pp. 154-65. The considerable agency given to scribes in all these analysesand, indeed, othersrelies on the demand for clear exemplars not to have been made or met. Were clear exemplars available, the need for scribal agency would be greatly reduced. 84 On the notion of quaternion gatherings as a mental quantity of organisation, as well as a practical physical unit, see Stone, 'Made to Measure?' and Williams, 'An Author's Role'. On the arrangement of piece-per-opening in the various genres and Machaut manuscripts, see Bain, 'Why Size Matters'. 85 On the condensed exemplar presentation of lyrical texts, see analysis of the Loange above. 86 Vg and B copy these same virelais in just such a strict, three-per-opening manner, even though this at times results in a suboptimal use of space.
of setting allows for a relatively straightforward estimation of the copying space they would require. Thus, when Machaut came to prepare exemplars for this genre, he would not have had difficulties in planning a 'gathering's worth'. Perhaps he expected the workshop to continue the copying technique of the dit section, matching gathering structure to genre, but this did not happen in practice. I suggest that while it was the lyrical 'overshoot' that caused the subsequent mismatch between generic groupings and gathering structure, it was the technically ground-breaking elements of copying the musical materials that made the workshop weary of estimating the space required for each one. This is supported by the ad hoc ruling practice adopted for the music of this source mentioned above and explains the gap between the amount of space used in MS C and the proposed quaternion exemplar-gatherings from which it was copied.

B R I D G I N G T H E S E A M : P L A C I N G T H E L A I S
The lais follow directly on from the virelais and ballades, also beginning mid-gathering (fol. 165 r ). Nonetheless, this generic unit stands out both structurally and visually. Structurally, the collection of lais stretches over the location presented above as the point of transition between CI and CII (fol. 186 v , at the end of gathering XXIII, where a single ballade interrupts the generic collection). 87 Its beginning also marks the only place in the manuscript where a change of text-scribe occurs mid-gathering. 88 Visually, it is unique in the Machaut manuscripts in coupling an illumination with each song. Indeed, MS C contains no other illuminations in the entire music section. Thus, the miniature for L1 on fol. 165 r is the first such insertion since the beginning of the Loange on fol. 121 r , and the fifteen lai illuminations that follow are the last of the manuscript. The series has been interpreted as presenting an overarching, unifying narrative for this generic section, suggesting that they (rather than the songs themselves) had been conceived as a unit. 89 It seems that the lais were particularly important for the commissioner. 90 There is no evidence of an interruption within the illumination process of the lais. Thus, illumination either took place after an interruption between CI and CII, or the transition was smoother than hitherto proposed. This point is important, as it suggests that any changes of plan and structural adjustments took place within the original production process. The ordering of this generic collection relies on the hypothesis that the exemplar used here mirrored the stable order of lais found in Machaut's later sources, but included the instruction to copy first the musical lais and move those not set to music to the end of the generic section. 91 Coupled with the visual programme, this scenario demands a clear set of instructions by the commissioner, as well as the workshop's deep engagement with the exemplar materials. While unusual in the context of MS C as a whole and of wider professional workshop practice, I concur with it here. The shape of the lai exemplars was never fully discussed. Still, the notion of a set order represented in a later collection but reworked here only works if it was physically evident. While a list commenting on chaotic materials could have sufficed, a gathered set of exemplars accompanied by instructions as to how to deviate from them makes more sense to me. Counting the non-musical works as well, which appear early on in the generic order in later manuscripts, I suggest the CI lais would have fitted on three exemplar quaternions. 92 rarely be conferred on single-song composition. K. Maxwell, 'The Order of Lays in the "Odd" Machaut Ms BnF, Fr. 9221(E)', in E. J. Cayley and S. Powell (eds.), Manuscript and Printed Books in Europe, 1350-1550: Packaging, Presentation and Consumption (Liverpool, 2013), pp. 32-47 offers a narrative explanation for the unique order of lays found in a later source, clearly relying on non-authorial initiative. While individual illuminations do not appear, the proposed additional engagement with the lais may suggest that the sense of their generic pre-eminence within the songs was not confined to Machaut's person. 90 Assuming this is Machaut, other parameters indicate the importance of this genre within his artistic endeavours. See, for example, the discussion of the Remède's lai in Leach, 'Machaut's First Single-Author Compilation', pp. 261-4 and, more generally, Smilansky, The Lais. For an alternative reading of the emphasis accorded the lais in C, see n. 63 above. 91 Ludwig's numbering system follows the later order, while that of C is given in Table 2.
