Scribes or Copy Editors? Scribal Behaviour and the Production of Manuscript Copies of Jean Froissart’s Chronicles in Fifteenth-Century Paris¹

By Godfried Croenen and Sofie Loomans

The Parisian book trade in the early fifteenth century

By the time copies of Froissart’s Chronicles were produced in Paris at the start of the fifteenth century, the book trade in the French capital had been a well-established industry for a good while. Clear legislative, organisational and financial structures shaped and regulated the production and sale of books in Paris at this time. Scholars like R. and M. Rouse, who have studied the Paris book trade during this period through a number of important case studies, have highlighted the central role played by the libraires in the whole system.² It is clear that many libraires made their way upwards in the trade, starting out as parchmenters, scribes, decorators or miniature painters before achieving the status of libraire or grand libraire. While their essential role as libraire was the sale of both new and second-hand books, and the organisation of the production of new manuscripts through a system of subcontracting, they probably often participated in the actual production of these books whenever they had the time and possessed the necessary skills to do so.³

Libraires, however, were first and foremost businessmen and they had to please their clients in order to maximise their profits. To grow their business they needed to make sure that new manuscripts, especially expensive luxury products, fulfilled the wishes of their clients, both as regards the specific desiderata of individual commissioners as well as in relation to the general and changing expectations and fashions of the book market.

The copies of Froissart’s Chronicles that were produced in Paris in the first decades of the fifteenth century clearly bear this out. Whilst there are some signs of organised production of copies of the Chronicles in the Low Countries during, or shortly after, Froissart’s lifetime, the surviving copies are relatively few and quite different in appearance to the books produced in Paris in the early fifteenth century.⁴ By comparing the early Low Countries’ copies (for example the Rome manuscript of Book I, or the Leiden manuscript of Book II) with their Parisian counterparts, we can see that the Paris book people introduced a whole series of new elements, including a more complex page layout in two columns, a system of multi-layered segmentation of the text using initials of different shapes and sizes as well as paraphs and rubrics, and in many cases a programme of illustration.

The Parisian book people also intervened in ways that are less obvious because they can only be detected through the painstaking textual comparison of different manuscript copies. Like the scribes of many other medieval works, those who copied the Chronicles did often not simply reproduce the text of the exemplar they were working from, but also acted as a kind of copy editor, trying to improve the text in various ways.⁵ This could range from correcting apparent errors, both logical and linguistic, to updating the language and ensuring consistency across the different parts of the text, especially if the manuscript was to form part of a larger set, for example one including different books of the Chronicles. In the latter case the scribes may even add cross-references to their texts.⁶ Often the changes introduced by these scribes seem to have been made in a rather haphazard manner, in the course of copying the text, but in some cases there is evidence that changes were part of a more systematic undertaking.

The scribal interventions in the manuscript copies of Froissart’s Chronicles have received relatively little attention. The general tendency in earlier textual scholarship on this text, as in much other philologically-inspired work, has been to try to identify and isolate scribal interventions as much as possible in order to eliminate them from published text editions.⁷ In some cases, however, scholars working on Froissart’s Chronicles have been inclined to accept manuscript readings as reflections of authorial decisions, even in cases where the evidence suggests that they are much more likely to have been the result of scribal activity.⁸

The availability of transcriptions and reproductions of parallel manuscript copies of Froissart’s Chronicles on the Online Froissart now provides ample material for the study of the behaviour of commercial scribes. Such studies may give us invaluable clues as to what these scribes, or their masters the libraires, were trying to achieve. This essay will focus on this particular question in relation to some manuscripts of the so-called ‘A’ redaction of Book I, which have hitherto been largely neglected by earlier scholars.

Morgan, MS M.804, Toulouse MS 511 and related manuscripts

One of the manuscripts for which a complete transcription is now available on the Online Froissart is the beautifully illustrated copy of Books I and II of the Chronicles held at the Morgan Library and Museum in New York, where it is found under the shelfmark MS M.804.⁹ A. Varvaro, in his 1994 study of the iconography of the Book I manuscripts, devoted several pages to comment on this manuscript. A full edition of the Book II portion of this copy was published by P. Ainsworth in 2001, while a preliminary study on the manuscript’s history by the same scholar appeared in 1999. R. Sanderson’s PhD thesis, which contains a complete transcription of the Book I part of this manuscript, was completed in 2003.¹⁰

Before the end of the twentieth century, though, scholarship on this manuscript was very limited and the most sustained attention devoted to it came from art historians.¹¹ The manuscript was virtually unknown to the nineteenth-century editors of Froissart’s Chronicles, baron Kervyn de Lettenhove and Siméon Luce. And while Gaston Raynaud, who edited Book II of the Chronicles for the Société de l’Histoire de France in 1894, was aware of its existence, he was not in a position to study in any detail the manuscript itself or the relation of its text to that of other witnesses.¹²

The availability of the manuscript in a public collection has not only allowed the transcription and edition of the text, but also makes it possible, now, to carry out detailed codicological investigation of the physical book.¹³ Comparative study of the codicological and other characteristics of the Morgan manuscript has shown that it is closely connected to a number of other contemporary Froissart manuscripts, in particular to several manuscripts of Book I. The most obvious way in which the Morgan codex is linked to these other codices is through its palaeographical features. The Morgan manuscript is entirely copied by a single scribe, whom we will call Scribe T. Scribe T can also be found in two other Froissart manuscripts. One of these, manuscript Toulouse 511, contains the ‘A’ version of Book I and is likewise written entirely in the hand of Scribe T. The other manuscript in which Scribe T’s hand can be found is London, British Library, MS Arundel 67, vol. I. Here Scribe T has only copied a small part of the text, the first of two quires that were added to the original manuscript (fol. 358r–365v). The latter manuscript also contains the ‘A’ version of Book I, but it belongs to a different textual family from the Morgan and Toulouse manuscripts.¹⁴

Toulouse 511 has a twin, Glasgow Hunter 42, a manuscript that is very similar to the Toulouse codex. The Hunter manuscript is not written by Scribe T, but it is in a hand showing some similarities with the writing of Scribe T. In terms of layout, text and illustration, however, the Hunter copy is almost identical to Toulouse 511, even though it remains unfinished and therefore lacks any of the planned miniatures and initials, as well as many of the rubrics, for which spaces have been left by the scribe. The close similarity between Hunter 42 and Toulouse 511 indicates that they were probably produced more or less simultaneously by or for the same libraire.

While the text and rubrics in the twin copies are different from Morgan M.804, there are a number of other elements which allow us to confirm that there are nevertheless close connections between the production of the twin manuscripts and that of the Morgan copy. Apart from the palaeographical evidence — scribe T having copied both the New York and Toulouse manuscripts — there is the layout of these three copies. The Glasgow and Toulouse manuscripts have identical measurements for the text block on the page. The area of the page which contains the two columns normally measures in both cases 276 by 187 mm. This is only slightly smaller than the measurement for the first part of the Morgan manuscript (fol. 1r–258r, containing Book I), where the written space on the page measures 282 by 192 mm.¹⁵ Such similarity may indicate that the same persons were involved in the measuring and preparation of the parchment for these books.

Furthermore, the pages in all three manuscripts have been prepared in a way which is virtually identical and otherwise not very common. In all three copies time was saved at the ruling stage by not applying a full ruling on all the pages. Instead, the person who prepared the parchment applied four prickings to indicate the four corners of the written space on each page, which allowed a frame ruling. He or she added further prickings in the top and bottom margins to guide the vertical ruling of the columns. The ruling was then executed in leadpoint, not on both sides of the unfolded double leaves, as was common practice, but apparently only on individual folded pages. This can be deduced because the rulings on the recto and verso sides are different and can therefore not have been executed simultaneously (each side of an open double leaf represents one recto and one verso side of a finished manuscript). In these three manuscripts there is very often only vertical ruling for the columns appearing on the verso side, and only horizontal (or sometimes both horizontal and vertical) ruling on the recto side.

As the number of ruled lines and the distance between the ruled lines varies slightly from one page to the next and from one quire to the next, we can also assume that the horizontal ruling (except for the ruled lines at the top of the written space) was executed without the help of further prickings and presumably simply by using the already ruled lines above as a guide, tracing further lines simply in parallel to the lines that had already been ruled, starting with the top line for which prickings were available. The absence of any trace of prickings for the rest of the horizontal ruling in the side margins, and the fact that the lower lines hardly ever run precisely through the prickings provided for the bottom line of the written space, confirm this interpretation of the codicological evidence. Because the writing on the recto side was always executed first and because parchment is relatively translucent, this reduced method of ruling the pages nevertheless seemed sufficient to allow the scribes to copy the text on the verso sides more or less neatly.¹⁶

This particular and slightly peculiar way of ruling the pages in what must otherwise have been expensive commissions, points to a shared practice, possibly even the intervention of the same person — whom we could tentatively identify as the libraire — in the preparation of the parchment for all three copies. Further connections between the three manuscripts are to be found in the decoration and illustration. Apart from some larger initials accompanying the miniatures, both Morgan M.804 and Toulouse 511 are decorated with two-line champ initials, which seem to have been painted, if not by the same artist, then at least in a very similar style. In Hunter 42 the initials have not been executed, but space has been left for two-line initials, which could originally have been planned as champ initials as well. Finally, the programme of illustration in all three copies is very similar, with those in the Toulouse and Glasgow copies probably identical, if one allows for the unfinished state of Glasgow Hunter 42 and for the loss of some illustrated pages in Toulouse 511.¹⁷ The close similarities in the illustration of these three manuscripts (as well as in further manuscripts of Book I), betray close connections between all these copies.¹⁸

