Chapter 3 - Supplementing

Summary

This chapter demonstrates the influential role of printed books in defining Chaucer’s canon and the implications of that newly formed canon for older books. It considers a series of texts which early modern readers added to manuscripts (and some early printed books) to update and improve them: short poems including Prophecy, Words to Adam, and Bon counsail; Robert Henryson’s Testament of Cresseid; various Plowman-themed texts; the Tale of Gamelyn; and the Retraction. The chapter argues that print made available an array of genuine and apocryphal works which readers could extract, assemble, and reconfigure in line with their own tastes and understanding of the Chaucer canon. The evidence collected in the chapter shows the persistence of particular narratives about Chaucer’s works which were promoted in print: that he was a poet of fin amour, that he condemned Criseyde to a wretched death, that he assigned his Plowman an anticlerical tale, and that the Retraction was a later monkish forgery. The changeability of the manuscript books chronicled in this chapter reflects a concurrent reshaping of Chaucer’s reputation in the period and the variability of his literary canon itself.

‘Every age’, Helen Cooper has observed, ‘remakes Chaucer in its own image, just as he remade old books in a new image’.¹ Print was a major site of this early modern refashioning of the medieval past, and older manuscript books inherited by the later period bear material traces of those readings which circulated in and were promoted by the new medium. Accordingly, some medieval manuscripts reflect the early modern remaking of Chaucer as it happened. Books, whose pages could be annotated, expanded, and excised, were a convenient medium for readers to revise and augment the canon with various, even narratively or ideologically opposed, texts. Old manuscripts and printed books deemed to be imperfect or mutilated were the most obvious candidates for readerly perfecting, but they were not unique in this regard. Sometimes, readers were explicitly encouraged to perfect books by the editors themselves. The 1687 reprint of Speght’s edition concludes with an addendum, appended ‘[w]hilst this Work was just finishing’. In this ‘Advertisement’ the editors explain that, very late in the printing process, they ‘hapned to meet with a Manuscript’ containing the ends of the incomplete Cook’s Tale and the Squire’s Tale.² ‘[C]oming so late to our hands’, continue the editors, ‘they could not be inserted in their proper places, therefore the Reader is desir’d to add them, as here directed’. The Advertisement prints the missing lines, which are preceded by instructions to the reader on where to insert them: ‘Immediately after what you find of the Cooks Tale, add this:’ and ‘Immediately after these words, at the end of the Squires Tale, … Let this be added’.³

At least one reader took these instructions seriously. On 10 December 1663, Samuel Pepys visited St Paul’s Churchyard to peruse and purchase books and reports ‘seeing Chaucer’ but finally opting to buy some other titles. But he did come to own a folio Chaucer by the following summer, when he recorded taking it to be bound and clasped on 8 July 1664:

So to Paul’s churchyard about my books – and to the binders and directed the doing of my Chaucer, though they were not full neat enough for me, but pretty well it is – and thence to the clasp-makers to have it clasped and bossed.⁴

Pepys’s Chaucer – a copy of Speght’s 1602 edition – is held at Magdalene College in Cambridge, in the binding he describes in the Diary. Its calf covering, blind tooling, and brass clasps all signal Pepys’s penchant for adding distinguished bindings to his books.⁵ In 1664, Speght’s 1602 Chaucer, despite being the most complete and recent edition, was itself an old book, and significantly older than the thirty-one-year-old Pepys. It is little wonder, then, that he oversaw the perfecting of this prized copy once it had been superseded by the 1687 reprint. Sometime after that book’s publication, he had a new leaf added to his printed Chaucer, which reproduced the Advertisement concerning the Cook’s and the Squire’s Tale. Copied out by an amanuensis, the transcribed text supplies the conclusions wanting in the 1602 edition and thus follows the later edition’s instructions that ‘the Reader’ should ‘add them as here directed’. Yet Pepys’s scribe also diverged from those directions by making the additions not in ‘their proper places’ in the newly reprinted edition of 1687, but as a means of bringing the older volume up to date.

Although buying the latest and most complete edition seems an easy solution to the problem of finding oneself with an outdated copy of Chaucer – a solution vigorously promoted by the printed book trade – it was not the only option available to readers. Early modern collectors also had the possibility of finding creative ways to supplement their existing copies, and the example of Pepys illustrates that early modern perfecting was not exclusive to centuries-old manuscript or printed books. Pepys’s Chaucer, with its newly furnished but spurious endings for two of the tales, shows the ability of new editions to circumscribe as well as to stretch the limits of the accepted canon, and to render older editions obsolete. Both the directive in the 1687 edition and Pepys’s insertion of the text into an older copy confirm the contemporary sense of the book as ‘relatively malleable and experimental – a thing to actively shape, expand and resituate as one desired’.⁶ Old printed books and medieval manuscripts inherited by the early modern period were no exception to these practices.

Medieval manuscripts of Chaucer’s works could suffer from an appearance of obsolescence and outmodedness, but they also possessed the authority of age. It was, after all, the fact that the editors ‘hapned to meet with a Manuscript’ that facilitated the enlargement of the 1687 edition with the conclusions for the two incomplete tales. Manuscripts which lacked newly printed Chaucerian texts might be treated as simultaneously authoritative and somewhat out of date. Their antiquity made them old enough to be treasured as valuable and rare objects, but also old enough to benefit from further expansion and supplementation. It is a peculiar characteristic of many of the volumes discussed in this chapter that they embody both medieval and early modern attitudes to Chaucer and his works. As a consequence of their material adaptability, these copies often toggle between the beliefs and tastes of the people who made them and those who later read and reinterpreted them. They reveal an early modern understanding of the material book as open to – and importantly, capable of being improved by – readerly revision and renovation.