For the copying technique, see Earp, 'Scribal Practice', pp. 138-41, which also includes an explanation for the one exception to this rule. 92 It is particularly hard to assess the exemplar space required for the lais as the presentation technique of such works can vary greatly. Non-musical lais are prone to the problem of lineation discussed above for the loange and residual strophes of the virelais. The unusual copying procedure could go some way towards explaining the surprising position of the lais in this source (though I propose alternative explanations below). As with the visually rich Remède, one may well have expected the illuminated lais to open the music section. 93 As discussed in relation to the narrative work above, evidence of the early circulation of Machaut's lais supports an authorial, reader-focused explanation for not placing the most attractive and familiar works first. 94 Still, the same order of generic presentation could have transpired through the workshop deciding to tackle the technically unprecedented musical section of this source in order of increasing difficulty. First, the more usual materials were copied (dits, then poetry), followed by those music exemplars to be reproduced as supplied (first the easier to lay out, monophonic and text-rich virelais, followed by the polyphonic ballades), leaving lai exemplars, which required active interaction and reordering, to the end. The problems encountered in copying the Loange and the lack of confidence in pre-planning the music's layout resulted in the inability to reorder genres at a later stage.
The seam between CI and CII opens up three procedural possibilities, all of which create difficulties for the current narrative for the creation of the manuscript as a whole. One option, tentatively suggested by Earp, imagines the copying process of the CI materials to have been completed before the supposed break between the layers occurred. 95 As his understanding of this source necessitates the association of the motets with CI, Earp imagined an original gathering containing the text-only lais and first five and a half motets being discarded and its contents recopied after the decision to insert CII was made. 96 Consequently, this view sees the copying of this manuscript's text as having been completed by the time of Bonne's death in September 1349, presumably, however, at least partially unilluminated. This complicates the narrative concerning the transition between the two sections. As the process of illumination would have been time-consuming and not necessarily immediate, the notions of 93 1(a) and 2(a)) and the omission of M4 from this manuscript. Both elements are discussed further below.
an abrupt downing of tools in reaction to Bonne's death or of a hasty completion at some later point are both weakened. Other questions also arise: if the source was so near completion, why wait? Why not turn it immediately into either a commemorative artefact of its originally intended recipient, or the facilitator for a new patronage relationship? The subsequent CII addition becomes entirely opportunistic, inserted to update the 'completeness' of the manuscript. The substantial effort requiredthe diminished visual presentation of the CII songs and the loss of the clear generic principle of organisationare taken as a price worth paying for this, regardless of who was responsible for the decision. These problematic characteristics are also taken to evince haste in the execution of CII and the possibility of Machaut's detachment from the process, but while up-to-date completeness would perhaps be meaningful for Machaut, would it be for his readers? If he was not involved, how could anyone guarantee completeness? If completeness was so important, why was the Navarre not integrated into the dits section? 97 As illumination would take much longer than copying CII, why the rush? Why does CII not reflect more closely the visual qualities of CI? This version of events would also diminish the viability of the notion that a single narrative and visual programme stretched over the entire lai section, or at the very least, that the one currently on view was Machaut's original plan.
The second option is that the non-musical lais were not copied and then recopied, but were held back for copying later, making sense of the empty space on fol. 186 v . This, however, would suggest that the downing of tools here was primarily a reaction to a problem in the supply of materials for copying. It imagines an expectation of a need to copy works not yet present before the manuscript could be considered complete. According to this view, at least elements of CII were part of the original manuscript planindeed, integral to its designand Machaut continued his involvement in the project until its completion. To my eyes, the most likely materials for which waiting would have been worthwhile would be the final lais of the preenvisaged series and an entirely missing generic collection; but I will return to the rondeaux below. The third option is closely related to the second, but relies not on an expectation of new materials, but on a division between materials of different quality: between clean and problematic exemplars. Here, the unusual copying of the Lay de plour (L22)a work beginning on fol. 187 r , and thus marking the beginning of CIIis perhaps telling. Earp suggested that the problematic yet workable truncated end of its music can be explained by imagining this copy relying on an exemplar akin to Machaut's 'autograph'. 98 While he interprets this to claim L22 must have been a recent composition at the time of copying (i.e., closer to option 2), one can just as easily associate the difference with a transition from reliance on materials designed as exemplars for professional copying to personal copies destined for the author's archive (option 3). 99 Both options 2 and 3 transfer responsibility for the contents and presentation of CII to the original commissioner, and I find them practically simpler, generating fewer problematic implications. An awareness that the series of lais contains more works than were to be found in the CI exemplars would be another reason to place it after the virelais and ballades rather than at the head of the music section. The continuous copying technique adopted caused this to be the only location where a seamless transition between the layers would have been possible. Viewed in this light, workshop practicalities explain the surprises in this manuscript's structure. While neither option specifically contradicts coincidence with Bonne's death, neither requires it in order to make sense. Indeed, neither requires a break between the copying of the two layers.