Scribe T and his base manuscripts

The codicological evidence cited above therefore seems to indicate a close affinity between the three main manuscripts discussed so far. Such a connection might typically be explained by a common origin, in particular by a situation in which all three manuscripts would have been produced by a single team of craftsmen, or several teams partly comprised of the same craftsmen, all working for the same libraire. Libraires often aimed for economies of scale by having the same scribes and artists produce so-called twin manuscripts, multiple copies of the same text manufactured using a single setup, including the same base text, a unified layout, and a single set of instructions to be used for all the copies, with scope for minor variation to satisfy particular clients’ wishes.¹⁹

If manuscripts were copied in this way from the same base text, they almost inevitably show close agreement on textual variants. That is why codices produced as twin manuscripts can often be identified because they belong to the same textual family that has been identified by a textual editor. Both Luce and Kervyn de Lettenhove have classified the Toulouse and Glasgow manuscripts of Book I in the same textual families, as one would expect, given the high degree of textual convergence in these two copies.²⁰ Textual comparison of the text of the Toulouse and Glasgow manuscripts shows that the agreement between them is indeed extremely high. One must have been faithfully copied from the other, or, more likely, both must have been closely copied from a common base text. Luce and Kervyn de Lettenhove have also assigned some further manuscripts to the same manuscript families. These other manuscripts are not contemporary with the Toulouse-Glasgow pair of codices and therefore cannot be considered as twin manuscripts of these two. Logically they must be descended from either one or the other, or from one or more further lost twin manuscripts that were produced from the same base text.²¹

On the face of it the New York manuscript does not seem to fit into this textual history, because the version of Book I that it contains is often quite different from the Toulouse/Glasgow text.²² Whereas the Toulouse and Glasgow manuscripts contain a redaction which regularly deviates quite significantly from the standard readings, providing a version of Book I that is to a large extent abridged, the New York manuscript often reproduces the standard readings much more faithfully than these two other manuscripts. Given the close codicological connections between the three manuscripts one could hypothesise that the New York copy may have been the text from which the Toulouse/Glasgow version was derived, but detailed textual analysis shows that this cannot have been the case.

The complex relationship between the standard version of Book I, the text transmitted in the New York manuscript, and the Toulouse/Glasgow version can be illustrated by passages such as the one below found towards the beginning of SHF § 304, where both the New York and Toulouse/Glasgow versions in turn preserve elements of the standard text (here represented by the Chicago manuscript) that are not in the other manuscript version:²³

‘Sitos comme elles furent passees, il conmenchierent a gueriier fortement, ly rois de France a conforter monsigneur Carle de Blois son nepveu, et ly rois d’Engleterre madamme la contesse de Montfort, ensy que promis et en convent li avoit, et estoient venu en Bretagne depar le roy d’Engleterre doy moult grant et moult vaillant chevalier et parti dou siege de Calais atout CC hommes d’armes et CCCC archiers, che estoient messires Thumas Dangourne et messires Jehans de Hartecelle.’ (Chicago, MS f.37, vol. 1, fol. 122r–v)

Sitost comme les treves furent passees, le roy de France conforta grandement le dit messire Charles son nepveu et aussi faisoit le roy d’Engleterre la contesse de Monfort. Et estoient venuz en Bretaigne depar le roy d’Engleterre II grans et moult vaillans chevaliers et partiz du siege de Calais atout IIC hommes d’armes et IIIIC archiers, che estoient messires Thumas Dangourne et messires Jehans de Hartecelle.’ (MS M.804, fol. 111r)

‘Sitost comme elles furent passees, ilz commencierent a guerroier forment. Et estoient venuz en Bretaigne depar le roy d’Engleterre monseigneur Thomas Dagorne et monseigneur Jehan de Hartevelle, et estoient partiz du siege de Calais atout IIC hommes d’armes et IIIIC archiers.’ (Toulouse, MS 511, fol. 112r)

Comparison of this short passage in the three different versions shows that both the New York and the Toulouse manuscripts abridge the standard text — and do so independently from one another. At the start of the passage the Toulouse manuscript is closer to the standard version than the New York version. Instead of having the pronoun ‘elles’ as subject of the verb in the subclause — as in both the standard text and the Toulouse manuscript —, the New York text deviates from the standard reading and gives the more specific variant ‘les treves’. New York then omits the main clause ‘il conmenchierent a gueriier fortement’, which is retained in the Toulouse manuscript. Instead it changes the infinitive ‘conforter’ into the conjugated verbs ‘conforta’ and ‘faisoit’ and turns them into the main verbs.

For the rest of the quoted passage, however, the New York text is much closer to the standard version than the Toulouse/Glasgow redaction, which heavily abridges the standard text. The New York manuscript abridges only slightly by omitting single words: ‘madamme la contesse de Montfort’ becomes ‘la contesse de Monfort’ in the New York witness and the phrase ‘moult grant et moult vaillant’ is reduced to ‘grans et moult vaillans’. The phrase ‘ensy que promis et en convent li avoit’ is entirely omitted from the New York text, but otherwise this witness retains all the constituent elements found in the standard version and gives them in the same order. The Toulouse manuscript, however, omits much more and also changes the order of the text, by bringing the names of the two English captains forward, before mentioning the number of troops they brought with them.

Passages such as this one — of which there are many — demonstrate that the Toulouse/Glasgow version of Book I cannot have been based directly on the New York copy, nor indeed can the Morgan manuscript have been copied from a manuscript which contained the Toulouse/Glasgow version. Whilst the Toulouse/Glasgow version very often abridges the standard text considerably, there is a large number of variants which demonstrate clearly that the Toulouse/Glasgow version must have been based on a manuscript which contained a full version of the standard text — or at least a version much more complete than what is found in the New York manuscript —, because it retains many elements of the standard version that are not present in the Morgan copy. To give just a few further short examples: in § 256 the Toulouse text has the standard reading ‘un bon port de mer’ (fol. 97r), whereas the New York text abridges to ‘un port de mer’ (fol. 94r); in § 301 Toulouse has the doublets ‘disoit et faisoit dire’ and ‘neccessaire et prouffitable’ (fol. 110v), as in the standard text, while New York reduces these, respectively, to ‘faisoit dire’ and ‘neccessaire’ (fol. 109v); and in § 302 the Toulouse manuscript has the standard reading ‘certains articles et traitiés’ (fol. 111v) whereas New York only has ‘certains articles’ (fol. 110v).

Readings such as these may not be deemed significant on their own, but taken together, and given that they appear in large numbers, they provide very strong evidence that the Toulouse/Glasgow version was not based on the New York manuscript. Even more convincing are the cases in which entire passages are missing from the New York version — as compared to the standard version — and where the Toulouse/Glasgow version has retained the missing text. A good example of this is found in § 313, where, following Edward III’s capture of Calais in 1347, Froissart reports on the king’s intention to repopulate the city with native Englishmen, a passage that is abridged in the Toulouse manuscript but entirely omitted in the New York manuscript:

‘... Et ensy a tous les aultres pour mieux repeupler la ville. Et estoit se intention, lui retourné en Engleterre, que il envoieroit la XXXVI riches bourgois, leurs femmes et leurs enfans, demorer de tous poins en le ville de Calais. Et par especial il i aroit XII bourgois, riches hommes et notables de Londres, et feroit tant que la dicte ville seroit toute repeuplee de purs Englés. Laquelle intension il accompli.’ (Chicago, MS f.37, vol. 1, fol. 127bis r)

‘... et ainsi a tous autres seigneurs.’ (MS M.804, fol. 115v)

‘... et aux autres seigneurs pour repeupler la ville. Et estoit son entention, lui retourné en Engleterre, qu’il envoieroit la XXXVI riches bourgois de Londres et feroit tant que la ville seroit repeuplee de purs Englois, laquelle entention il acompli.’ (Toulouse, MS 511, fol. 115v)

A more complex example is provided by § 62, which is highly abridged in both the New York and the Toulouse manuscripts, in the former to 81.9% of its original length and in the latter to a mere 29.8%. Both manuscript versions here shorten the standard version of the text, but they do so in divergent ways, leaving out different elements, with the effect that the resulting versions in both copies, leaving aside the differences in the scale of the reduction, are quite dissimilar.

It seems very plausible, then, that the Toulouse and Glasgow manuscripts were both copied from a common base manuscript that is now lost, which contained the original abridged Toulouse/Glasgow version of Book I. The conclusion to be drawn from the textual examples above is that such a base manuscript cannot have been based on the New York copy. It is also clear that the New York manuscript, in turn, cannot have been based on Toulouse’s and Glasgow’s base copy, nor indeed directly on either of these surviving manuscripts.

Textual comparison of the New York Book I text, made possible because its full transcription is now available on the Online Froissart, combined with codicological research on the entire manuscript transmission of Book I, has led to the identification of what in all likelihood is the manuscript from which most of the Book I part of the New York manuscript was directly copied: Paris, Bibliothèque nationale de France, ms. fr. 2655.²⁴ This manuscript has long been known as one of the earliest copies of the ‘A’ version, but despite this insight, it was largely ignored by both Luce and Kervyn de Lettenhove, neither of whom realised that it is probably a sub-archetype of the ‘A’ redaction, from which a substantial proportion of the surviving Book I manuscripts descends.²⁵ Partial comparison of the text of this witness with other surviving manuscripts of the ‘A’ version, in particular with the early fifteenth-century tradition, shows that it is one of the earliest and best witnesses of the two or three main families of the ‘A’ redaction.²⁶ All the significant variant readings that can be found in MS fr. 2655 are also present — unless they are affected by textual omissions — in the New York manuscript, in the Toulouse/Glasgow version and in the other manuscripts belonging to the same family.