Chapter 2 argued that the missing leaves, blanks, and gaps supplied and filled by early modern readers in Chaucer’s medieval manuscripts constitute a form of perfecting which privileged ideas of bibliographical completeness influenced by models found in print. The following discussion enlarges that scope to consider books which did not show signs of damage or glaring incompleteness, but which were nonetheless perceived as wanting or inviting expansion. However curious completed, patched, or repaired volumes may initially appear to modern scholars, it is not difficult to understand the motivations that led pre-modern readers to fill the gaps in texts and to perfect their old books, especially when seemingly authoritative exemplars could be easily located. It is less obvious, but essential to literary history, to imagine why readers chose to supplement seemingly complete old books with new texts, and why they chose the texts they did. As this study illustrates, in the early modern period successive printed editions set the standard for what a correct, complete, and authoritative Chaucerian book should be. The present chapter argues that print made available expansive and apparently complete versions of the canon which readers extracted, assembled, and reconfigured in line with their own tastes and beliefs about what Chaucer wrote. This early modern sense of Chaucer’s works as contained in definitive printed volumes is given visual expression in the mid-seventeenth-century triptych commissioned by Lady Anne Clifford (1590–1676) and known as the Great Picture. Amongst the forty-eight labelled volumes in the painting is a copy of Chaucer in folio, which is titled ‘All Geffrey Chaucers Workes’ on the fore-edge. Of note here is the promise of exhaustive coverage signalled by ‘all’, a word which does not appear in the title of any editions of the Workes, but which almost certainly refers to the contents of a 1602 copy of Speght owned by Clifford.⁷ To examine the ways that readers transformed their Chaucer manuscripts in line with the expansive and seemingly definitive prints is to witness the reshaping of the poet’s post-medieval reputation, and to understand the role played by the new medium in forging his reception.

Despite its appearance of comprehensiveness, however, the print canon in which most early modern people read their Chaucer was far from fixed. Chaucer’s name and fame had a magnetic effect in the early modern period, causing editors and readers to attach new and varied texts to him. With this expansion of the canon, old books fell out of fashion faster.⁸ The redrawing of the lines between accepted and apocryphal works had significant material effects on how Chaucer was read, as readers updated and supplemented their older copies to reflect a canon that was regularly in flux. The expanding of Chaucer’s manuscript books by medieval and early modern readers alike therefore points to the more fundamental variability of his literary canon itself. The histories of textual transmission explored in this chapter show canonical texts rubbing shoulders with texts today excluded from the canon and readers grappling with the ambiguity of their presentation in print. The bibliographical history of spurious, apocryphal, and ambiguously positioned texts – most of them once assigned to or implied to be by Chaucer in authoritative editions or commentaries – offers a corrective to modern scholarship’s preoccupation with attribution and authenticity and a reminder that Chaucer, as a historically constituted entity, has always been subject to reinterpretation.⁹ The texts that early editors and readers once attributed to him, and the array of justifications they had for such choices, form a vital chapter in the history of Chaucer’s literary afterlife.

3.1 Commonplacing Chaucer

One defining feature of the monumental editions of Chaucer printed in the sixteenth and seventeenth centuries is their array and inclusiveness.¹⁰ Although Chaucer’s name appears prominently on the title pages of the folio volumes of his works, the editions produced by Thynne, Stow, Speght, and their collaborators also promoted a Chaucer under whose authoritative umbrella other Middle English works could conveniently cluster. This progressive expansion of the print canon was the result of a dual impetus: first, to recover those works of Chaucer which had never before been published; and second, to increase the editions with thematically or linguistically similar works in order to market the number of new texts on offer. Kathleen Forni notes that ‘early editors do include genuine works never before printed, but also poems overtly attributed to, or known to be by, other authors’. She cautions, however, against the temptation to ‘dismiss these editions as simple miscellanies modelled on the manuscript canon, since these books were sold as Chaucer collections’.¹¹ In the early prints, many works therefore occupied an unsteady middle ground between genuine and apocryphal status, and contemporary readers were left to grapple with the ambiguity that such configurations present. Sometimes this precarious canonicity was a result of codicological instability, as in the case of short poems, lyrics, or ballades. Julia Boffey has advanced the view that these texts may have been ‘registered by Chaucer on perilously unattached single leaves and fragments’, that the early copies circulated in ‘sometimes confused and confusing forms’, and that the mid-fifteenth-century manuscript collections and Caxton’s early prints ‘offered these essentially ephemeral poems a securer environment’.¹² The brevity that made these texts prone to loss, corruption, and variance also assured their material portability – that is, their capacity to be easily extracted, re-copied, and thus preserved in new bibliographic contexts.

Chaucer’s late medieval and early modern reputation for axiomatic wit owes much to the genuine and apocryphal works, rich in proverbs and sententiae, that circulated under his name in fifteenth-century manuscripts, in Caxton’s early quarto editions of the Parliament of Fowles and Anelida and Arcite, and in the folio canon inaugurated and progressively expanded during the sixteenth and seventeenth centuries.¹³ Jennifer Summit has shown, for example, that Pynson’s 1526 edition of the Book of Fame took pains to frame Chaucer as a moral authority – a designation made explicit in its inclusion of ‘certayne morall prouerbes of the foresaid Geffray Chaucers doyng’ – and has demonstrated that the printer appropriated the writing of Christine de Pizan in this effort to establish the poet’s sententiousness.¹⁴ The enthusiasm of Renaissance readers for Chaucerian aphorism is witnessed by marginalia in the printed Chaucer folios surveyed by Alison Wiggins, who observes that readerly attention to proverbs and sententiae in the Wife of Bath’s Prologue and Melibee makes those tales ‘by far the most frequently and heavily annotated of Chaucer’s works’.¹⁵ Francis Beaumont, a friend of Speght and supporter of the 1598 edition, deemed Troilus to be ‘so sententious, as there bee fewe staues in that Booke, which are not concluded with some principall sentence’.¹⁶ In the same edition, Speght himself lamented the fact that a lack of sufficient time had prevented him from noting ‘[s]entences also, which are many and excellent in this Poet’ and which ‘might haue ben noted in the margent with some marke’.¹⁷ In the subsequent edition, Chaucer’s sententiae were indeed marked out with printed marginal hands or maniculae, and the book’s status as a repository of ‘Sentences and Proverbs’ was heralded on its new title page.¹⁸

It is in this context of the editorial promotion of Chaucer as a moral authority and writer of aphorism that certain supplements made by early modern owners to their medieval manuscripts should be examined. In CUL, MS Gg.4.27 (Gg), an anthology of Middle English poetry, the book’s medieval makers appear to have focussed on an idea of Chaucer’s oeuvre as their chief organising principle. Joseph Holland’s project of perfecting that book, which is discussed throughout this study, is a perceptive elaboration of the manuscript’s Chaucerian theme, for his additions show a keen awareness of Chaucer’s authority as a literary figure.¹⁹ At the same time, he also recognised and added to the formal and generic miscellaneity of the original manuscript, which already included Lydgate’s Temple of Glass as well as shorter items which were also unassigned to any author.²⁰ In addition to an extract from Henryson’s Testament (to which I will return) Holland had his scribes supply the manuscript with lyric poems and poetic excerpts which he read in Speght’s first edition. They appear in Holland’s book under the following titles:

‘Chaucer, touchinge gen[tle]nes of Birthe: or who is worthy to be called gentill’ (Gentilesse, IMEV 3348; fol. 1^v and Canterbury Tales 111.1117–24)

an untitled extract from the Parliament of Fowles (IMEV 3412; fol. 4^v)

an untitled extract from the poem then known as ‘Chaucer’s Prophecy’ (Prophecy, IMEV 3943; fol. 4^v)

‘Bon counsail’ (Yit of the Same, IMEV 3521; fol. 35^r)

‘Chaucer to his emptie purse’ (Purse, IMEV 3787; fol. 35^r)

‘Chaucers words to his Scrivener’ (Words to Adam, IMEV 120; fol. 35^r)

The excerpt from the Parliament of Fowles (ll. 22–5) had also appeared in a cartouche at the head of the architecturally-styled title page of two issues of the 1598 edition (see Figures 3.1 and 3.2).²¹ The chosen lines rehearse a well-known Chaucerian quatrain:

Out of the old fields as men sayth,

Commeth all this new corn fro yere to yere;

And out of old Books in good faith,

Cometh all this new science that men lere.

Holland was partial to these lines, for he also noted them with his characteristic annotation mark where they appear in Gg’s complete text of the Parliament (fol. 481^r). It is not hard to account for their appeal to an antiquary, collector, and bibliophile like Holland. For him, and no doubt for Speght, Chaucer’s words encapsulated the guiding ethos of the antiquarian project: to recover ‘new science’ or knowledge that lay dormant in the ‘old Books’ collected, preserved, and studied by Stow, Holland, and their circle. On Islip’s title pages the lines are contraposed against a quotation from Ovid’s Metamorphoses placed at the foot of the page, and both extracts are neatly framed within the classicising woodcut border selected for the occasion. The Ovidian line ‘Seris venit usus ab annis’ (‘Experience comes with riper years’), a reproach delivered to the youthfully brazen Arachne by Minerva in Book 6, echoes the Parliament’s sentiment about the value of the old.²² The oeuvre of Chaucer, like that of Ovid, is iconographically rendered here as an enduring literary monument worthy of memorialisation and quotation. The Chaucerian quatrain would have been recognised in the Renaissance as a simile or similitude, a rhetorical device whose brevity and epigrammatic wit made it popular for commonplacing.²³ Chaucer had been singled out in this regard by George Puttenham’s immensely influential Arte of English Poesie (1589): ‘his similitudes comparisons and all other descriptions are such as can not be amended’.²⁴

Figure 3.1 Title page of Speght’s 1598 edition (STC 5078) with a cartouche containing a quatrain from the Parliament of Fowles.

Fondation Martin Bodmer [without shelfmark], sig. [a]2^r. Digitised and reproduced courtesy of the Bodmer Lab, University of Geneva.

Figure 3.2 An extract from the Parliament of Fowles, the short poem Prophecy, and praise of Chaucer from Speght in Holland’s manuscript. CUL MS Gg.4.27(1), fol. 4^v.

Reproduced by kind permission of the Syndics of Cambridge University Library.

With the exception of the lines from the Parliament, which could be seen in pride of place on the Islip title pages, the verses that were newly transcribed into Gg were not especially marked out in the 1598 printed edition. Rather, they represent a series of telling literary choices made by Holland as he sought to supplement his manuscript. Beneath the lines from the Parliament, the scribe copied the six-line tetrameter poem Prophecy (‘When faith faileth in priests saws’), while the rest of the page (fol. 4^v) is dedicated to a description and transcription of Lydgate’s praise of Chaucer. Prophecy was frequently re-copied in the medieval and early modern periods, making it ‘the most popular apocryphal work’.²⁵ It is variously assigned to Chaucer or Merlin in early manuscripts, and was ‘overwhelmingly attributed’ to Chaucer in the early modern period.²⁶ First printed by Caxton at the end of Anelida and Arcite (c. 1477), perhaps as filler material, the poem’s three rhyming couplets foresee a world in which moral failures – such as ‘robbery’, ‘lechery’, and loss of faith – will ‘Be brought to grete confusion’.²⁷ It was later reprinted without explicit attribution by Thynne, Stow, and Speght, all of whom consistently placed it before the beginning of the Canterbury Tales within the Workes. In this position (as elaborated later), it may have served once again as filler material used to complete a gathering. Holland seems to have appreciated the poem’s nostalgic sententiousness as a pairing for the Parliament’s lines about the emergence of new learning from ‘old Books’. Speght, who did not provide an attribution for Prophecy, followed the earlier folios in placing the text in the preliminaries (sig. ¶4^v). As Weiskott points out, from the skewed perspective of a literary history which has Chaucer serve as flagbearer for English pentameter verse, the popular tetrameter Prophecy appears to be a metrical anomaly. But Holland, like the early scribes and editors who paired it with pentameter poems, registered its oracular tone and its long view of history as an appropriate accompaniment to his favourite lines from the Parliament.²⁸

Another newly added poem with an edifying moral message, Gentilesse, was positioned at the very beginning of the manuscript when Holland owned it.²⁹ In its dispensation of worldly wisdom about the nature of true nobility – ‘For unto vertue longeth dignite’ (l. 5) – the transcription into Holland’s manuscript of Gentilesse affirms Chaucer’s posthumous standing as a paragon of moral instruction and courtly counsel. Underneath the transcript of the poem is an eight-line extract on the Chaucerian theme of ‘gentlenes’ from the so-called pillow lecture of the Loathly Lady to the Knight in the Wife of Bath’s Tale.³⁰ Both Gentilesse and the Lady’s words emphasise the ‘noblesse’ of character which makes men true ‘gentlemen Icalled bee’ and which, unlike inherited titles or riches, may not be bequeathed by one’s ancestors. In Speght, Gentilesse is embedded amongst a series of ‘Certain Balades’ under the heading ‘A Balade made by Chaucer, teaching what is gentilnesse, or whom is worthy to bee called gentill’.³¹ In Gg, both texts appear under a heading which has been slightly adapted: ‘Chaucer, touchinge gen[tle]nes of Birthe: or who is worthy to be called gentill’. As an amateur herald, one of Holland’s chief scholarly preoccupations was the study of ancient coats of arms, including that of his own family.³² His activity as an antiquary during the late sixteenth century and first years of the seventeenth coincided not only with a period of Chaucer’s soaring print popularity, but also with a growing enthusiasm in England for arms, pedigrees, and genealogies.³³ In light of the role of the College of Arms and its heralds as the granters of arms and arbiters of claims to gentility, Holland’s interest in Chaucer’s passages on ‘gentlenes of Birthe’ is worth pausing over. His taste for aphorism, his notion of Chaucer as a moral author, and his own interests in gentility and genealogy all informed his selection. In Chaucer, Holland saw a poet whose own ‘gentilesse’ was evinced both by his sententious works and a noble and royal lineage.³⁴ It is conceivable that Holland strove to cultivate a similar mode of gentility for himself – one defined as much by the arms he was granted in 1588 as by the pursuit of ‘vertuous liuing’ and the reading of edifying literature.³⁵