If, as I propose, we opt for the second or third option and choose to eliminate the break between CI and CII, the questions raised in relation to the first option are defused and other implications arise. For example, understanding at least some CII materials as forming part of the original manuscript plan suggests that such compositions must have pre-dated the manuscript's commissioning, making them potentially as early as works copied into CI. Disassociating the two layers from notions of 'early' and 'late' has the potential to change our chronological understanding of both this manuscript's creation process and Machaut's creative activities. For example, the motets can still be considered pre-1349 compositions even if we imagine them being copied after, rather than before, the CII songs, and conversely, the composition of the rondeaux does not have to post-date that year due to their location within the manuscript. Moreover, the very necessity to use 1349 as a linchpin for dating is undermined. 98 Most of the music for its last strophe is missing, but a knowledgeable reader would recognise that it should reproduce the music of strophe I in transposition. For all this to be viable, however, alternative interpretations must be found for the unique characteristics of the CII section. Elizabeth Eva Leach's assertion that the quality of scribal work did not diminish here makes it possible to infer that no change in attitude or working practices occurred at this point. 100 This implies that the differences seen in the manuscript are not the fault of the workshop, but of the materials supplied to it. The most straightforward way of explaining them is through a marked deterioration in the quality of the exemplar, 101 that is, a transition from relying on clearly copied gathered exemplars to working with Machaut's personal, unstructured versions. As this change is physical rather than temporal, it does not matter whether such materials arrived at the workshop late (option 2 above), or whether they were available in this form all along (option 3). Indeed, a combination of the two is also possible. Before assessing why this could have happened and how the workshop may have reacted, let us look at the orderly motet section, which up to now was considered part of the older layer of this manuscript.

A C C O M M O D A T I N G C I I , R E C O P Y I N G M O T E T S
MS C's motet collection did not seem to challenge the workshop, and it was probably copied from a set of structurally unified exemplars. After all, models for collections of motets were already available, and the simplicity and regularity of its one-work-per-opening layout left less room for error. The evidence of programmatic ordering suggests that the organising hand was that of the commissioner rather than the scribe. It was probably always intended to end the manuscript. 102 As certain motets have been dated to the 1320s and 1330s, the currently accepted view of MS C's creation places the collection as part of the temporal orbit of CI. Earp suggested it was copied early on as a stand-alone group of gatherings for inclusion after gathering XXIII, with its first gathering recopied as a result of the integration of the CII additions. 103 Having removed the temporal need to associate the motets with CI, we have little to go on when considering whether they were copied before or after the CII songs. One clue is the omission of the motet De Bon Espoire/Puis que la douce/Speravi (labelled M4 due its stable position in Machaut's later collections) from this manuscript. Earp's explanation sees the omission as a small error brought about by the recopying mentioned above: the turning over of two leaves rather than one. 104 I suggest that this explanation works better within a single rather than a double copying process. After all, Scribe A would have been more likely to notice the omission if he or she had already copied the missing work once before. Other issues for consideration include the structural anomaly of gathering XXVI and the empty spaces left at the end of the motet collection (gathering XXVIII) and on fol. 186 v (end of gathering XXIII). I will consider them here in the order in which they appear in the manuscript. According to Earp's later-integration theory, fol. 187 r originally opened a gathering on which the first copy of the motet collection's beginning was to be found (see Figure 1(a)). 105 It is rather surprising that the independent, multi-gathering copying process envisaged should not result with a structure mirroring that of a dit, that is, starting with regular quaternion gatherings and adjusting the final gathering as is necessary. That this did not occur can be explained either by a compiler's decision to end the manuscript with a full gathering, or as the result of through-copying comparable to the songs of CI. The first option fails to explain why this easy to lay out section left the last opening of the final gathering empty. The second option was discussed in relation to the lais and is contradicted by the interruption of the group of lais on fol. 186 v . As Earp concedes, the placing of the non-musical lais L8, L9 and L11 before the motets in a hypothetical early version of gathering XXIII does not explain this anomaly. 106 It more likely implies an expectation for the integration of further additions into the manuscript before work on CI ceased. 103 Earp, 'Scribal Practices', pp. 140-1. 104 Ibid., pp. 141-2. 105 Ibid., p. 141. 106 Ibid. A different explanation that would support Earp's view of the original copying of the motets into a quaternion which began with two and a half empty folia would point at the structural 'overshoot' discussed in relation to the loange. This, though, would have two implications: (a) it would strengthen the notion that the workshop expected more The structure of gathering XXVI is rather mysterious: while its contents would fit perfectly on a quaternion, it uses only two full bifolia at its end preceded by four cut bifolia, or single sheets (see Figure 1(b) above). It seems clear that two elements have been grafted together here, as the transition from cut to full bifolia more or less correlates with the transition from CII songs to motets. The question is whether, as Earp suggests, the motet structure was re-planned to end with the beginning of M6 (as gatherings XXVII and XXVIII were already in existence), or whether the current structure was planned for the motet collection's first copy.