A first example of such a reading is found near the beginning of Book I, in SHF § 3, where Froissart discusses the succession to the French throne and refers to Philip VI of Valois as ‘monseigneur Phelippe, filz jadiz de monseigneur Charles de Valoys, frere jadiz de ce beau roy Phelippe’ (our emphasis).²⁷ In MS fr. 2655 this authorial version of the phrase has been reduced and changed into ‘monseigneur Phelippe, nepveu jadis a cel biau roy Phelippe de France’ (our emphasis), a reading which also appears in the New York and Toulouse copies, as well as in other manuscripts of the same family.²⁸

Another variant which occurs in MS fr. 2655 and which is undoubtedly an error and therefore non-original, is found in § 260. Here Froissart states that there were 8,000 or 9,000 inhabitants in Saint-Lô at the time of Edward III’s 1346 military campaign in Normandy and northern France.²⁹ In many manuscripts these numbers are given in Roman numerals, as in the Chicago manuscript, which reads ‘VIII M ou IX M’ (fol. 102r). In MS fr. 2655, however, the Roman numerals are incorrectly written out, with the superscript numeral ‘M’ (1,000) given as ‘XX’ (20), which results in the total numbers being incorrectly listed as ‘VIII^XX ou IX^XX’ (‘160 or 180’). Both the New York and the Toulouse manuscripts reproduce this textual error, strongly suggesting that they both derive from a common source in which this shared error was already present.

Another significant variant typical of this textual family is found in § 261 where, as the result of an eye skip, several lines of text are missed out in MS fr. 2655, with the same omission occurring in the New York and Toulouse copies :

‘.. et grant fuison de bonnes gens d’armes que li rois de France y avoit envoiiés pour garder le ville et le passage contre les Englés. Ly rois d’Engleterre avoit bien entendu que la ville de Kem estoit durement grosse et riche et bien pourveue de bonne gens d’armes.’ (Chicago, MS f.37, vol. 1, fol. 102v)

‘... et grant foison de bonnes gens d’armes. (MS fr. 2655, fol. 131r)

‘... et grant foison de bonnes gens d’armes.’ (MS M.804, fol. 95v) ‘... et foison de gens d’armes.’ (Toulouse 511, fol. 98v).

A further example of an eye skip occurs in § 269. Here the repetition of the adverb ‘la’ seems to have caused the omission of several words in MS fr. 2655 and subsequently also in its presumed descendants:

‘... et eurent la li Englés de premiers I moult dur rencontre. Car tout chil qui estoient avoecquez monsigneur Godemars la envoiiet pour deffendre et garder le passage ...’ (Chicago, MS f.37, vol. 1, fol. 107r)

‘... et orent la envoiéz pour deffendre et garder le passage ...’ (MS fr. 2655, fol. 136v)

‘... Et eurent la envoiéz pour garder et deffendre le passage ...’ (MS M.804, fol. 95v)30

Common textual variants shared between the three manuscripts and which group them together against the other families also include a number of further instances where the text as it appears in MS fr. 2655 is clearly incorrect and the variant reading is therefore non-authorial. In § 266, which narrates Edward III’s 1346 chevauchée, the place name ‘Pont a Remi’ (Pont-Remy; Chicago, MS f.37, vol. 1, fol. 105v) is incorrectly rendered as the nonsensical ‘Pont Atemi’, ‘Pont Athemy’ and ‘Pont a Temy’ in MS fr. 2655 and the Morgan and Toulouse manuscripts respectively.³¹ The next section, § 267, relates how Philip VI of Valois was in Amiens, where he gave Godemar du Fay the order to defend the ford of La Blanchetaque. This passage ends with Godemar taking his leave of the king: ‘Sy se parti li dis messires Godemars dou roy’ (Chicago, MS f.37, vol. 1, fol. 106r). In the MS fr. 2655 and the Morgan manuscript we find a version of this passage in which the word ‘roy’ seems to have been garbled in the transmission process and changed to ‘kay’ (quay): ‘Si se parti le dit monseigneur Godemar du Fay du kay’. In the Toulouse manuscript the text is shortened and the phrase ‘dou roy/du kay’ omitted, possibly because the author of the abridged version present in this manuscript became aware of the inconsistency he found in his base text.

Amongst the significant variants found in MS fr. 2655 are also additions to the standard text, which appear both in this manuscript and in the others of the same family. In § 257, for example, the standard version has the reading ‘ne veurent miez leissier le chastiel ensy’ (Chicago, MS f.37, vol. 1, fol. 101r), to which in MS fr. 2655 is added a subclause, which seems to be repeated from a passage just above: ‘ne voudrent mie laissier le chastel, qui estoit moult fort, ainsi’ (MS fr. 2655, fol. 129v; our emphasis). In the New York manuscript the added phrase also appears, though with a slightly different word order: ‘ne vouldrent mie laissier le chastel, qui moult estoit fort ainsi’ (New York, MS M.804, fol. 94v; our emphasis), while the Toulouse manuscript does not add the whole subclause but still adds the adjective: ‘ne vouldrent mie laissier le chastel fort ainsi’ (Toulouse, MS 511, fol. 97v; our emphasis).

Many more such variants can be found. Some of these cases provide strong evidence for the hypothesis that the variants in the New York manuscript and in the Toulouse/Glasgow version of Book I go back directly to the actual surviving MS fr. 2655 rather than to a textual ancestor which this witness would have shared with the rest of the textual family. This is in particular the case where the scribe of MS fr. 2655 has left out words which he probably could not read in his base copy, leaving blank spaces instead, presumably so that he or someone else could return to those passages at a later stage to complete them. In § 2, for example, the scribe of MS fr. 2655 was unsure of the surname of Edward III’s younger brother John of Eltham and simply gave his first name ‘Jehan’ and then left a blank space. In the New York manuscript Scribe T reproduced the blank space, but in the Toulouse copy the anomaly is ironed out by omitting the space altogether: the fact that part of the name was missing would only be apparent to someone who had access to MS fr. 2655 and who would take the trouble of collating the different manuscript copies.

In other cases of this kind, such a simple solution was not possible and Scribe T was forced to be more interventionist. In § 10 the scribe who copied MS fr. 2655 seems to have been confused by the name ‘Hesbaignons’ (inhabitants of the Hesbaye) and left a blank space for this word. In both the New York and Toulouse manuscripts the blank is filled in with the incorrect reading ‘Behaignons’ (Bohemians), a scribal emendation undoubtedly based on the occurrence, just before this passage, of the erroneous reading of the placename ‘Behaigne’ (Bohemia), instead of ‘Hasbain’ (Hesbaye), an error probably caused by the visual similarity between the letters ‘b’ and ‘h’ in the cursiva libraria script.

A similar intervention on the part of Scribe T can be seen in § 31. Here Froissart talks about the shortage of fodder for the warhorses during Edward III’s Scottish campaign of 1327 with the result that they had to make do with what could be found and had to graze rather than be fed: ‘Et les chevaus mengier terre par le wason ou bruiere et fueilles d’arbres’ (Chicago, fol. 37, vol. 1, fol. 11r). The scribe of MS fr. 2655 seems to have been unsure about the word ‘wason’ (turf), possibly because of its Picard spelling with a letter ‘w’ at the start instead of the more standard ‘g’, and simply left a blank. In the New York manuscript part of this sentence appears completely rewritten: ‘Et les chevaulx mengier terre ensemble le petit d’erbes qu’ilz trouvoient, ou fueilles d’arbres’. In the Toulouse copy a different solution was adopted and the sentence so heavily abbreviated that the word ‘wason’ was no longer needed: ‘et les chevaux aussi fors fueilles d’arbres’.

In § 268 Scribe T adopted a similar approach in the passage in which the name of the ford of La Blanchetaque is explained by reference to the white gravel found on the riverbed: ‘Et au dit passage (...) a gravier de blanque marle, forte et dure’ (Chicago, fol. 37, vol. 1, fol. 106v). In MS fr. 2655 the word ‘marle’ (marble) is left out, with a blank space appearing instead. In both the New York and Toulouse manuscripts the text is patched up, but in different ways in each case, in the first instance by adding a reference to the white colour of the gravel, and in the second by supplying a synonym ‘perre’ (stone) for the missing word ‘marle’:

‘... a gravier de blanche [blank space] forte et dure ...’ (MS fr. 2655, fol. 136r)

‘... a gravier de blanche couleur, forte et dure ...’ (MS M.804, fol. 98v)

‘... a gravier de blanche perre, forte et dure ...’ (Toulouse, MS 511, fol. 101v)

It is theoretically possible that the omission of words and their replacement by spaces happened in the transmission stage before the copying of MS fr. 2655, and that the scribe of MS fr. 2655 simply reproduced the physical gaps left in his exemplar. Nevertheless it is much more likely that the omission happened in the copying process of MS fr. 2655 itself, because professional scribes, when faced with such clearly signalled omissions, mostly tried to patch up their text either by leaving out more words, so that the result was a syntactically correct phrase, or by trying to fix the omission through the addition of words that seemed appropriate for the context, or by rephrasing the passage. This is exactly what Scribe T seems to have done when copying from MS fr. 2655.