The last of Holland’s interpolated leaves (fol. 35) presents three additional short poems which extend the theme of morality and add an authorising biographical note to the manuscript (see Figure 3.3). The copied texts are all one stanza long and are titled ‘Bon counsail’, ‘Chaucer to his emptie purse’, and ‘Chaucers words to his Scrivener’. Together, these three stanzas written in Chaucerian rhyme royal fill the recto of fol. 35, but they are not printed as a set in Speght’s edition. The poem called ‘Bon counsail’ in Holland’s Gg is the same text which Speght titled simply ‘Yet of the same’, and which follows a one-stanza poem titled ‘A Saying of Dan John’.³⁶ In the printed editions, then, the poem is clearly attributed to Lydgate.³⁷ The generic title ‘Bon counsail’ in Holland’s manuscript, however, obscures these Lydgatean origins and focusses instead on the poem’s advice to embrace humility over folly. The two remaining poems on fol. 35^r both name Chaucer in the title and were no doubt of biographical interest to Holland. In Speght, ‘Chaucer to his emptie purse’ is buried within an earlier section of ‘Balades’ (sig. 3O6^v) first printed by Thynne. Meanwhile, the poem which appears in Speght as ‘Chaucers wordes vnto his owne Scriuener’ becomes ‘Chaucers words to his Scrivener’ in Gg. Chaucer’s witty reprimand of an errant scribe, the poem appears at the terminus of the first part of Speght’s edition (sig. 3T4^v), where it immediately precedes the explicit ‘Thus endeth the workes of Geffray Chaucer’ and the woodcut title page border used to introduce Lydgate’s Siege of Thebes on the facing page. The appearance of the poem at the end of Holland’s Chaucer manuscript echoes its placement in the printed editions from 1561 onward where, as Cook has argued, it allows Chaucer an authoritative ‘last word’ in his own voice.³⁸ Unlike Chaucer’s address to his purse, for which there is evidence of transmission of the text and variants in at least a dozen medieval manuscripts, the poem now known as Adam Scriveyn is extant in a single manuscript witness.³⁹ The decision of Stow, and later Speght, to print the poem at the very end of Chaucer’s works in the folios gives that text ‘the summative power of an envoy’⁴⁰ – a valedictory function that Holland aimed to replicate when he had it transcribed into the end of his manuscript compilation.

Figure 3.3 Short poems added by Holland to CUL MS Gg.4.27(1), fol. 35^r.

Reproduced by kind permission of the Syndics of Cambridge University Library.

Holland’s sense of the Chaucerian canon was profoundly shaped by a knowledge of the body of texts and their paratexts that circulated in the print canon. As a result, it has been suggested that he ‘seems to have regarded the collection as analogous to an early printed edition of Chaucer’s “Workes”’ and that the additions are material ‘which a sixteenth-century reader had come to expect in a copy of Chaucer’.⁴¹ This assessment is fair, but it understates the literary judgement and selectivity that characterise Holland’s act of enlarging Gg. By 1600, he had access to dozens of printed Chaucerian lyrics which might have been appropriate for inclusion in his own manuscript. That he chose only a handful of particular lines from long poems and certain lyrics shows that Holland’s book, for all its comprehensiveness in relation to other Chaucer manuscripts, was not intended to be an exhaustive collection that would achieve the same coverage as the capacious printed folios. Rather, as in any commonplace book of the period, the short texts added to Gg preserve a record of reading and literary taste, which in this case cohered in a particular vision of Chaucer as a sententious author.

Like Holland, the person who supplied missing text in a late fifteenth-century copy of the Canterbury Tales now at Trinity College in Cambridge during the sixteenth century also took the opportunity to add a revealing series of supplementary items to the manuscript.⁴² In the beginning of this book, an extra quire has been added to accommodate the missing text of the General Prologue and the Knight’s Tale. Its first leaves were then filled out with additional lyrics now known to be apocryphal: Eight Goodlie Questions (IMEV 3183); a pair of Hoccleve poems, To the Kings Most Noble Grace and To the Lordes and Knyghtes of the Garter (IMEV 3788 and IMEV 4251); and Prophecy (IMEV 3943).⁴³ The placement and sequence of these items exactly reproduces their appearance in the printed book from which they were transcribed, Thynne’s 1532 edition. In that printed book, these texts appear in a prominent if awkward place, between the list of contents (‘table of al the workes’) contained in the volume and the beginning of the first substantial text, the Canterbury Tales.⁴⁴ In Greg Walker’s reading, this sequence of short texts comprises a ‘vitally important’ part of the edition’s ‘poetry of moderation’: tactful political and moral counsel directed at the increasingly tyrannical Henry V111, to whom the book is dedicated and in whose household Thynne worked as ‘chefe clerke of your kechyn’.⁴⁵ Whether the positioning of these texts reflects a purposeful insertion at a visible point in the book which might catch the King’s attention (and his conscience), or whether it results mainly from a more prosaic need to fill the last leaf of the gathering with any printed text, their transcription into the Trinity manuscript confirms their attractiveness to early modern readers, and establishes Thynne’s edition as responsible for later interpretations of these apocryphal texts as Chaucer’s.⁴⁶ Knight notes that their placement in Thynne ‘confers on them the status of a preface’, and it is this bibliographical and rhetorical function that another reader reproduced in an older manuscript.⁴⁷ They may have occupied a ‘curiously liminal position’ in Thynne’s Workes, but that very placement caused the later book owner to select this sequence of short poems as the best material for the newly supplied preliminary leaves in their own manuscript of the Canterbury Tales.⁴⁸ That choice to enhance the book by supplementing it with these poems is also recorded in the manuscript’s table of contents, where the three items occupy the prime position at the head of the list. The trio of added texts again underscores Chaucer’s early modern reputation for proverbial wisdom and shows that the form of lyrics continued to make them particularly convenient for extraction as filler material or ‘makeweights’ in manuscripts as well as in printed books.⁴⁹ More broadly, these material interpolations show the effectiveness of print in promoting particular versions of Chaucer – in this case, a sententious one – to early modern readers.