In both readings, it seems the compiler had difficulties estimating the amount of space required for the last few songs before the motets were to begin, either because their state made this difficult, or because it was not clear whether more would require copying. The recopying hypothesis necessitates the re-planning of the first five and a half motets. This requires five and a half openings and thus the preparation of a ternion. In order to explain the stub of the outer bifolio of this group, one would have to imagine the text scribe noticing the omission of M4in itself not surprising as he or she would have run out of exemplar materials before reaching the last verso of the gatheringbut deciding to cover up the error by cutting the last leaf of the gathering rather than owning up to the mistake and recopying the relatively small amount of material for the third time (see Figure 2(a)). While possible, I would argue that the multi-agent character of professional manuscript production makes such underhand tactics difficult to get away with. The other optionthat the motets were copied after the songs of CIIrequires only the preparation of two bifolia here, as the empty verso of the single leaf that is now fol. 206 could easily have been taken into account (see Figure 2(b)). The existing structure of two bifolia followed by two quaternions would thus provide the exact number of openings for twenty motets to be copied, using the last gathering to its full. This option removes the need to arrive at M6 at a specific point in the structure, and to have copied M4 once already. Thus, it is possible for the omission to have gone unnoticed. The final empty opening may just as likely have been seen as a planning rather than a copying error. While both readings are technically viable, the prior copying of the motets requires a few more production stages and involves a couple more open questions than the We have seen how the contents of CI seem to mirror a logical exemplar structure consisting of dits in self-contained units; two quaternions and a quinternion for the Loange; a quaternion of ballades and another of virelais; and three quaternions of lais. Tellingly, CII does not contain enough works in any single genre to fill another dedicated exemplar quaternion. This raises the possibility that the primary difference between CI and CII songs may only be the ease with which they could be integrated into structured, generically-based quaternion exemplars. 107 In this context, it is interesting to note that all the CI ballades are set to two voices, even though the Remède testifies that Machaut was already experimenting with four-part ballade composition by this point in time. 108 While potentially carrying chronological implications, this organising principle can also be read as a technique to control space and layout. According to this view, the prime parameter for inclusion in the quaternion exemplar of the CI ballades was Machaut's (or his secretary's) ability to create a consistent presentation. The easiest way to achieve this was to organize first all the works that conformed to the most popular setting size. Chronological and narrative ordering would thus have been a secondary consideration. 109 If this notion is accepted, strict compositional chronology and generic developmentalism is taken out of the equation. While some CII songs may have been brand new, others could have been just as old as the CI compositions. Two-part ballades may pre-date or post-date each other regardless of the section in which they appear, and three-part songs need not have been composed only after the creation date of the works in CI. This leads us to the rondeaux. Apart from the one incorporated in the Remède, all the musical rondeaux in MS C appear within CII. The current understanding of this section thus suggests they were set to music after 1349. This would be rather counterintuitive. After all, the rondeau is the only literary fixed-form genre for which we can securely identify a tradition of polyphonic musical settings clearly predating Machaut. 110 Furthermore, since Machaut had already written many rondeaux texts, I find it hard to believe that he would have needed the experience of composing dozens of settings in other, more complicated textual genres before tackling this relatively simple one. The appearance of rondeaux as motet tenors further weakens this stance. 111 Finally, it is also telling that the Remède rondeauplaced last in the descending order of difficulty presented in this ditis a mature, three-part work unlikely to be a first attempt. 112 Only a few years later Machaut recommended this genre as a good starting point for a budding (admittedly, literary) artist in the Voir dit. 113 Nine rondeaux appear in CII. Being short and non-strophic, their notation requires little space. Altogether, only six sides are needed, far short of the sixteen sides offered by a quaternion. Following on from the analyses of the separation between CI and CII ballades, virelais and lais, I contend that this is the very reason they do not appear in CI: Machaut had simply not yet come up with a mechanism to organise them into a unified gathering. Once again, an awareness of the absence of a whole genre (perhaps two, if the motets were yet to be copied) makes more sense of integrating the CII materials regardless of the extra work this involved and its detrimental effect on the uniformity and look of the manuscript as a whole. 114 As previously recognised, the songs in CII can be divided up into two sections. 115 Omitting B19, the first twenty songs appear in small generic groupings (see Table 2). As discussed above, the lais bridge the two sections and some of the non-musical works appearing here may in fact have belonged to the CI materials. Virelais and ballades follow in the same order in which they were presented in CI, with the new genre (new to the manuscript, not in terms of composition)the rondeauxat the end. The exchange in the order of R2 and B24 was probably due to a planning error. When ruling fol. 201 v , the scribe estimated correctly that the music of B24 would require four and a half lines but failed to allow room for double underlay under the first two lines, destined to contain the musical repetition of each strophe's A section. A short rondeaua genre not requiring text stackingwas copied instead, leaving a stave's worth of space empty. 116 The correct arrangement for the final ballade of the group was then rather inelegantly ruled underneath.