Chapter § 263 provides an even stronger textual argument to support the hypothesis that MS fr. 2655 was indeed the very manuscript from which Morgan M.804 was copied, and also served as base copy for the composition of the abridged Toulouse/Glasgow version of Book I. Here again we find an omission, in this case the name ‘Vrenon’ (Vernon), which is missing from a sentence appearing around the middle of this section: ‘Mes il allerent a Vrenon’ (Chicago, fol. 37, vol. 1, fol. 104r). Again, the scribe of MS fr. 2655 left a blank space, possibly to be filled in later. In this case, a cursive hand has effectively supplied a name in the margin: ‘Gysors’ (MS fr. 2655, fol. 133r). The suggested emendation is not correct, but nevertheless has made its way into the text of both the New York and Toulouse manuscripts.³²

A similar situation in which a change in MS fr. 2655 made it into the text of the other manuscripts occurs in § 264. Here Froisssart’s text contained the phrase ‘et retourna messires Godefrois sus le viespre deviers le grosse host dou roy’ (Chicago, fol. 37, vol. 1, fol. 104v). The scribe of MS fr. 2655 started copying out this phrase from his exemplar, but after having written the words ‘la grosse’ he seemed to change his mind, expunctuated the two words and instead wrote the near-synonymous combination ‘le grant’. The New York text has the latter reading, which suggest that it was directly derived from MS fr. 2655 and not only related to it via a common ancestor, as MS fr. 2655’s direct ancestor must have had the authorial reading, the phrasing which the scribe had initially written and then subsequently crossed out.³³

The textual evidence that MS fr. 2655 is not simply related to the other manuscripts in its family but should be considered a direct textual ancestor is therefore significant. Two further, independent elements support such a conclusion. The first is connected to the programme of illustration found in the Morgan copy and which is also present, with a few changes, in the Toulouse and Glasgow manuscripts. MS fr. 2655 itself is not illustrated, but a miniature was planned for the opening page, where the top half of the page is left blank. No further spaces were left for miniatures, but throughout the manuscript someone has entered instructions which indicate which chapters should be illustrated. These instructions are in the form of a small letter ‘h’ written in a cursive hand at the start of specific sections, obviously an abbreviation for ‘histoire’ (illustration). These instructions can evidently not have been meant for MS fr. 2655 itself, since the text of this manuscript was already copied and no spaces had been left for these miniatures. Instead, they must have been entered into the manuscript when a new copy was made from it, in which a more extensive programme of illustration was to be executed.³⁴

When comparing the placement of these instructions with the Morgan, Toulouse and Glasgow copies, it is striking that they agree very closely with the programme of illustration present in all the manuscripts, but especially with the illustration of the Morgan copy: whenever there are differences between the Morgan copy, on the one hand, and the Toulouse and Glasgow manuscripts, on the other, the places where the instructions appear invariably agree with the Morgan illustrative programme. It seems therefore difficult to escape the conclusion that the illustration in the Morgan manuscript was executed on the basis of these very instructions in MS fr. 2655, and that this manuscript therefore served as a direct exemplar for the former.

The final element in support of the argument that Scribe T used MS fr. 2655 for the production of the Morgan manuscript is provided by palaeographical evidence, which demonstrates that Scribe T, the scribe who copied Morgan M.804, had direct access to MS fr. 2655. On fol. 169v this scribe added the phrase ‘le prince de Gales’ in his normal cursiva libraria handwriting in the margin. This addition made its way into the text of the Morgan manuscript (fol. 121r). Furthermore, several of the instructions connected to the illustrations entered in MS fr. 2655 were probably also written by Scribe T, but the fact that they only consist of a single letter makes it difficult to ascertain this.³⁵ All this would tend to show that Scribe T not only used MS fr. 2655 as a base text for his copying work, but also that he was involved in the preparation of MS fr. 2655 as base text for the new manuscript copy.

Abridgement in the Morgan manuscript and the Toulouse/Glasgow version of Book I

The above analysis has shown that the surviving manuscript MS fr. 2655 was in all probability the base manuscript from which Morgan M.804 was directly copied, and also the manuscript that was used in drafting the abridged scribal version of Book I that is found in the Toulouse and Glasgow manuscripts. With this knowledge it becomes possible to study in some detail how the scribe of Morgan M.804 and the author of the Toulouse/Glasgow redaction went about composing these versions, and in particular how they used their exemplar.

What has already transpired from the discussion of these two scribal versions of Book I is that both the Morgan copy and the Toulouse/Glasgow redaction abridge the text significantly. If we consider the section SHF § 254–286, for example, we can use the transcriptions available on the Online Froissart to calculate with some degree of precision the extent to which the text of MS fr. 2655 has been shortened in these two versions.³⁶ For this section MS fr. 2655 contains 17,201 orthographic words in total, which is slightly below the word count for the standard version.³⁷ This indicates that the text contained in MS fr. 2655 is probably on the whole a faithful reproduction of the standard text. Word-by-word collation of the text of MS fr. 2655 against the standard version of these chapters as found, for example, in the Chicago manuscript clearly confirms this. For the same section, however, the Morgan copy only contains 15,287 orthographic words, or 88.87% of the word count of MS fr. 2655, which implies a reduction of the text found in the exemplar by more than 10%. Furthermore, the Toulouse manuscript has an even lower word count, with only 11,671 words, or 67.85% of the number of words present in MS fr. 2655, which amounts to a reduction by about one third of the original.

The pattern of abridgement appears even more clearly in Fig. 1, which represents visually the degree of textual reduction for the same section, but broken down chapter by chapter. The graph shows the percentages calculated on the basis of the word count found in the Chicago manuscript, which is shown on the graph as 1 (or 100%). The line representing MS fr. 2655 fluctuates very closely around that representing the Chicago manuscript throughout the whole section, except for § 261 where it drops down to 90% because of some textual omissions that occur in this chapter, probably the result of accidental copying errors. The Morgan and Toulouse manuscripts, however, show considerable variation from the standard text. While at times reproducing fairly accurately the text found in the exemplar, they mostly reduce the base text to varying degrees, by as much as two thirds, for the Morgan manuscript, and three quarters, for the version in the Toulouse manuscript, in chapter § 284.³⁸

Fig. 1: Word count per chapter shown as percentages based on the word count in the standard version of Book I (SHF § 254–286).

The textual examples discussed above have already illustrated several of the means by which the abridgement in both the Morgan copy and the Toulouse/Glasgow version has been achieved. On the simple word level, a study of the vocabulary in both versions shows that they almost constantly abridge by omitting adjectives or nouns deemed superfluous or not essential to the understanding of the text, or by replacing nouns or proper names by pronouns, when the context allows this. To give just a few further examples to those given above for the Morgan text: in Toulouse the noun phrase ‘li rois’ of the standard version is often replaced by the pronoun ‘il’ (for example in § 263; cfr. Chicago, fol. 104r and Toulouse, fol. 100r); and in § 305 the phrase ‘a monsigneur Jehan et a monsigneur Tangis’ (Chicago, fol. 123r) is rendered simply as ‘leur’ in the Toulouse manuscript (fol. 112v).

The author of the Toulouse/Glasgow version went even a step further and regularly substituted for certain words synonyms that are slightly shorter, a phenomenon that does not show up in the word counts, but which is something that would have had a marginal impact on the length of the copied text on the page. So the preposition ‘pardevers’ is often replaced in the Toulouse/Glasgow version by ‘devers’ or even ‘a’, and verbs composed of a root plus a prefix are often changed into the simple root form: for example ‘remander’ is replaced by ‘mander’.

In order to reduce the text even more, both the Morgan and Toulouse/Glasgow versions regularly simplify the syntax considerably. This affects infinitive constructions and set phrases, which can both be replaced by simpler constructions, often a single conjugated verb, while relative subclauses are replaced by past participles or shorter phrases. In § 317, for example, the passage ‘et ne fist nul semblant a ses compaignons de chose que il eust enpensé a faire’ (MS fr. 2655, fol. 163r) is reduced in the Toulouse/Glasgow version to ‘et tint son affaire secret’ (Toulouse 511, fol. 116v). This phrase has largely the same meaning but is considerably shorter than the original. The phrase ‘que on nomme/clamme’ followed by a name is often replaced in the Toulouse/Glasgow version by the single word ‘appellé’ followed by the name.

Lists of two synonyms or near synonyms are regularly reduced to a single term whilst longer lists are reduced to shorter lists or even to a single word. An interesting case of this is found in § 259:

‘... dont les plaintes et les nouvelles vindrent au roy de France ...’ (MS fr. 2655, fol. 130r)

‘... dont les plaintes vindrent au roy de France ...’ (New York, MS M.804, fol. 94v)

‘... Ces novelles vindrent au roy de France ...’ (Toulouse, MS 511, fol. 98r)

In the Toulouse/Glasgow version the passage has been rephrased, with the relative subclause introduced by ‘dont’ turned into the main clause. What is more interesting to note, however, is that from the doublet ‘les plaintes et les nouvelles’ found in the base text, the Morgan version has retained the first element and the Toulouse/Glasgow version the second. This illustrates again that both versions use the same techniques to reduce the length but that the actual composition of the two versions must have taken place independently.