Early readers of the Gg and Trinity copies thus responded to print’s presentation of Chaucer as a moral authority and treated their manuscripts like personal anthologies that could be expanded to accommodate this sententious matter. Another medieval book, the Fairfax manuscript, appears at first glance to have been supplemented from print following another principle – not with the aim of assembling a repository of gnomic wisdom and pithy sayings by and about Chaucer, but according to a thematic focus on courtly love. In the Fairfax manuscript, the same seventeenth-century hand that filled in gaps in the Book of the Duchess and the House of Fame also supplied a new text, the Ten Commandments of Love (IMEV 590), a lyric poem of fourteen rhyme royal stanzas about how women should conduct themselves in matters of the heart. This apocryphal work was first printed by Stow in 1561, and was included in the printed editions of Speght, Urry, and subsequent editions until Thomas Tyrwhitt’s edition (1775–8).⁵⁰ In Fairfax, the poem has been copied from Speght’s 1598 edition onto two blank leaves supplied by the original scribe for the ending of the House of Fame (fols. 184^r–185^v).⁵¹ In Speght, Ten Commandments is printed with other short works under the heading ‘Here followeth certaine workes of Geffray Chaucer, annexed to the impression printed in the yeare, 1561. With an adition [sic] of some things of Chaucers writing, neuer before this time printed, 1597’.⁵² Seven lyrics, including Gentilesse, precede Ten Commandments, and the running head ‘Certain Balades’ unites these disparate titles, which are generally courtly in nature.⁵³ Whatever Speght and his predecessors may have believed about the authorship of the Ten Commandments of Love, it is clear that the text’s placement in the 1598 edition would have led some readers to assume that it was written by Chaucer. Ten Commandments is written in rhyme royal and belongs to the lyric tradition of advice to lovers. It may have been these formal and thematic characteristics, which echo those found in other Fairfax texts, that caused the annotator to select it for inclusion in the manuscript. But it is also possible that the annotator took Speght’s edition as an authority on Chaucer’s text and the canon, and included Ten Commandments on the assumption of its genuineness. The early modern addition of Ten Commandments to Fairfax offers an instructive counterexample to the book’s incomplete House of Fame, for which the same annotator also supplied an ending.⁵⁴ While the latter is a visibly incomplete text which its seventeenth-century annotator reasonably set about to conclude, Ten Commandments was not conceived as part of the medieval manuscript. In both cases, however, the early modern annotator identified obvious or inviting gaps in the manuscript book and chose to fill them in, completing and enhancing the original manuscript in the process.⁵⁵ The origins of the Gg, Fairfax, and Trinity annotations in contemporary editions reveal the active role of printed books in shaping literary taste.

At the same time, the status of Gg and Fairfax as anthologies is a reminder that the reconfiguring of texts into new collocations was a literary and readerly activity older than print. The modular potential of the written text was well known to Chaucer, whose poem which he called ‘the love of Palamon and Arcite’ was conceived as a stand-alone work before it was revised to become the opening text within the collection of his Canterbury Tales.⁵⁶ Just as the introduction of texts into new bibliographic contexts during the early modern period was not without precedent, neither was the extraction of sententiae seen in the books studied here. The copying of aphoristic or amatory poems from sixteenth- and seventeenth-century printed copies onto blank pages or supplied leaves in medieval manuscripts recalls not only the diligent commonplacing and note-taking practised by humanist readers, but also the widespread medieval practice of compiling florilegia.⁵⁷ As such readers knew well from the Latin root of read (lego, legere, to read, to gather, to choose), thoughtful excerption was essential to the utility of reading as a means of moral education.⁵⁸

These continuities notwithstanding, the act of extracting printed texts or their fragments for inclusion in medieval manuscripts belongs to a context of reading and compilation which is different from that which initially produced anthologies like Gg or Fairfax in the fifteenth century. The printed books in which early modern people read their Chaucer were widely disseminated and presented a panoply of standardised and authorised texts which (as we have seen) could be extracted, re-copied, and used to supplement other books. In gathering texts under the authenticating rubric of Chaucer’s name and presenting them as the individual parts of a defined canon, the printed folios brought to the fore concerns which were largely secondary to medieval compilers and readers of manuscripts.⁵⁹ More so than the older manuscripts, the printed volumes of Chaucer’s Workes emphatically participated in the monumentalising of the author and his writings. They altered the historical and material circumstances in which the poet was read and venerated, and they presented a canon of his works which was ripe for extraction and quotation at the hands of readers. The use of those printed books as a basis for supplementing fifteenth-century manuscripts with new texts offers striking evidence for the effectiveness of the book trade’s promotion of Chaucer and his works.

Manuscript miscellanies compiled during the sixteenth century by George Bannatyne (c. 1568) and by readers in the Tudor household of Anne Boleyn (in the 1530s and 1540s) also document the use of printed poems in editions of Chaucer as material for excerpting and adaptation. In the Scottish Bannatyne manuscript, a set of mostly apocryphal poems transcribed from Thynne’s c. 1550 edition (or a later one based on it) is attributed to Chaucer and grouped under a section titled ‘Ballatis of Luve’.⁶⁰ In this position, a Chaucerian author-figure is made to ventriloquise differing stances on the contemporary querelle des femmes and contributes to the ‘inherent moral trajectory’ of the manuscript as designed by its compiler.⁶¹ In the early Tudor Devonshire manuscript, verse extracts from texts printed in Thynne’s edition were likewise copied in order to voice varied positions in the debate about women’s morality, and were sometimes freely adapted to this end. The poetic selections are from Chaucer’s Troilus and Criseyde and Anelida and Arcite, Hoccleve’s Letter of Cupid, and Sir Richard Roos’s La Belle Dame Sans Merci, but none of these extracts is attributed to an author in the manuscript.⁶² Whereas the compiler of Bannatyne relied on extracts bearing Chaucer’s name to authorise competing points of view in the querelle, the courtly Devonshire manuscript witnesses the elision of the author’s presence – sometimes twice over, as in the case of Troilus, where the extracts are detached not only from their prominently named author, but from the much longer and famous poem from which they derive. These sixteenth-century compilations provide valuable context for the early modern practice of copying poems of love or counsel from print into manuscript. Collectively, the supplemented medieval books and the early modern anthologies show that such practices of extraction and supplementation were not anomalous. They provide evidence of a widespread readerly desire to extract, adapt, and reconfigure the contents of Chaucer’s printed books to new ends.