The remaining seven songs are more mixed in genre, even though one can detect a further two rounds of insertions following the same virelai-ballade-rondeau pattern. If we believe the CII materials arrived at the workshop late, it is tempting to conclude that the first group of twenty works (minus the non-musical lais) consists of old songs that Machaut was not able to integrate into quaternion-exemplars while preparing the CI materials and that the remaining seven (and B19) are newer works added to the collection in either one or two stages. If, instead, we imagine all of MS C's materials being already available at the commissioning stage and being dealt with in diminishing order of organisation and quality, there remains no reason to affiliate any temporal meaning with the order of the songs. Vestiges of the order of Machaut's bundle may still be present here. Still, the supposed need to locate and move new lais from within the bundle towards the CI lai section suggests it is as likely that the libraire 'improved' on a chaotic situation in order to reduce the number of  *denotes works without musical setting. 1 I take all songs which were not copied as presented in the CI exemplar (that is, also lais contained there but held up for copying later) to be CII songs. As my explanation for their copying (and that of the motets) removes the need for a break between the two layers, the exact point of transition has fewer implications in any event.
changes in layout and text-stacking requirements. Either way, this has substantial implications concerning the dating of at least the group of twenty ordered works, if not of CII as a whole. For example, even a conservative reading would imply that at least six rondeaux (R1, R2, R5, R6, R7 and R9)if not all ninecan be regarded as potentially much earlier works, requiring us to adjust our understanding of style-development in this period. Even dated compositions are not immune to this. R6, for example, is currently dated to 1350 or 1352 on the basis of combining the supposed dating of CII and its encryption of the name Jehan/Jeahanne. 117 Removing the temporal requirement means that the work could also refer to the 1332 wedding of John, duke of Normandy and Bonne of Luxembourg, or to John of Luxembourg's second marriage to Beatrice of Bourbon in 1334. 118 All the lais and virelais in MS C are set monophonically and current thought suggests the very late 1330s as the point in which Machaut began setting ballades. 119 Therefore, the revised dating may now enable viewing the rondeaux as his first locus for experiment with setting vernacular song to polyphony.
In practice, I suggest this section's creation can be explained as a result of the workshop's hesitation in dealing with inferior exemplars combined with a need to remain faithful to two initial instructions: that the motets should be placed at the end of the manuscript and that the unity of the emphasised lai section was to be maintained. Putting the motets aside, the workshop first exhausted the materials available in clear exemplars, leaving the lais to the end so as to place them at the transition between the two kinds of exemplars. The availability of another six lais made it clear that at least one quaternion's worth of materials was still available, hence the continuation of the standard gathering-structure in the MS C copy (gathering XXIV). This first CII gathering still left a lai and a half and twenty-one songs to be copied, allowing for an easy calculation that yet another quaternion will be filled: this number of virelais had already taken over a gathering's worth of space to copy in CI, and it was clear many of the remaining works would have required more space to copy than that monophonic genre. This resulted in the current gathering XXV, which contains most of the first round of generically grouped additions (see Table 2 above). Now, 117 See Earp, Guide to Research, p. 34. 118 Technical features such as the use of iambic minim-semibreve patterns and altered minims have been used to associated it with the 1350s, but such assertions are circular, as our knowledge of their temporal popularity relies exactly on this kind of song and manuscript dating. Similar stylistic and notational use can be found in many CI works, such as, for example, V1, V2 and V3. 119 See Plumley, The Art of Grafted Song, ch. 6. only nine and a half relatively short songs remained to be copied: clearly not enough for a complete new gathering. At this point, the remaining songs were copied on single sheets with the intention of grafting them on to the beginning of the motet collection. As grafting multiple sheets to a quaternion would have resulted in an unwieldy and hard to integrate gathering, the decision was made to have a shortened gathering at the beginning rather than at the end of the motet collection, which was easy to lay out. As long as the verso of the last single sheet remained free to begin the motet section, two and a half quaternion gatherings would have housed it perfectly. In practice, one motet was erroneously left out, resulting in an empty opening at the manuscript's end. The unusual characteristics of this section can thus be explained as the reasoned response of the workshop to changes in the quality of the materials from which it worked, with these changes themselves being a reasonable result of Machaut's difficulties in organising the remaining few pieces of each musical genre that did not suffice to make up a gathering.

