Timothy A. Shonk

   In the long history of the study of manuscripts containing works by Geoffrey Chaucer, Harley 7333 has been treated as something of a poor cousin. In spite of the increasing attention to manuscript studies in the past two decades and the resultant flood of publications, Harley has received very little attention. Since the initial study of the Chaucer manuscripts by Manly and Rickert in 1940, Harley 7333 has not been investigated at length in any published study of which I am aware.

   And there is some reason for this. While Harley 7333 is unique in the corpus of Chaucer manuscripts in some ways--the largest in size of the Chaucer manuscripts (approximately 450 mm./17¾" tall and some 330 mm./13" wide), the largest in terms of its collected holdings, and the only manuscript in which The Canterbury Tales makes up less than half the book1--the text is so corrupt that it carries no authority for literary scholars. The Parson's Tale, moreover, breaks off incomplete, as do other Chaucer pieces, such as The Parliament of Fowls, which lacks its final stanza, and Anelida and Arcite, which lacks its final 120+ lines.

   The manuscript, furthermore, neither presents itself as a sumptuous manuscript, like the famous Ellesmere, nor holds a place of importance in the date of its composition, like the equally famous Hengwrt manuscript. Harley 7333 dates from about 1460. But in my view the Harley manuscript holds great interest for those who study medieval bookmaking. It represents one of a relative few larger collections of secular literature produced in the rural regions--Harley in Leicester, some 100 miles from London. The manuscript also holds promise for Chaucer scholars, offering valuable clues about the manner in which Chaucer manuscripts were being produced, "published," if you will, outside of London.

    Harley 7333 follows no straight line of descent from any of the earlier Chaucer texts, deriving from perhaps four exemplars from different branches of Chaucer manuscripts. Some of the exemplars, now lost, are of particular interest. In the Nun's Priest's Tale through the Parson's Tale Harley follows another significant

   Even more significant is that Harley's first segment of The Canterbury Tales is almost certainly related to the work of a relatively famous, though still rather enigmatic, book producer in London--John Shirley. Shirley rented several tenements in Bartholomew's Close, in which he copied, had copied, sold, and "loaned," through a kind of primitive circulating library, literary texts for some twenty-five years.5 Shirley, who had been a professional scribe all his life while also serving as esquire and secretary to the Earl of Warwick,6 seemed particularly interested in producing manuscripts of the poetry of Chaucer and of John Lydgate from the Abbey of Bury St. Edmunds, some of whose manuscripts were copied by scribes in the region and some sent "directly to Shirley for production and ‘publication’" in London.7

   Harley 7333 displays, in its early folios of The Canterbury Tales and in pieces occuring later than the Tales, markers of the Shirley influence: "characteristic rubrics and orthography which they [copies of Shirley manuscripts] preserve."8 The Shirley influence is also seen in the long wordy preface to The Canterbury Tales (see illustration 1) and in "the tautological headings for the first prologues and tales."9 A number of marginalia in the earliest tales, ending with the Miller's Tale, read "nota per Shirle." At least in part, then, this Chaucer segment of Harley 7333 is indebted to the exemplars being produced in the shops of one of London's earliest and most successful "publishers" of Chaucer.

   These points of textual influence have been generally accepted by the scholarly community. However, since Harley 7333 has not really been studied with any rigor as a codex since Manly and Rickert, scholars have adopted their conclusions without really putting them under scrutiny. My recent analysis of the manuscript leads me to believe that some of what Manly and Rickert deduced is in error, errors which have a significant bearing on both the methods of production of this manuscript and on its evidence for the "publication" of Chaucer in the provinces.