From the printed editions, the readers I have been discussing in this chapter selected appropriate texts with which to enhance their medieval manuscripts, but they often did so in ways that complement, accentuate, and make meaningful pre-existing features of the older books – the sententiousness of Gg, for instance, or the thematic emphasis on love in Fairfax. The three prefatory lyrics added to the Trinity manuscript of the Canterbury Tales, meanwhile, witness a process of purposeful supplementation which aimed to retrofit the medieval book according to a textual order seen in the later print. The willingness of their owners to introduce such additions suggests a contemporary sense of the medieval manuscripts as open-ended; for all their antiquity and value, they were prime for iterative expansion and personalisation with newly available Chaucerian texts.

These readers were at home in a culture of compilation that saw groups of disparate texts at turns gathered into (sometimes only loosely defined) print and manuscript anthologies and dispersed again by recopying and excerption. Works such as Prophecy, Ten Commandments, the Parliament of Fowles, and Troilus and Criseyde might convey some semblance of stability from within the covers of a folio of Chaucer’s printed Workes, but such appearances are often illusory.⁶³ The attribution of Prophecy and Ten Commandments to Chaucer is ambiguous in the prints, and the status of the former as preface or filler material is undetermined; the Parliament of Fowles sees its crowning similitude excerpted as an epigraph on a printed title page; and (as I discuss next) the relation of Chaucer’s great tragedy to the slighter companion poem that follows it in print is cast into doubt. Moreover, texts collectively designated ‘works’ and assembled into authoritative editions were still liable to be removed from that bibliographical context and made to serve in new configurations. In particular, the brevity and pithiness of the lyric form often lent itself to the extraction from print into manuscript charted here, and to the reticence about attribution which often accompanied this textual mobility.

3.2 Chaucer’s Troilus and Henryson’s Cresseid

With the radical expansion of Chaucer’s canon during the early modern period came a good deal of readerly interest in its contents and exclusions. But as with the lyrics and short poems, the placement and clustering of some longer narrative works in the printed folios made their connection to the poet ambiguous at best and misleading at worst. A manuscript now held at St John’s College in Cambridge illustrates the consequential effects of that editorial uncertainty on Chaucer’s literary reputation. The book is a fifteenth-century manuscript of Troilus and Criseyde, MS L.1 (L1). Like other medieval manuscript copies of Troilus, the text in L1 is not attributed to Chaucer by means of an incipit or explicit. The manuscript’s text was corrected by two annotators in the course of the seventeenth century, but the most substantial early modern addition to L1 is the Testament of Cresseid (IMEV 285), which was copied onto eight leaves of thick parchment inserted after the final quire of the original scribe.⁶⁴ This Middle Scots poem, composed in the fifteenth century by Robert Henryson (d. c. 1490), invents for its heroine a tragic fate of her own: her rejection by Diomedes, descent into prostitution and leprosy, and eventual death. Following Chaucer’s rhyme royal, the poem’s eighty-six stanzas are seven lines long, except for the seven stanzas comprising ‘The Complaint of Cresseid’, an embedded lament written in the style of Chaucer’s nine-line ‘Lament of Anelida’ stanzas (aabaabbab). This, at least, is the form of the Testament as it is read and studied today, in a text based on the 1593 edition printed by Henry Charteris in Edinburgh.⁶⁵ But it was another, textually inferior version which William Thynne included as a belated addition to his 1532 edition of Chaucer’s Workes – one that circulated widely during the sixteenth century and would be reprinted in all editions until 1721.⁶⁶ Although Thynne’s text is the earliest surviving witness of the Testament, its reliability has been questioned on account of its heavy anglicisation of Henryson’s Middle Scots and its muddling of the nine-line stanzas.⁶⁷ Despite its dubious textual value, this was the version with which many early modern readers were familiar, and it was the text selected to supplement a fifteenth-century Troilus manuscript in the seventeenth century.

On the one hand, the motivations behind this reader’s choice to import the Testament into a fifteenth-century manuscript of Troilus appear transparent; the transcription supplies ‘evidence of abiding interest of the story’.⁶⁸ Not only does the poem rely on Chaucer for its literary form and material, but its narrator frames his work with several hallmarks of Chaucerian dream vision: professed inexperience in love, bookishness, and a narrative self-awareness in relation to his sources.⁶⁹ On the other hand, criticism on the poem has long recognised that the Testament is not a mere continuation or sequel to Troilus but ‘an alternative ending’ which, in its pivot to the suffering of Cresseid, refracts and redefines the moral landscape of the earlier Middle English work.⁷⁰ Henryson’s narrator initially takes up Chaucer’s Troilus at the point of the lovers’ separation but decides to ‘nocht reheirs [not retell]’ (l. 57) the distress and heartbreak of the hero. Setting aside his Chaucer book, the narrator opts to read an enigmatic ‘vther quair [other book]’ (l. 61) in which he finds ‘the fatall destenie / Of fair Cresseid, that endit wretchitlie’ (ll. 62–3). More strikingly still, the poem professes at this point to doubt the authority of its literary sources:

Quha wait gif all that Chauceir wrait was trew?

Nor I wait nocht gif this narratioun

Be authoreist, or fenyeit of the new

Be sum poeit, throw his inuentioun

(ll. 64–7)

These lines have often been read as a repudiation of Chaucer’s literary authority and as Henryson’s own claim to vernacular ‘inuentioun’ and to the mantle of ‘poeit’.⁷¹ The Middle Scots poem looks back to Chaucer, but its look is slightly askance. It elides the narrative details of Troilus’s death and stellification at the end of Chaucer’s Book v and, as Spearing argues, ‘more boldly offers an antithetical misreading … in which Cresseid dies and Troilus remains alive’.⁷² This acknowledged friction and incompatibility between the two texts should inform our interpretation of a historical reader’s choice to pair them in L1. More than a straightforward comparison, the Testament invites juxtaposition with the earlier tragedy, and consciously presents itself as a worthy challenger to Chaucer’s Troy poem.