A good illustration of the skill deployed by the author of the Toulouse/Glasgow version to reduce the text he was working from to its very essence is found in § 62. The first part of this chapter talks about the duke of Brabant’s knight Leonius of Kraainem and how he was sent to the French royal court to appease King Philip VI of Valois. This entire first section is dropped from the Toulouse/Glasgow version, a decision that was justifiable because further on in the text chapter § 72, where the same storyline is picked up again, largely repeats the essence of the information found in § 62 in the standard version. The omission of part of § 62 therefore does not affect the reader’s ability to comprehend that particular story line. The second part of § 62, which narrates how Edward III’s envoys return to London and report on their efforts to find allies for their king, is kept in the Toulouse/Glasgow version but is reduced by the omission of many of the original words from the text. In the Toulouse manuscript the reduced chapter § 62 only contains 208 words compared to the nearly 700 words in the standard version, of which no less than 466 words account for the second section alone, corresponding to the passage that is kept in the Toulouse/Glasgow version. This particular chapter is therefore reduced to 29% of its original length, or to 44% if only the second part of the whole chapter is considered. This reduction, with some loss of details, but with no loss of essential information, is achieved almost exclusively by omitting words, phrases and sentences and with minimal rephrasing.

The ‘copy editor’ of the Morgan and Toulouse/Glasgow versions

The degree of shortening that can be found both in the Toulouse/Glasgow version and in the Morgan version of Book I is considerable and cannot have been easy to achieve. Both versions betray the work of one or more highly skilled ‘copy editors’, who were able to retain the important and essential information they found in their base text and to reduce the length of the text without impinging on the legibility of the result or altering significantly the narrative style of the text. We may conjecture whether these ‘copy editors’ were the actual scribes of these manuscripts, or whether they were instead different agents in the production process.

The text of Morgan M.804, at least, gives the impression that for this version the person responsible for the abridgement was the same person as Scribe T, the scribe who copied the sole manuscript in which this particular manuscript version of Book I survives. Indications which support this are the fact that Scribe T seems almost certainly to have worked directly from MS fr. 2655 when copying the Morgan manuscript, rather than from a prepared copy based on that manuscript, as has been pointed out above.

A further indication to support this interpretation is the improvised nature of some of the rubrics present in Morgan M.804, in particular in those cases where the scribe simply decided to copy in red ink parts of the main text of the chapters, thereby creating the visual illusion of rubrics, without actually having real functional chapter headings (fol. 10r, 18v, 20r, 29v, 31r, etc.). This phenomenon becomes much easier to understand if we assume that the scribe had to work directly from MS fr. 2655, in which there are no rubrics except for the opening one, and that he did not otherwise have access to a list of rubrics from which he could copy. The scribe would then have had to compose the rubrics as he went along, which at times he may have found too much of an effort, because it involved reading the whole of the next chapter ahead of his copying it out. Simply writing some of the normal text in red ink would have saved him valuable time, while it would also have resulted in some further reduction of the overall length of the text, which seems to have been an important consideration. If these assumptions are correct, then Scribe T must have been responsible for the composition of the rubrics, which is a clear indication that he was not simply a scribe but that he had editorial and compositorial skills as well.

Finally, the fact that the abridged redaction of the text preserved in the Morgan copy has not been put to use by the person who drafted the Toulouse/Glasgow text, and that there is no indication that any further copies were produced of the Morgan version, makes it likely that the effort to reduce the text for the Morgan copy was not part of a larger-scale systematic operation and that it had a more occasional character. It was therefore in all likelihood limited to the production process of an individual codex, and seems to have been actually subsumed into the copying stage of the text.

The way in which the textual material was processed in the Toulouse/Glasgow redaction is in some crucial ways quite different from what seem to have happened with the Morgan version. The fact that at least two contemporary copies were made of this version, which were to all intents and purposes very similar, suggests that in this case a much more systematic approach was taken. From the very high degree of similarity between the Toulouse and Glasgow manuscripts and the fact that they contain a clean and largely error-free text, we can confidently deduce that the base copy which was used for their manufacture was well prepared and laid out, and included not only the text itself but also a number of different paratextual features, such as the rubrics, actual miniatures or sketches, or descriptions of miniatures to be executed, and clear indications for the placement of the various miniatures and initials.³⁹

The seriousness with which the author of the Toulouse/Glasgow abridged redaction went about his task confirms that it was clearly a planned process that must have taken some time to complete. Whilst Scribe T may have been able to use his technical skill as an experienced copy editor to abridge while he was writing down the text of Morgan M.804, this cannot have been the way in which the abridger of the Toulouse/Glasgow redaction went about his work. The scale of the reduction he achieved and the way he did this, often by changing the order of the material, imply that he was very well acquainted with the whole text and not just with the actual section he was working on. The example cited above of his decision to omit part of § 62 because the essential information of that passage would be found further on in § 72, shows that he must have had the time to verify such compositorial matters.⁴⁰ Also, his fairly frequent reorganisation of material, which often involved inserting textual fragments into earlier or later chapters, or changing the order of the narrative elements inside a section (as earlier or later chapters, or changing the order of the narrative elements inside a section (as happened for example in § 259, § 269, § 283, § 289 and § 290), must have required time as well as the opportunity to experiment with different solutions. We can therefore assume that his draft copy probably included fairly extensive textual deletions and insertions, possibly inserted sheets or slips tacked into the draft copy, as well as, possibly, instructions for the final order in which the text should be copied.⁴¹ It seems logical to assume that the Toulouse and Glasgow copies were not copied from such a potentially messy draft document, but instead from a clearly laid-out fair copy that was used as the working exemplar.

MS fr. 2655 gives at times an inkling of how the process of drafting the Toulouse/Glasgow version might be imagined. Earlier in this essay a marginal scribal correction which occurs in that manuscript and which contains the name ‘Gysors’ was discussed. On the same page in the manuscript there is another scribal addition in the margin, but this one is not a correction or addition to the text. Instead someone has here entered an instruction which anticipates precisely the kind of operation that is typical of the Toulouse/Glasgow redaction: ‘Cy faut mettre la chevauchee du roy de France’ (MS fr. 2655, fol. 133r). While this suggestion for the insertion of § 266 in the middle of § 263 was not taken up by the copy editor who drafted the Toulouse/Glasgow redaction, it is clear that he was not the only person who was exercised by the various different ways in which one might rearrange the text of Book I of Froissart’s Chronicles.

Another indication that the copy editor of the Toulouse/Glasgow version went through his text very carefully is to be found in the fact that he was able to detect and fix a number of textual problems in his base text. Chapter § 18 contains a clear example of such textual reparations. This passage narrates in gruesome detail the execution of Hugh Despenser the Younger in 1326 by castration, the opening of his body and the removal of heart and lungs, and quartering. In this section of the text all the manuscript witnesses of the so-called ‘A’ redaction of Book I have a textual defect, the loss of a passage of about 25 words, caused by an eye skip triggered by the repetition of the words ‘on les getta ou feu’ (first with reference to Despenser’s genitalia, then to his innards). As the omission occurs in all the ‘A’ manuscripts, it must have featured already in the archetype of the ‘A’ manuscripts. It also occurs in the two sub-archetypes, MS fr. 2655 and Brussels, MS II 2551, and from there it has found its way into the whole manuscript tradition of the ‘A’ redaction. The passage can be easily restored because § 18 is present in almost identical forms in all the other surviving redactions of Book I, as well as in the Chronicle of Jean le Bel, Froissart’s direct source for this section.⁴²

Even without access to a complete version of this chapter, it would have been obvious to a careful reader — as the author of the Toulouse/Glasgow redaction clearly was — that something is amiss when the text says that Despenser’s genitalia were removed and ‘thrown in the fire because he was false and treacherous in his heart’. The ‘copy editor’ responsible for the Toulouse/Glasgow redaction must have felt that this was an inconsequential statement and correctly guessed that there was probably something missing between the different parts of this sentence. Although he did not manage to restore the original phrasing of the lost sentence, he certainly did a good job in improving the text and restoring its internal logic:

‘Quant li membres et les coilles li furent copees, on les jetta ou feu et furent arses. Apriés on li fendi le ventre et li i osta on tout le cuer et le coraille et le jetta on ou feu pour ardoir par tant qu’il estoit fauls de coer et traittres’ (Chicago, MS f.37, vol. 1, fol. 6v, italics indicate the words omitted in the ‘A’ tradition)

‘Quant le vit et les coilles li furent coppees, on les getta ou feu [****] pour ardoir, par tant qu’il estoit faulz de cuer et traitre’ (MS fr. 2655, fol. 8r, asterisks indicate where the textual loss occurred)

‘Quant le vit et les couilles lui furent coupéz, on les getta ou feu [****] pour ardoir. Et aprés lui fu le cuer coupé hors du ventre et getté ou feu, pour tant qu’il estoit faulx et traitre de cuer’ (Toulouse, MS 511, fol. 13v, asterisks indicate where the textual loss occurred, italics indicate the added words)

The comparison of the original passage with the incomplete version found in most ‘A’ redaction manuscripts and the emended text that characterises the Toulouse/Glasgow redaction indicates that there is no direct textual connection between the added sentence in the latter and the original phrasing in the former. While the choice of words is very similar, and while the gist of the text is more or less the same, there is no real similarity at the level of word or phrase. Also, crucially, the addition in the Toulouse/Glasgow version is inserted not at the point where the text loss originally occurred, but just after the phrase ‘pour ardoir’, which originally followed the phrase that had been lost. This strongly suggests that the variant is not the result of a scribal change but the result of scribal restitution aiming to restore a lost reading.