C O N C L U S I O N S
In the preceding overview, I attempted not to be too prescriptive or gloss over the interpretative nature of some of my statements, at times offering more than one possible new solution to the questions raised by the materials. When combining such statements into a single narrative, however, choices must come into play. Two points in time suggest themselves as likely catalysts for the inception of the project, and for the same reason: both mid-1346 and late 1349 saw a material change in Machaut's career prospects with the death of John and Bonne of Luxembourg respectively. A strong case can be made for either, and I will happily swap position in the light of further research. Still, to my eyes, the later date is slightly harder to posit as a point at which Machaut would have lacked avenues of immediate patronage, having established links with John of Normandy and Bonne's court over the preceding three years. It is marginally less conducive in explaining the emphasis on the Remède, or the gap between the inception and completion of the Navarre. For the time being, therefore, I opt for the earlier date and suggest what seems to me the most likely series of events behind the creation of MS C.
To reiterate, the impetus for undertaking this project was not entirely artistic, but related to Machaut's career prospects following the Battle of Crécy in late August 1346, where his most constant and long-standing patron -John of Luxembourgmet his heroic death. 120 John's daughter, Bonne, wife to the heir of the French throne, would have seemed the most appealing dedicatee for a manuscript commissioned shortly after the battle, regardless of the degree to which she had offered direct patronage to the poet by this point. 121 The imagined bargain must have been that the gift of such an exquisite collection of poetry and music merited the acceptance of its author into the new patron's official (rather than practical) service. The requirement for speed in integrating himself into a new household, coupled with the social status of the intended recipient, led Machaut to choose a professional workshop already associated with book-production for the French royal family, as well as to use as much of his personal archive as possible as exemplar materials. These included all the dits and the Loange. In coming to deal with the other short works, Machaut concentrated on arranging the motets into a unified group and had as many works of other genres as would fit recopied onto complete quaternion gatherings. This amounted to three gatherings of lais and one each of ballades and virelais. The remaining works were kept in their existing, single-leaf copies. He was never to set a sufficient number of musical rondeaux to fill an entire quaternion.
Machaut's willingness to send what became the CII songs in a suboptimal state of layout and organisation seems surprising. Perhaps he neither wanted to begin a new gathering for each genre and send it mostly empty, nor to prepare smaller exemplars that would have to be redesigned once further works had been written. It could be that he prioritised some genres to the detriment of others. As discussed, the motets and lais demonstrate the clearest evidence of intentional ordering in this source, implying more attention was given to them. He may have expected the workshop to maintain an equivalence between gathering-structure and genre, rendering this presentation technique not unreasonable. 122 Still, the suggestion that the workshop's difficulties stemmed from a lack of clarity in the Loange exemplarscompounded, perhaps, by that of some virelais and lai textsmakes reliance on suboptimal materials here part of a pattern, rather than an exception. The existence of both kinds of exemplars explains the visual and organisational disparity between CI and CII without the need for haste, a temporal gap, Machaut's non-involvement, or external influence. The very separation into two layersnow more narrowly defined only according to the kind of exemplars used in each sectionwas the result of the workshop's copying process.
At this point (perhaps still in 1346) all the materials for the source were delivered to the workshop, following normal commissioning procedure. The relevant negotiations took place, settling on the illumination programme, level of execution, the pace and the price of the work. As part of this, some specific instruction concerning the more unusual illumination of the lais and placement of the motets (if not a complete general ordering of the manuscript) must also have been provided. 123 This would already have ensured that the workshop knew how many lais were to be copied and in what order. It would also be clear that the order of works in their exemplars was not to be followed and that the copying of lais should alternate between structured and unstructured exemplars. Structurally, practically and artistically, this section required considerable attention. It is not clear whether the rethinking of the order of laisrequired to accommodate the current understanding of their copyingwas the result of this negotiation process or had already occurred between the preparation of exemplars and the commissioning of the manuscript. 124 Now the workshop was allowed to take over and, in turn, took it upon itself to solve any difficulties encountered. While there was no explicit need for it to maintain the gathering arrangement of the old dit exemplars, doing so would have facilitated the division of work between scribes and artists and made it easier to maintain 122 The preventative measures taken in A, whereby blank space was left between throughcopied forms fixes sections (and, indeed, used for the copying of omitted songs and correcting errors) can be seen as a continuation of this expectation and a reaction to the problems encountered in C. See Earp, 'Scribal Practice', pp. 68-72. 123 The order of single short works was probably not specified. After all, within CI it was set by the physical structure of the exemplars, and the CII songs do not show evidence of such detailed ordering. 124 Perhaps a deeper engagement with motet order caused Machaut to want to re-examine more complex organisational techniques within the lais as well, or, as discussed further below, the composition of L22.