Many early modern readers looked past those contradictions, however, and seem to have treated both texts as authentically Chaucerian.⁷³ The sixteenth- and seventeenth-century folio editions of Thynne, Stow, and Speght themselves appear to register uncertainty on this crucial point. None of these collections attributes the work to Henryson – nor, indeed, does any explicitly name Chaucer as its author. Instead, in these editions the poem nestles suggestively between Troilus and the Legend of Good Women. The latter poem bears its own intertextual relationship to Troilus, since Chaucer’s portrayals of women dishonest in love – a group in which the poet includes Criseyde – furnishes the pretext for the Legend’s literary catalogue of virtuous women.⁷⁴ Whether or not Thynne and the editors who followed him believed the Testament to be an authentic work of Chaucer, printing the three poems in immediate succession in the collected works made sense to them.⁷⁵ They are silent, however, on the point of the Testament’s authorship. In Speght’s 1598 volume, for instance, the poem’s surrounding paratexts deftly thread it between Chaucer’s genuine works. Its incipit reads, ‘Thus endeth the fifth booke, and last of Troilus: and here foloweth the pitefull and dolorous Testament of faire Creseide,’ and its explicit likewise emphasises continuity with the work that follows: ‘Thus endeth the pitifull & dolorous Testament of faire Creseide: and here followeth the Legende of good women’.⁷⁶

One of Speght’s editorial innovations was the addition of ‘Arguments’ or summaries to the major texts in the volume, a feature advertised on the 1598 title page. Both of his editions supply an Argument for Troilus and Criseyde but not for the Testament. The 1602 Argument is representative:

In this excellent Booke is shewed the feruent loue of Troylus to Creseid, whome he enioyed for a time: and her great vntruth to him againe in giuing her selfe to Diomedes, who in the end did so cast her off, that she came to great miserie. In which discourse Chaucer liberally treateth of the diuine purueiance.⁷⁷

Speght’s Argument here is rife with ambiguity, for although he refers chiefly to the single ‘excellent Booke’ called ‘Troilus and Criseyde’, his summary includes incidents proper to the Testament (Diomede’s spurning of the heroine and her ‘great miserie’). Equally inconclusive is the fact that he says this happens ‘in the end’, leaving open the possibility that Speght is referring to the outcome at ‘the end’ of the love story as a whole, rather than to ‘the end’ of Chaucer’s Troilus. Speght’s editorial paratexts and very language thereby evade the question of the Testament’s authorship, neither affirming that Chaucer wrote it nor naming Henryson. Instead, he appears to treat the poems scholars now call ‘Troilus and Criseyde’ and ‘The Testament of Criseyde’ as different but related parts of a larger composite text, also called ‘Troilus and Criseyde’. This move sees Chaucer’s Troilus and Criseyde come to represent Troilus’s side of the tragedy, while Henryson’s poem relates its other half, namely Cresseid’s fate. Such a reading might be bolstered by the shape of Chaucer’s Troilus itself, beginning as it does with one aim – ‘The double sorwe of Troilus to tellen’ (1.1) – and concluding with its male protagonist’s lament, death, and stellification.

Even as it diverges from Troilus in its portrayal of a punitive moral universe and in its pivot to Cresseid’s suffering, the presentation of the Testament in Speght’s editions strongly implies Chaucer’s authorship. But irrespective of their views on the poem’s authorship, Speght and his fellow editors clearly identified the Testament as a worthy companion piece, or perhaps counter-narrative, to the longer work. By adding the Testament to his own book, the L1 copyist adopted and endorsed this contemporary reputation of the two texts as complementary and may have likewise assumed the work to be Chaucer’s. This interpretation is confirmed by the explicit furnished for the transcription. In the 1602 Speght edition no explicit had been printed, but the new scribe supplied one in a display script at the poem’s conclusion: ‘EXPLICIT LIBER TROILI, & CREISEIDOS’ (see Figure 3.4).⁷⁸ For the L1 copyist, too, it seems that the texts formed two halves of a single work called ‘Troilus and Criseyde’. The addition of the Testament to L1 and the furnishing of an explicit which treats the two texts as one ‘liber’ demonstrates the influence of the printed edition on this reader’s understanding of both poems and their authorship. MS L1 thus preserves valuable evidence of the staying power and interpretation of the apocrypha introduced into the canon by Thynne in 1532 and retained in subsequent editions. Fox has argued that several editions attributing the work to Henryson also appear to have been published in Scotland during the late sixteenth and early seventeenth century.⁷⁹ Despite the text’s apparently wide circulation under Henryson’s name in those books, Chaucer’s association with the Testament would only be dislodged in the eighteenth century. At the same time, however, L1 also registers a trace of doubt about the poem’s provenance. On an otherwise blank flyleaf, a seventeenth-century hand (possibly belonging to another owner) has written out the contents at the head of the page:

Chaucer’s Troilus and Criseid.

The Testament of Criseid.

(fol. i^r)

This arrangement of the two titles and the conspicuous absence of an author’s name in the second line might suggest some questioning of Chaucer’s putative role as author of the Testament. Another instance of readerly ambivalence concerning the poem’s authorship appears in a copy of the 1532 Workes now at the Beinecke Library, where a contemporary annotator has observed that while the content of the poem points to Chaucer as the author, ‘this meetre is not his’.⁸⁰

Figure 3.4 An early modern parchment supply leaf containing the end of the Testament of Cresseid.

Cambridge, St John’s College, MS L.1, fol. 128^v. By permission of the Master and Fellows of St John’s College, Cambridge.