While the scale of shortening that is apparent in the Toulouse/Glasgow version of Book I must often have required fairly substantive reorganisation of the material that goes much further than what is found in the Morgan text, the other aspects of the abridgement in the Toulouse/Glasgow redaction are very much reminiscent of the style of the abridgement of the text found in Morgan M.804. If the person responsible for the abridgement in the Morgan codex was indeed Scribe T, as suggested above, then there are good reasons to believe that he was also the ‘copy editor’ who prepared the abridged redaction preserved in the Toulouse and Glasgow manuscripts. This would explain why often the same or very similar choices were made in the omission of words or shorter phrases. Since the textual analysis above has demonstrated that there cannot have been direct textual influence from one abridged version to the other, the similarity in the resulting abridgement must have been the result of a close similarity in the editorial techniques used. Assuming that the same person was at work would make such similarity very likely.⁴³

The other important factor to consider in this context is that Scribe T is the unmistakable element that links the base text from which both abridged versions were made (MS fr. 2655), the Morgan version (through its unique manuscript, M.804) and the Toulouse/Glasgow version (through Toulouse 511). If he was the person who prepared the base copy MS fr. 2655, who then simultaneously copied and abridged its text into the Morgan codex, and also copied a different abridgement into the Toulouse manuscript, then it seems very likely indeed that he was also the person who prepared the base text for the Toulouse manuscript itself. The fact that another scribe made a second copy of that abridgement does not diminish his importance in the whole process.

Conclusion

The analysis of textual and codicological features of Paris, MS fr. 2655, New York, Morgan M.804, Toulouse, MS 511 and Glasgow, MS Hunter 42 has brought out the often complex relations between these four codices and the rest of the manuscript tradition of the ‘A’ redaction of Book I. In this essay we have argued that the volume Paris, MS fr. 2655, which was only partially completed because it lacked a frontispiece miniature, was at some point adapted and prepared for use as a workshop exemplar. We have also argued that Morgan M.804 was such a copy that was made from this base manuscript, and that, as part of his copying work, Scribe T, the scribe who wrote out the entire Morgan copy, reduced the size of the text through omission and localised rephrasing. We have then made the case that at a different time Scribe T used the same base manuscript to compose an entirely new abridged redaction of Book I of Froissart’s Chronicles, in which the text was reduced even more than in Morgan M.804. A fair copy of this abridgement was then made, which has not survived. It is likely that this fair copy was not in the first instance meant to be sold but was primarly intended to be used as a workshop exemplar. At least two copies were made from this new base text which are still extant today, Toulouse 511, copied by Scribe T, and Glasgow, Hunter 42, copied by another scribe.

The conclusions adduced about the base manuscripts that were used in the production of Morgan M.804, Toulouse 511 and Hunter 42 provide a sound basis for the further exploration of scribal behaviour in the production of the early fifteenth-century codices of Froissart’s Chronicles and in the production of manuscripts in Paris during this period in general. The analysis of the two abridged versions of Book I found in these three manuscripts also yields some clear insights into how at least some Parisian scribes approached the texts to be copied by them, and about what they considered to be their tasks.

The textual analysis has highlighted the relative complexity of the composition of the abridged versions, in particular of the Toulouse/Glasgow redaction. The conclusion must be that whoever was behind this project found it important enough to devote considerable resources to it. Given what we know about the Parisian book trade, this person was probably a libraire. In this case it is very tempting to identify Scribe T as the libraire behind this project. This scribe seems to have played such a prominent role in the whole production process of these codices that it is difficult to imagine that he was not also instrumental in the various operational as well as commercial decisions that were made.

Thinking about the economic realities of the production of these luxury manuscript suggests several obvious, and possibly complementary reasons that may have played a part in why these two independent abridgements (Morgan and Toulouse/Glasgow) came about. The first is linked to the client aspect of the libraire’s business: if potential clients were keen to own a copy of Froissart’s Chronicles but found the text a bit long-winded or simply too long to read given the various other calls on their time, then the prospect of owning a version which was considerably shorter than the standard text but was in all other respects virtually identical to the ‘real thing’, may have appealed to potential clients. The ability to procure such shorter copies in order to satisfy his clientèle may therefore have been important for a libraire.

The other elements that may have played a role are linked to production costs and the libraire’s profit margins and turnover. A libraire who was involved in the production of manuscripts containing very long texts such as Froissart’s Chronicles may have tried to increase his profit margin by reducing his production costs. In order to do so he may have decided to shorten the text, which would have allowed him to reduce the amount of parchment necessary for the production of each book, to reduce the wages he needed to pay to his scribes, and to reduce the cost for the decoration of each copy because the shorter text required fewer initials.

Alternatively, it may also be that because of the lower production costs, it was possible to offer the shorter copies at marginally lower prices, compared to copies that contained the full text, thereby reaching a clientele that could not afford the more extensive and expensive versions. This commercial strategy may have increased the libraire’s turnover. Either way, the composition of the Toulouse/Glasgow abridgement clearly required an initial investment, and it must be assumed that the libraire involved in this operation estimated that it would pay off in the longer run, either through increased profit margins and/or if sufficient copies of the cheaper text could be produced and sold.