   Manly and Rickert assume, first of all, that this manuscript was produced by a group of Austin canons at the Abbey of St. Mary de Pratis. They cite as evidence apparent ecclesiastical censorship in such items as the omission of the end of the Pardoner's Tale to avoid the vulgar quarrel, the omission of the Shipman's Tale and its link, and the depiction in the Reeve's Tale of the wife as the daughter of a swanherd rather than of a parson.10 Curiously, though, as Charles A. Owen notes, passages in the General Prologue and in the Canon Yeoman's Tale one would presume to be equally offensive to such ecclesiastics have not been edited

   On f. 150 remains a note that "Doctor Peni writ this boke." Manly and Rickert identify a John Peny, an LL.D. of Oxford, who was a canon at Leicester

   Manly and Rickert also identify from "six to nine or more hands"17 at work on the manuscript. Here I am certain that they are in error. As a result of their rather general descriptions of the hands, they see sometimes as many as three hands involved in writing pieces of the same quire (the final quire, for example). Certainly this does not seem evidence of the type of efficient, organized production of manuscripts one might expect to come from an abbey of the Austin Canons. In contrast, I see three hands at work. My reading of the division of labor by three hands shows a system more conducive to accurate and efficient copying and falls more neatly into the seven booklets which Manly and Rickert have identified as comprising the manuscript:

   Some significant points must be noted. Scribe I, the major scribe, shares booklets with the other two on occasion. Scribes II and III appear together only once, this single instance representing scribe III's lone appearance, suggesting a one-time assignment in the work of one not involved intimately with the production.

   A problematic fit of the scribes’ work occurs on f. 57r. Scribe III's only stint is interrupted abruptly on f. 56v, before concluding the Miller's Tale. Scribe II reenters at this point to add a single leaf concluding the Miller's Tale and presenting the prologue to the Reeve's Tale (see illustrations 3-5), but he ends this brief stint on f. 57v at line 20--an odd occurrence to be sure, but one quite revealing about how the manuscript was compiled. As noted earlier, it appears that a Shirley exemplar was used at least through the Miller's Tale, a different exemplar employed for the next group. It seems, then, that scribe III copied a small portion of the Tales from an exemplar defective in some manner (a lost or damaged final leaf perhaps). Production halted for a time until an exemplar was found to fill the gap, an exemplar which seemingly shows no Shirley influence. Scribe II thus wrote the linking lines and scribe III returned to write one more quire before disappearing from the codex. This incident becomes all the more interesting when, two leaves later, on f. 58v, scribe III copies a catchword linking to f. 59r. But no new quire begins here. It seems that the new exemplar employed by scribe III broke quires at this stage, and the scribe dutifully copied the words on the page before him. The change of exemplars is further indicated by the erasure of the final line of 56v. It appears that scribe III ended the first part of his stint here, unaware of what was to follow. Rather than copying the first few words of what would be the first line of 57r, scribe III found himself at a loss for a catchword. Thus, to avoid repetition, he marked his stopping point by erasing the final line he had written (note the uneven columns and the swirls of remaining letters which correspond to those letters in the new catchword) and wrote the entire line as a catchword, the line that would follow on 57r when the new material was found.

   A similar occurrence takes place on f. 134r, where scribe II once again added a singleton at the front of a quire. The preceding quire had ended with 2 blank folios and a blank column b on f. 133v, where Chaucer's Mars and Venus ends incomplete. Scribe II added the singleton to begin Anelida and Arcite, which scribe I resumed on f. 134v. Thus scribe II may have served as an organizer or final compiler of sorts, obtaining the exemplars and filling in when the original exemplars were insufficient for whatever reason or replacing a lost or damaged leaf.

   In contrast to Manly-Rickert's view that the manuscript was necessarily completed in a scriptorium by a group of ecclesiastics, I find it equally possible that the work was completed on contract, perhaps from professional craftsmen working in the area, producing booklets for the Abbey or copying independently booklets later purchased by the Abbey for compilation into the codex.

   A mass of evidence makes it clear that lay professional book producers were at work in London as early as 1330, as I have demonstrated in an earlier publication on the famous Auchinleck manuscript.18 But what evidence is there for such professional book production in the rural regions? And, particularly relevant to Harley 7333, were book producers for religious houses necessarily resident ecclesiastics or were they lay craftsmen?