the visual integrity of each unit within the final product. 125 The dits were logistically the simplest part of the production (apart from the Remède, which clearly received special attention) and seemingly progressed without difficulty. The less-than-optimal presentation of the Loange exemplar caused problems in determining lyrical lineation and resulted in an error in estimating the space required for this section. Perhaps compounded by the need to use specialist scribes for the music and the music's novel layout requirements, Scribe B ended up privileging the clear, new music exemplar over the unordered single leaves, copying them consecutively with no regard for genreto-gathering equivalence. Among lesser problems created by this decision (such as the lack of differentiation between the end and beginning of musical genres), it became obvious that in order to fulfil Machaut's specific instructions regarding the illumination of the lais, these could not open the music section, but had to be moved to the seam between CI and CII (as newly defined here). Lais available on both new and old exemplars were thus copied on a continuous series of quaternion gatherings, leaving more than half a gathering empty for the unordered songs of other genres. Once this space was filled up with a more or less logical arrangement of small generic groups, a few more songs were still to be copied before the motets, which the commission specified should end the manuscript, and which offered no layout problems. As estimating space and controlling the layout of the songs proved a challenge, it was decided to maintain the appearance of structural design by constructing two and a half gatherings for the motets, placing the half-gathering at the beginning of the generic section and grafting to its beginning as many single sheets as would be required to copy the remaining songs. 126 Regardless of the degree to which the workshop's adjustments irked Machaut, a commissioning date of 1346 would probably mean that MS C would have been finished, presented to and accepted by Bonne at some point between 1347 and her death in 1349. The economic and military upheavals following Crécy and the ravages 125 The 148 folia of the dits and Loange was divided 60:88 between the two scribes. The ninetyone illuminations planned for this section were allocated thirty-three to thirty-four to twenty-four between the three artists. In both cases, it was the practitioner who contributed the least to this part of the manuscript who took on the lion share (or indeed all) of the work in its remaining seventy-six folia. Once again, this would have enabled work to continue seamlessly between the two sections, without having to imagine any breaks or delays. See Table 1 above. 126 As the number of remaining songs to be copied was already known, it was already clear they would not take more than half a quaternion's worth of space. This meant that even after grafting multiple single sheets the resulting gathering would not be bulkier than the quaternions surrounding it.
of the Black Death, however, may well have caused the settling of the bill to have taken a few more years and even to post-date her demise. Two options are available here: recompensing Machaut for his expenditure or picking up the bill directly from the workshop. As 1349 payments ordered by John II and received by Jean de Wirmes have already been associated with MS C, the latter option is more likely here. 127 The proposed new timetable only serves to remove some of the misgivings around making these associations.
To my eyes, this narrative has a number of advantages over the current model. Primarily, it engages more deeply with what we know of the practicalities of the professional Parisian book trade. Even if my suggestions are not accepted, this element cannot be ignored. It necessitates consideration of now lost exemplars and contractual arrangements. The new narrative also simplifies the commissioning process and the relationships between the actors involved in it. It explains the decision to include the CII materials, removing any sense of inferiority or temporal displacement attached to them. It does so while decreasing the number of production steps and coincidences needed for MS C to have acquired its current state and deals with more of its structural and visual anomalies. Once commissioning occurred, it did not require direct involvement in the physical process by either Machaut or an intended patron, or a reaction by the workshop to happenings beyond its metaphorical walls. It even sits better with external archival evidence, imprecise and non-specific as it is. I therefore find it more convincing.