Notes

¹ Godfried Croenen’s research for this essay has been generously supported by a research fellowship of the American Bibliographical Society and by a research grant of the Bibliographical Society (Cambridge). Sofie Loomans’ contribution to this essay was originally part of her MA dissertation, completed during the academic session 2012–2013 at the Katholieke Universiteit Leuven under the supervision of Michèle Goyens. This essay draws on chapter 2, which contains the transcription of Toulouse, MS 511, SHF § 254–318, and on chapter 3, in which the Toulouse and New York versions of these chapters are compared against each other. Both authors would like to thank Peter Ainsworth and Michèle Goyens for their support and feedback on earlier versions of this essay. We are also grateful for feedback from Mike Kestemont and Richard and Mary Rouse.
² See Richard H. Rouse and Mary A. Rouse, Illiterati et Uxorati: Manuscripts and Their Makers, Commercial Book Producers in Medieval Paris, 1200–1500, 2 vols (London: Harvey Miller, 2000). For a brief introduction see Godfried Croenen, ‘Pierre de Liffol and the Manuscripts of Froissart’s Chronicles’, in The Online Froissart, ed. by Peter Ainsworth and Godfried Croenen, v. 1.5 (Sheffield: HRIOnline, 2013).
³ See for example Richard H. Rouse, ‘Pierre le Portier and the Makers of the Antiphonals of Saint-Jacques’, in Patrons, Authors and Workshops: Books and Book Production in Paris around 1400, ed. by Godfried Croenen and Peter Ainsworth (Leuven: Peeters, 2006), pp. 47–67.
⁴ Godfried Croenen, ‘Philologues, historiens de l’art et codicologues: l’étude des manuscrits des Chroniques’, in Jean Froissart, Chroniques, Livre III. Le manuscrit Saint-Vincent de Besançon, Bibliothèque municipale, ms. 865, ed. by Peter F. Ainsworth, I (Geneva: Droz, 2007), pp. 32–39.
⁵ For a discussion of scribes as copy editors in relation to the manuscript tradition of the Arthurian romances, see Elspeth Kennedy, ‘The scribe as editor’, in Mélanges de langue et de littérature du moyen âge et de la renaissance offerts à Jean Frappier, 1 (Geneva: Droz, 1970), pp. 523–31. A. Grafton has recently published a wide-ranging study of the corrector in the early printed period: Anthony Grafton, The Culture of Correction in Renaissance Europe (London: British Library, 2011). He brings out the many other tasks, besides correcting print proofs, that were carried out by the early-modern correctors in the book production process, in particular as producers of texts and paratexts.
⁶ A striking example of an internal reference added to one of the manuscripts that will be discussed below, New York, Morgan Library, MS M.804, is found on fol. 175r of that manuscript. In § 514 of Book I Froissart records a conversation between the Captal de Buch and Falcon Herald, in which the Captal asks his interlocutor for information on enemy troops. While in the standard text the herald lists thirteen commanders by name, in the Morgan copy this passage is abridged by reference to the earlier chapter in which the same lords had already been mentioned: ‘Lors lui nomma les souverains si comme ilz sont escrips ou IIIe chapitre precident qui parle comment messire Bertrain de Claquin se mist sur les champs’.
⁷ This can, for example, be seen in S. Luce’s introduction to his edition of Book I, where part of the logic he deploys to categorise the manuscripts of the ‘A’ redaction, is precisely to identify the useful manuscripts (all those assigned to the ‘première classe’), and to distinguish them from those that are judged not to be directly useful for the edition (the ‘seconde classe’, containing manuscripts with an abridged text, the ‘troisième classe’, with fragmentary manuscripts, and the ‘quatrième classe’, containing manuscripts of epitomes), Siméon Luce, ‘Introduction au premier livre des Chroniques de J. Froissart’, in Chroniques de J. Froissart, ed. by Siméon Luce, I (Paris: Jules Renouard, 1869), pp. XXXIV–XXXV.
⁸ Kervyn de Lettenhove in his discussion of the manuscripts very often evinces a strong belief in the capacity of the different manuscripts to tell us something about authorial changes. This implies that he believed that many of the differences between the manuscript copies found their origin in the author himself and not in the later scribes. See baron Kervyn de Lettenhove, ‘Introduction. Troisième partie: Description des manuscrits’, in Œuvres de Froissart publiées avec les variantes de divers manuscrits, ed. by baron Kervyn de Lettenhove, I^II-III (Brussels: Devaux, 1873), pp. 185–461, passim. G. Diller seems to concur with this, as he is of the opinion that Froissart’s constant re-writing of Book I explains in part the considerable degree of textual variation found in the manuscripts of this text; Peter F. Ainsworth and George T. Diller, eds, Jean Froissart, Chroniques: Livre I (première partie, 1325–1350) et Livre II, Lettres gothiques ([Paris]: Livre de poche, 2001) (p. 62). M.-H. Tesnière, in her work on the scribe Raoul Tainguy, has been keen to accept as authorial changes to the text what are most likely Tainguy’s own interpolations in Froissart’s Chronicles, Marie-Hélène Tesnière, ‘Les manuscrits copiés par Raoul Tainguy: un aspect de la culture des grands officiers royaux au début du XV^e siècle’, Romania, 107 (1986), 282–368, in particular 304–306. On Tainguy’s interpolations, see Siméon Luce, ‘Un copiste interpolateur sous le règne de Charles VI: Raoul Tainguy’, in La France pendant la guerre de Cent ans. Épisodes historiques et vie privée aux XIV^e et XV^e siècles, ed. by Siméon Luce (Paris, 1890), pp. 245–59.
⁹ Book I was transcribed by Robert Sanderson as part of his 2003 Liverpool PhD thesis; book II was transcribed by Peter Ainsworth.
¹⁰ Alberto Varvaro, ‘Il libro I delle Chroniques di Jean Froissart. Per una filologia integrata dei testi e delle immagini’, Medioevo Romanzo, 19 (1994), 3–36, in particular 26–28; Ainsworth and Diller, eds, Jean Froissart, Chroniques: Livre I (première partie, 1325–1350) et Livre II; Peter Ainsworth, ‘A Parisian in New York: Pierpont Morgan Library MS M.804 Revisited’, Bulletin of the John Rylands University Library of Manchester, 81 (1999), 127–51; Robert David Sanderson, ‘Linking Past and Future: An Application of Dynamic HTML for Medieval Manuscript Editions’ (PhD thesis, University of Liverpool, 2003).
¹¹ See in particular Millard Meiss, French Painting in the Time of Jean de Berry: The Limbourgs and their Contemporaries, with the assistance of Sharon Off Dunlap Smith and Elizabeth Home Beatson, 2 vols (London / New York: Thames and Hudson / The Pierpont Morgan Library, 1974), pp. 368–372. However, see also the historical study of the manuscript by Léon Mirot, ‘Notes sur un manuscrit de Froissart et sur Pierre de Fontenay, seigneur de Rance, son premier possesseur’, Bibliothèque de l’École des Chartes, 83 (1922), 297–330.
¹² Gaston Raynaud, ‘Introduction au deuxième livre des Chroniques de J. Froissart’, in Chroniques de J. Froissart, ed. by Gaston Raynaud, IX (Paris: Renouard, 1894), pp. I–XIII (here p. VII, note 1).
¹³ A full codicological description will be published on the Online Froissart.
¹⁴ The Arundel manuscript is most closely related to Paris, Bibliothèque nationale de France, ms. fr. 2662. While there are elements that link the Arundel copy to the four main manuscripts discussed in this essay (Paris, Bibliothèque nationale de France, ms. fr. 2655, Morgan, M.804, Toulouse 511 and Glasgow Hunter 42), the connections are not straightforward and there remains considerable uncertainty about the precise relationship between the Arundel and Paris fr. 2662 manuscripts and the group of four manuscripts. The Arundel manuscript will therefore not be taken into consideration for the remainder of the discussion. See, however, below notes 18 and 43.
¹⁵ Although the Morgan copy of Books I and II is now bound as one volume, differences in layout and the fact that a table of rubrics is inserted before the text of Book II indicate that this manuscript was originally meant to be bound as two separate volumes.
¹⁶ Scribe T, or the person who was in charge of preparing the parchment for the books he copied, was certainly not unique in using reduced ruling methods. A variant of the system described here can be found in a contemporary manuscript of the Livre de Pontus et de Sidoine copied by Raoul Tainguy, Cambridge, University Library, Hh.3.16. Here a full ruling is found on one side of the leaves only, with a simple frame ruling on the other side.
¹⁷ On the illustration, see below. For further details on the illustration of Toulouse 511 and the planned illustration in Hunter 42, see the codicological descriptions of these manuscripts on the Online Froissart.
¹⁸ A. Varvaro has formulated the hypothesis that these similarities may be explained by a common source. He identifies this source, tentatively, as an authorial copy of Book I, which would have contained a programme of illustration that originated with Froissart himself, elements of which would have found their way into nearly all of the illustrated copies of Book I produced in Paris in the early fifteenth century and later (Varvaro, ‘Il libro I delle Chroniques’). Research on the illustration of Book I has provided no evidence to support this hypothesis, except possibly in relation to the first scene of the common frontispiece miniature. Rather, it is now clear that several separate and different programmes of illustration circulated in early fifteenth-century Paris. Shared elements in the illustration of the surviving copies are to be explained either by direct influence (direct copying or the use of the same set of instructions) or simply by the narratological importance of certain pasages in the text of the Chronicles. Godfried Croenen, ‘Froissart illustration cycles’, in The Encyclopedia of the Medieval Chronicle, ed. by Graeme Dunphy, 1 (Leiden / Boston: Brill, 2010), pp. 645–50. The programme of illustration present in the Morgan, Toulouse and Glasgow copies seems to have directly influenced the illustration found in two other contemporary manuscripts: London, British Library, MS Arundel 67, vol. I, and Paris, Bibliothèque nationale de France, ms. fr. 2662, as well as in the later Burgundian copies Paris, Bibliothèque de l’Arsenal, ms. 5187 and Paris, Bibliothèque nationale de France, ms. fr. 2643. Finally, the same programme of illustration also seems to have had some influence on two manuscripts which can be related to the Parisian libraire Pierre de Liffol, Besançon, Bibliothèque municipale, MS 864 and The Hague, Royal Library, MS 72 A 25.
¹⁹ See the case study of the manuscripts connected to the libraire Pierre de Liffol: Godfried Croenen, Mary Rouse and Richard Rouse, ‘Pierre de Liffol and the Manuscripts of Froissart’s Chronicles’, Viator, 33 (2002), 261–93.
²⁰ Luce does not give any precise textual arguments for how he arrived at his third family of the second class (MSS A30 to A33). Kervyn de Lettenhove sees the presence of SHF § 320 (which is absent in most manuscripts of the ‘A’ version of Book I), as well as a particular version of the opening rubric, as typical for this family. It is known that both editors often used the rubrics as their guide when studying the textual relations between witnesses. In this case the rubrics are virtually identical, as can be seen from a comparison of the two tables of rubrics. The Online Froissart contains a range of transcribed text from both manuscripts, to allow fuller comparisons. Currently available are the following transcriptions: for Glasgow SHF § 0–3, § 18, § 47, § 318–320, the final chapters and the first page of the table of rubrics; for Toulouse SHF § 0–98, § 181–182, § 186, § 254–320, part of the Grandes Chroniques de France interpolation, § 776–end and the complete table of rubrics. The transcriptions from the Glasgow manuscript were made by Godfried Croenen (SHF § 47, § 318–320, final chapters and table of rubrics) and Natasha Romanova (SHF § 0–3, § 18); those from the Toulouse manuscript by Sofie Loomans (SHF § 254–320), Godfried Croenen (SHF § 0–98, § 181–182, § 186, Grandes Chroniques de France interpolation; § 776–end) and Hartley Miller (table of rubrics, rubrics fol. 255r–268v).