   While oversimplifying the trend certainly, J. P. Tatlock summarizes the movement of book production out of the monasteries and into the hands of lay professionals: "In Chaucer's day the time was long past when almost all book-making was in the hands of ‘the old monks.’ ... it is impossible to imagine that secular reading matter multiplied much except through secular and commercial routes."19 In monastic libraries Doyle finds "ample evidence that religious houses and their members employed named and unnamed scribes in the fourteenth and fifteenth centuries, to write documents and books, both over a period and occasionally."20 Doyle concludes that "in the last century and a half of the middle ages in England, whatever may be true earlier or elsewhere, monastic book production can-

   Recently, scholars have found evidence that professionals were employed in the rural areas, sometimes by religious houses, to produce manuscripts. A most notable example of professional book production in the provinces is the activity in the area of Bury St. Edmunds, where, for some decade and a half, a group of professionals produced a series of manuscripts, primarily poems of John Lydgate. The most prolific scribe of this area is described by A. S. G. Edwards as "a co-ordinating scribe capable of drawing upon the services of a number of proficient artists and decorators to adorn his work," a group which comprised "a commercial infrastructure capable of sustaining the demand for manuscript copies."22 Roy Vance Ramsey has discovered a professional scribe at work near Norwich, one Geoffrey Spirleng, who owned two scribal shops near city hall, copied MS Glasgow Hunterian U. 1.1 in the late fifteenth century, and trained other scribes as apprentices.23 A more pertinent example may be one Henry Mere, depicted by Malcolm Parkes as a "professional scribe and layman"24 who did much if not all of his business with the Christ Church priory in Canterbury, producing work for the convent. Parkes’ scribe probably copied other texts elsewhere than at Canterbury and may have been a peripatetic scribe.25 A final example, offered by Phillipa Hardman, seems very close to the kind of work I think was being done with Harley 7333. Hardman describes the collaboration of three scribes (two identifying themselves as Recardum Heege and John Haughton) in producing booklets to be compiled into a "Library in Parvo" in the North Midlands in Nottinghamshire.26 While Hardman refrains from deeming these scribes professional, Thorlac Turville-Petre suggests that one of these booklets may have been produced for the main scribe's own pleasure or "on commission for a local family."27

   Certainly the Austin Canons were active in bookmaking in the fourteenth century, with some individuals active in the fifteenth, such as John Capgrave.28 And, as Manly and Rickert affirm, copying in some sort of scriptorium was done at the Leicester abbey by William Charyte, who copied some texts, besides "buying many other books."29 But Charyte's hand appears not at all in the Harley codex, and a catalogue of holdings done in 1477 makes no specific references to the major works of the codex, though a very few pieces may be noted by more generic references, which may suggest that some of Harley's pieces were in the abbey before Harley was assembled.30

   Given the relationship between lay producers and religious houses, there are, I think, some better explanations for the manuscript's idiosyncrasies than that of monastic production, whether in a scriptorium or not. For example, Manly and Rickert note that the decoration of the booklets is inconsistent, evidence for what they deem seven different "styles" corresponding to the seven booklets of the manuscript.31 For instance, in the first booklet the foliation is in red, in various colors thereafter. Styles of paraphs change frequently, occasionally within the booklets themselves. Sometimes headings and titles are in red, sometimes in the scribe's ink, underscored or picked out in red. While it is beyond the scope of

   Still, the inconsistency in decoration seems to me to argue against the production of the booklets within a religious house. There is no indication, for instance, of a "house style," which one might expect if the pieces were being copied and decorated by scribes and artists working within and perhaps trained within the same walls. Moreover, any signs of supervision of the copying of the text is "fitful" at best.32 Scribal signatures within the quires, especially those incorporating The Canterbury Tales, denote that at one time the folios were arranged in an order different from what has survived (see, for example, ff. 42-68 and 173-99, in which new systems of signatures begin in mid-booklet).