If this hypothesis were to be afforded wider acceptance, the slight changes in attitudes towards MS C's internal relationships and dating might have wide-ranging interpretative implications. Beyond Machaut's activities in the late 1340s, the removal of a late terminus post quem for the composition of the CII songs may affect our entire understanding of the development of French musical genres and style in the first half of the fourteenth century. The rondeaux could now be seen as Machaut's first locus of experimentation with polyphonic vernacular song, offering a greater sense of continuity and development between it and earlier traditions. In the same vein, we would no longer be forced to associate two-part song composition with an early stage of Machaut's compositional development and three-and four-part works with a mid-century aesthetic and technical shift, as the previously presumed temporal separation between CI and CII suggested. 128 Instead, a more organic view of compositional experimentation with different types of settings can be adopted, with the importance of setting size as a primary organisational characteristic coming to the fore only in the act of creating exemplars, where consistency of layout took precedence over chronology and narrative planning.
Further narratives can also be developed, for example, in relation to the Jugement du roi de Navarre. 129 The Navarre is a literary counterpoint to the Behaingne, but while the latter opens MS C, it does not contain the former. Furthermore, the Navarre's text dates the work to 1349 but presents the composition of the Lay de Plourwhich appears as the first lai of CIIas part of the narrator's atonement, having lost a mock-legal argument. 130 This implies that the lyric emanated from and post-dated the dit, countering the codicological evidence. While some commentators read the omission of the Navarre from MS C as a political move, my potentially earlier date for the manuscript and its intended use in establishing Bonne's official patronage opens up another possibility. Due to its presentation in MS C, it has been argued that the Lay de Plour was composed not long before being copied into this manuscript. 131 If, as I suggest, we take John of Luxembourg's death in 1346 as the instigator of this manuscript's existence, the most obvious male death a lai of a similar date is likely to refer to is that of the Bohemian king himself, with Machaut ventriloquising his widow, Beatrice of Bourbon. 132 Bonne, regardless of whether or not she had been an active patron of her father's poet before 1346, would have had every occasion to hear about such a lament, as well as about Machaut's best-known dit to date, the Behaingne, which was explicitly dedicated to her father. This identification would have drawn Bonne's attention to the contradictory stances on female love described in the two works. If MS C was delivered to her complete, one can understand her commissioning of a new dit referring back to the two works with which she would have felt particular affinity as a reciprocal gesture, indicating Machaut's acceptance into her official service. Machaut's apparent modelling of the main female character in the Navarre on her person further links its creation with their professional relationship. 133 The new work's overturning of the judgement given by John in the Behaingne would then signify Machaut's transition from being her father's man in her court to forming part of her own retinue, rather than a slight on his earlier patron's rectitude. The incorporation of the lament on his death into the end of the new dit would have guaranteed that such overturning would not be seen as diminishing the late king's memory. 134 The pre-existence of the lai in relation to the dit thus changes from a problem to the very catalyst for the narrative work's creation, while the tension between the song's inclusion in MS C and the dit's omission from it is defused. Both the association of the Navarre with Bonne and its internal dating to 1349 would thus be able to stand, with Machaut making use of the work as an offering to a new potential patron following Bonne's premature death. 135 The reference within the Navarre to collecting and organising his works then changes from a general comment concerning Machaut's fictionalised parallel preoccupations with composing and commissioning to a knowing reference to this manuscript's role in the story's inception. 136 All in all, MS C emerges as an author-commission arising from a specific socio-economic circumstance where a need to establish a new patronage relationship required an artefact embodying cultural capital of wide appeal. While I would still link it with Bonne, its ordering and presentation served Machaut's needs as author more than those of a specific patron. 137 It did so not in a unifying, anthologising wayas did some of the later sourcesbut by attracting attention to his best-  135 The date in the text may still be more suggestive than literal, added to suggest the work post-dated Bonne's death and thus was always destined for Charles of Navarre. The difference between the two narratives is that the earlier completion date of C allows for this creative timekeeping to sound more plausible. See also n. 17  known and most unusual work: placing the Behaingne at its beginning, illuminating the lais and giving special visual attention to the Remède. Together with its ground-breaking contents came unique difficulties in layout and copying. As would have been expected of a highly professional workshop, production progressed independently over one continuous period, with problems solved as they arose. While the end result may not have mirrored Machaut's intentions exactly, it was still a successful commission. Bonne's premature and sudden death caused not only the short-term reworking of the Navarre into yet another socially expedient tool for establishing patronage, but probably resulted in the decision to change MS C from a stand-alone summa into the first of an extended 'complete works' manuscript tradition.
Obviously, I have no concrete evidence for any of this, but then again I hope to have demonstrated that the evidence for the currently accepted narrative is at best similarly unstable. My narrative takes into account the same set of hints and suppositions as the one advanced most importantly by Ursula Günther, Lawrence Earp, James Wimsatt and William Kibler, resolving some issues it leaves outstanding, while taking into consideration more of Machaut's commercial realities and known links. While my offer may well need adjustment and revision, is it not time to reopen the discussion?