²¹ A third possibility is of course that they were copied directly from the same base manuscript as Toulouse 511 and Hunter 42, which would make them their siblings — but not twins. Such a hypothesis would assume that the workshop exemplar was still available much later in the contexts in which these other manuscripts were copied, which seems less obvious but not impossible. Both editors put Tours MS 1041 in the same family as the Toulouse and Glasgow copies, Luce, ‘Introduction au premier livre’, p. XXXV; baron Kervyn de Lettenhove, ‘Introduction. Troisième partie: Description des manuscrits’, p. 291–94. Luce furthermore adds Paris, Bibliothèque nationale de France, ms. fr. 2651. Collations of the transcriptions currently available on the Online Froissart indicate that this is correct, but also that the whole of Luce’s second family of his second class (his MSS A23 to A29), as well as the incunable editions and two futher manuscripts unknown to Luce, Harvard MS fr. 152 and Rouen MS 1148, are in fact derived from the Toulouse/Glasgow version of Book I. Closer textual comparison of the later manuscripts to the Toulouse and Glasgow copies should shed further light on the precise textual relationships between all these witnesses.
²² The New York manuscript was unknown to Luce and Kervyn de Lettenhove and therefore does not feature in their discussion of the textual families of Book I. L. Mirot, who studied it in 1922, noted its textual similarities with manuscripts of Luce’s third and fifth families of the ‘première classe’ of the ‘A’ redaction manuscripts. Mirot, ‘Notes sur un manuscrit de Froissart’, p. 302.
²³ We are here using the Chicago manuscript as a very good witness of the ‘standard text’ of Book I. The term ‘standard text’ is here a shorthand for the authorial version of those parts of Book I that are common to the so-called ‘A’, ‘B’ and ‘C’ versions, i.e. § 12–320 and 371–697. The ‘A’ version is the same as Luce’s ‘première rédaction proprement dite’ and Kervyn de Lettenhove’s ‘seconde rédaction’. The ‘B’ redaction is Luce’s ‘première rédaction revisée’ and Kervyn de Lettenhove’s ‘troisième rédaction’, see Luce, ‘Introduction au premier livre’, p. XXIX–XLIII; and Kervyn de Lettenhove, ‘Introduction. Troisième partie: Description des manuscrits’, pp. 194 –234, 291–307 and 354–62. The ‘C’ redaction refers to the authorial version of Book I that only survives in the Chicago manuscript, see Godfried Croenen, ‘A ‘Refound’ Manuscript of Froissart Revisited: Newberry MS f.37’, French Studies Bulletin, 31 (2010), 56–60.
²⁴ The text in MS fr. 2655, as in the other early witnesses of the so-called ‘A’ redaction of Book I, breaks off before the end of § 697, so after this point the text that appears in Morgan M.804 cannot have been copied from MS fr. 2655. This is a textual problem that is not specific to Morgan M.804, though, but one that affects all of the manuscripts of the ‘A’ redaction whose text continues beyond § 697.
²⁵ Kervyn de Lettenhove praises this ‘excellent manuscrit qui a conservé l’orthographe de Froissart’, but incorrectly classifies it amongst the ‘complete’ manuscripts, emphatically — but incorrectly — denying any textual relationship between this copy and the Phillipps manuscript (now Brussels, Royal Library, MS II 2551), see baron Kervyn de Lettenhove, ‘Introduction. Troisième partie: Description des manuscrits’, p. 196, note 1 and p. 301–02. Luce rightly accepts the obvious relationship between these two manuscripts and classifies them in the same family, but fails to notice the relationship between MS fr. 2655 (his A7) and the families A18–19, A23–29 and A30–33, even though he lists some of the variants which are indicative of this genealogical relationship; see below notes 28 and 31. Groundbreaking work on MS fr. 2655, to which we are indebted, was undertaken by Susan Wales, ‘Froissart, Book I: Manuscripts and Texts’ (PhD Thesis, University of Sydney, 1988), in particular vol. I, chapter 5.
²⁶ Paris, MS fr. 2655 on the whole provides a very good version of the ‘A’ redaction of Book I, and the number of non-archetypal variants in this witness is therefore relatively small. The transcriptions from this manuscript currently available on the Online Froissart include SHF § 0, § 18, § 47, § 254-286, the complete Grandes Chroniques de France interpolation, § 421, and § 697. § 18 was transcribed by Natasha Romanova; all other transcriptions from this manuscript are by Godfried Croenen. A full reproduction of a black and white microfilm of this manuscript can be consulted on the Gallica website.
²⁷ Here cited from The Hague, MS 72 A 25, fol. 3r. This wording must be authorial, because it also appears in identical or nearly identical form in the ‘Amiens’, ‘B’, ‘C’ and ‘Rome’ versions of Book I, as well as in Jean le Bel’s Chronicle, see Jean le Bel, Chronique, ed. by Jules Viard and Eugène Déprez (Paris: Renouard, 1904–1905), I, p. 8.
²⁸ Luce noticed that the variant appeared in Paris, MS fr. 2655, as well as in the fifth family of his first class and the second and third families of his second class, but did not infer from this that there must be some sort of genealogical relationship between all these witnesses, Chroniques de J. Froissart, ed. by Siméon Luce, I (Paris: Jules Renouard, 1869), p. 218. It seems likely that this variant came about in two stages. First, an eye skip caused by the repeated word ‘jadiz’ could have led to the omission of the phrase ‘filz jadiz de monseigneur Charles de Valoys’, resulting in the obviously erroneous reading *‘monseigneur Phelippe, frere jadiz de ce beau roy Phelippe’. A scribe who came across this reading in his base text could have recognised the historical/genealogical error and then corrected it by substituting ‘nepveu’ for ‘filz’, thereby creating the variant that is now found in Paris, MS fr. 2655 and the rest of the textual family.
²⁹ The passage is closely based on the text of the Chronicle of Jean le Bel, who gives the total for the population simply as 8,000, which is the reading also found in the ‘Amiens’ redaction of Book I, see Jean le Bel, Chronique, II, p. 78; Jean Froissart, Chroniques. Livre I. Le manuscrit d’Amiens. Bibliothèque municipale n° 486, ed. by George T. Diller, 5 vols (Geneva: Droz, 1991–1999), II, p. 382. In the ‘Rome’ redaction the number given is slightly higher, with the reading ‘9,000 or 10,000’, Jean Froissart, Chroniques. Dernière rédaction du premier livre. Édition du manuscrit de Rome Reg. lat. 869, ed. by George T. Diller (Geneva: Droz, 1972), p. 686.
³⁰ This phrase does not appear in the Toulouse manuscript because of the abridged nature of the text found there.
³¹ These variants were noted by Luce, see Chroniques de J. Froissart, ed. by Luce (Paris, 1869), p. 390.
³² On the importance of this addition see Wales, ‘Froissart, Book I: Manuscripts and Texts’, in particular vol. I, p. 182–9. S. Wales is one of the few scholars who have highlighted the crucial role of MS fr. 2655 in the manuscript tradition of the ‘A’ redaction. Unfortunately, she was not aware of the Morgan manuscript, and so was unable to include it in her research, and she did not investigate the Toulouse manuscript either. But she pointed out the close links between the Glasgow manuscript and MS fr. 2655, both in terms of the textual features and in relation to the programme of illustration.
³³ In the Toulouse manuscript the passage is abridged and the variant does therefore not occur.
³⁴ The details of how this programme of illustration originated and how it was transmitted within this textual family invalidate to a large extent what A. Varvaro has written about the origin of the Parisian programmes of illustration for Book I, in particular his hypothesis that these programmes reflect an authorial programme of illustration, whose origins therefore go back much further than the production of MS fr. 2655; cfr. supra, note 18. See on this also Wales, ‘Froissart, Book I’, vol. I, p. 182 and vol. II, p. 37, appendix J. Wales was not aware of the very close similarities between the Glasgow and Toulouse manuscripts, nor of those between these two manuscripts and the miniatures found in the Morgan manuscript, see above note 32.
³⁵ It is theoretically possible, but not very likely, that MS fr. 2655 was also a product of Scribe T. This manuscript is copied in a textualis libraria script by two different scribes. The Morgan, Arundel and Toulouse manuscripts are all written in cursiva libraria. Scribe T could have been one of the two scribes who worked on MS fr. 2655, for which he would have used a different type of script that was not his normal one. It should be clear that such a hypothesis would be very difficult to prove. Moreover, the spelling habits used by the main scribe of MS fr. 2655 (who wrote the whole text except fol. 165r–171v), suggests that this was not Scribe T.
³⁶ For this section SHF § 254–286, the Online Froissart has full parallel transcriptions of three of the relevant manuscripts, MS fr. 2655, Morgan M.804 and Toulouse 511. Furthermore, there is word collation data available for § 254–261, 274 and 276, which permits an even more detailed comparison of these three manuscript witnesses.
³⁷ The word count for MS fr. 2655 represents 98.96% of the total of 17,381 words found in the Chicago manuscript. At least some of this variation is likely to be attributable to slight differences in the spelling in the manuscript copies, in particular differences in word separation. The rest is caused by some short omissions in MS fr. 2655.
³⁸ In § 277 the Toulouse/Glasgow version has a slightly higher word count than the standard version and MS fr. 2655, as the result of extensive rephrasing.
³⁹ It is of course possible that all these materials were not bound up in a single physical codex but divided over a set of materials which together would have formed the ‘base manuscript’. It is possible, for example, that a separate list of illustrations (actual designs or textual descriptions) was used, keyed in one way or another to the base text. Likewise, it may have been possible that the rubrics were kept separately on a roll or on a set of unbound leaves. For a general discussion of the former, see Anne D. Hedeman, Translating the Past: Laurent de Premierfait and Boccaccio’s De casibus (Los Angeles: J. Paul Getty Museum, 2008), pp. 36–41; for the latter, see Richard Rouse and Mary Rouse, ‘Some Assembly Required: Rubric Lists and Other Separable Elements in Fourteenth-Century Parisian Book Production’, in “Li premerains vers”: Essays in Honor of Keith Busby, ed. by Catherine M. Jones and Logan E. Whalen (Amsterdam — New York: Rodopi, 2011), pp. 405–16
⁴⁰ If Scribe T had produced the Morgan copy before starting the work on the Toulouse/Glasgow abridged redaction, then we may assume that he already possessed a certain familiarity with the text.
⁴¹ An interesting point of comparison is provided by the slightly earlier case of the autograph copy of Ranulf Higden’s Polychronicon, in which a large number of additions and corrections have been entered by the author at various compositorial stages, V. H. Galbraith, ‘An Autograph MS of Ranulph Higden’s “Polychronicon”’, Huntington Library Quarterly, 23 (1959), 1–18.
⁴² Transcriptions of chapter § 18 from all the surviving manuscripts of Book I, as well as full word-by-word collation data, are available on the Online Froissart. For Jean le Bel’s Chronicle, see Jean le Bel, Chronique, I, pp. 27–28.
⁴³ Further arguments for Scribe T’s role as copy editor/abridger are provided by the other manuscripts in which his hand has been recognised, MS Arundel 67, vol. I, containing Book I of Froissart’s Chronicles (see above, note 14), and the contemporary copy of the Grandes Chroniques de France, Kyn?vart Castle, MSS 85–86. The text copied by Scribe T in both manuscripts has been significantly abridged. The text of Book II of Froissart’s Chronicles found in the second part of the Morgan M.804 codex is also abridged, especially the storylines concerning Flanders, which are mere summaries of the original text. All this textual evidence shows that abridgement seem to have been part of Scribe T’s normal modus operandi.