   In contrast, the adorned capitals, which mark the beginnings of pieces or major segments of longer pieces, argue for some uniformity applied at a date later than the copying. The formation of the capital letters and their attendant decoration (blue lombards with red sawtooth adornments to the left sides extending into lengthy sprays in the margins--admittedly a rather common type of decoration), apparently done by the same hand, signify a later effort to unify the disparate booklets in a very basic appearance, a view supported by the fact that the capitals were among the last bits of decoration added, the red swirls of the embellishments often having to swerve around or over red and blue paraphs inserted earlier (see, for example, ff. 13r and 190r).

   Material omitted from the Tales has led to more speculation about the methods of producing the manuscript. The most famous examples are the omission of the Shipman's Tale, noted above, and the truncated version of the Parson's Tale, which ends abruptly, leaving blank folios ruled and numbered (see illustration 6). Manly and Rickert remain silent about these instances except for the previously mentioned suggestion of ecclesiastical censorship. Owen offers that the loss of the portion of the Parson's Tale may indicate that the scribe found copying this lengthy piece a "chore" and simply quit.33 To me, this hardly seems likely to have been the case in a religious house.

   I think there is a much better explanation, especially for the end of the Parson's Tale, which seems more indicative of the problems of producing Chaucer in the rural regions. Ralph Hanna has argued that the problems with links found in the Hengwrt MS of The Canterbury Tales show it was put together "from a series of separate exemplars, acquired separately, and from exemplars that did not, in all cases present tales with links."34 He concludes that the Hengwrt "production team acquired the exemplars of the poem piecemeal and were thus always at the mercy of their sources of supply," a view voiced earlier by Edwards and Pearsall.35 This sort of piecemeal supply would serve to explain the losses in Harley 7333. Indications in the Harley signatures and foliation show also that the arrangement of the Tales was in a state of flux as the booklets were being prepared. That the Parson's Tale breaks off abruptly and yet is followed by two ruled and foliated pages, as well as a stub, which was included in the foliation, suggests that the scribe or compiler was aware that his exemplum was lacking and that he perhaps intended to acquire the missing lines and fill in the quire at a later date. Given

   Harley 7333 is an important manuscript not for authoritative readings for a text but for the clues it offers for the publication of Chaucer outside of London. Such codices as the Harley manuscript also beg important questions about Chaucer's work as it appeared in the fifteenth century and the audience it enjoyed. In an analysis of Chaucer's audiences, Paul Strohm, for example, notes that "the role of ecclesiastical persons among the literary publics of late fourteenth-century England is still to be examined."36 Harley 7333 suggests that further examination of this audience, particularly in light of the emendations made in the text of the Tales (and perhaps many more alterations may exist than the few noted in this essay), may shed further light upon the role of these persons into the fifteenth century. In a different article, Strohm also notes the tendency of a fifteenth-century audience of different "strata" to perceive of Chaucer and his works with a greater sense of "traditionalism,"37citing Shirley's headnote to the Tales of Harley 7333 (see illustration 1) as evidence. As Strohm and others continue to revise the views of how Chaucer's work may have been received by later audiences and what that perception may yield for our own understanding of his work, rigorous analyses of the texts of such manuscripts as Harley become all the more important. To be sure, a knowledge of how the texts were being circulated and why those who obtained the manuscript copies sought them will bring us to an even surer understanding of Chaucer's reading public.

   Certainly, Chaucer's lines were not sacrosanct in the fifteenth century. Fifteenth-century scribes, like John Shirley in London, felt free to enter into the text and to make changes. Surprisingly, in Harley the Tales do not seem to be a centerpiece of the collection. Its beginning mid-quire suggests the order of an exemplar or the order in which the Tales became available to the copyist. Exemplars came to the copyists in bits and pieces, reflective of what some scholars see as the initial state of the Tales as it began to circulate n London. Thus gaps may occur, pieces left incomplete, as the copyists awaited more material or a different exemplar. Certainly a group such as the Austin Canons would have been interested in such material. But I am not convinced that they were interested so much in producing the piece as in acquiring the piece, by commission of the booklets when exemplars became available or by the purchase of them when copying was completed. Yet Harley 7333 indicates the popularity of Chaucer's work and the means, however roundabout, by which it circulated 100 miles from London.