The shared subject matter and literary heritage of Chaucer’s and Henryson’s texts also captured the attention of Joseph Holland, who oversaw the copying of four stanzas from the Testament (ll. 582–609) onto a parchment supply leaf inserted into Gg.⁸¹ In this case, the excerpted lines pertain to Criseyde’s last will and testament, an inset text which gives the poem its title. Sheridan has observed that early readers of the Testament would have recognised Criseyde’s ‘Testament’ (ll. 577–91) as ‘a special category of text embedded in the narrative’.⁸² The extract copied into Gg represents the tragic conclusion of the narrative, with a description of Criseyde’s dying act, her bequest to Troilus of a ‘roiall ringe’ that he had given to her as a love token, Troilus’s own grief on learning of her death, and his erecting of a marble tombstone and an epitaph for her grave. Holland’s excerption of the ten concluding lines (ll. 582–91) of the embedded ‘Testament’ into Gg confirms their distinctiveness to early modern readers, while the remaining eighteen transcribed lines (ll. 592–609) recount Troilus’s actions in the aftermath of her death and supply narrative closure. Holland’s scribe did not, however, copy the poem’s moralising final stanza (ll. 610–16) with its direct address to ‘worthie women’ who can learn from Cresseid’s terrible example.

While these acts of transcription in Gg and L1 may have been informed by a misguided idea about the authorship of the Testament, they also testify to a close readerly affinity to the story and in particular to the character of Criseyde/Cresseid, whose departure from Troy also sees her recede from the narrative of Chaucer’s poem. Neither manuscript was conceived to include the Testament but by the seventeenth century at least two collectors of old copies of Chaucer expanded their books to accommodate an account of Cresseid’s fate. On a broad cultural level, the work of these readers was undoubtedly influenced by the pervasiveness of the Troy story; in a much more direct sense, it was facilitated and encouraged by the juxtaposition of the texts in contemporary printed editions. At least until the lost edition of 1585 and possibly as late as 1593, the majority of sixteenth-century readers learned of the story’s outcome from printed copies bearing not the name of the Scottish Henryson but that of England’s national poet.

The ambiguous convergence of Troilus and Criseyde and the Testament of Cresseid in sixteenth- and seventeenth-century printed editions had far-reaching effects on the story’s reception. Authorised in editions of the Workes, the figuration of Cresseid by Henryson as a wanton, beggar, and leper filtered into the general consciousness and thereby into the Elizabethan literary tradition.⁸³ The proliferation of Cresseid-figures in early modern retellings is demonstrably indebted to the Middle Scots response. Unmistakeable allusions to Henryson’s Cresseid appear in popular poetry collections including George Turberville’s Epitaphes (1567), George Whetstone’s Rocke of Regard (1576), and George Gascoigne’s Posies (1575). Further echoes are found in dramatic works that include not only Shakespeare’s Troilus and Cressida (c. 1602), but also his Henry V (1599) and Twelfth Night (c. 1601), as well as Thomas Dekker and Henry Chettle’s lost play ‘called Troyeles and creasse daye’ (c. 1599), and Thomas Heywood’s 2 The Iron Age (c. 1613).⁸⁴ Around 1585, the courtier Gabriel Harvey included the Testament of Cresseid with its ‘winterlie springe’ in a list which praised Chaucer’s ‘description[s] of the Spring’.⁸⁵ Sometime between the mid-sixteenth and early seventeenth century, an anonymous author translated Chaucer’s Troilus and Henryson’s Testament into the Welsh-language Troelus a Chressyd, a dramatised fusion of the two works which relied on a contemporary printed edition as its source text.⁸⁶

The intellectual lineage of these early modern Troy stories may be traced to printed copies of Chaucer. By positioning the Testament as a product of Chaucer’s pen, Thynne and his editorial successors reoriented the narrative away from Troilus’s meditation on the transience of the universe in Book V and towards Henryson’s wrenching portrayal of human suffering embodied in his Cresseid.⁸⁷ Readers of Thynne and the subsequent folio editions then created adaptations informed by the misattribution, and charted a different reception for Criseyde in the process. Books like Gg and L1 show this new literary history in the making. In supplementing each manuscript with text taken from Henryson’s sequel, the readers of these old books reveal a literary taste tolerant to adaptation and even contradiction. In bringing the narrative itself to a more satisfying conclusion, they also express a new cultural interest in an imagined textual entity called Troilus and Criseyde which accommodates the fate of Criseyde/Cresseid as well as that of Troilus.

3.3 Chaucer’s Plowmen

The long history of supplementing Chaucer’s books with material from the Middle English Plowman tradition offers further evidence for the central role of print in propagating key textual traditions into the early modern period. In Oxford, Christ Church, MS 152, a late fifteenth-century manuscript containing the Canterbury Tales as well as Lydgate’s Churl and the Bird and Siege of Thebes, there is a spurious tale embedded amidst the Chaucerian material. After the abrupt conclusion of the Squire’s Tale in the manuscript, the rest of the quire was left blank by the first scribe to await finishing. A second copyist working around the same time filled these blank pages with a new text.⁸⁸ The supplied poem is a version of Thomas Hoccleve’s Miracle of the Virgin, also known as The Monk and our Blessed Lady’s Sleeves.⁸⁹ But as part of this textual interpolation, the new work is here attributed to Chaucer, not Hoccleve, and is introduced in terms similar to those of any other Canterbury tale. Thus, the poem begins with ‘The prologe of the Ploughman’ and the tale proper is introduced with the incipit ‘Here begynnyth the ploughmannys tale of owr lady’.⁹⁰ As its modern titles imply, the tale is an orthodox miracle of the Virgin; in the Christ Church manuscript, it has been refashioned to fit the mold of Chaucer’s storytelling game.⁹¹ ‘The prologe of the Ploughman’ consists of a spurious two-stanza link, unique in this copy, in which the Host invites the Ploughman to tell the next tale and the Ploughman vows to tell ‘A tale of Crystys modyr dere’.⁹² The Miracle of the Virgin, with its exemplum of the devout monk praying a Latin Pater Noster, has been recognised as a tale of ‘unimpeachable orthodoxy’ and even one which is ‘implicitly anti-Lollard’.⁹³ The suspected origins of the manuscript at Winchester College, a place known for its Marian devotional traditions, offers a compelling rationale for the supplementation of the blank leaves with a perfectly conventional tale.⁹⁴ It was fitting for an institution founded in honour of Mary to fill the lacuna thus; it was also advantageous for Chaucer’s fifteenth-century reputation that the theretofore silent Plowman named in his General Prologue should prove to be a Catholic conformist fully distinguishable from the notoriously reformist Piers Plowman.⁹⁵ The supplementation of the blank leaves (initially set aside for the conclusion of the Squire’s Tale) with an interpolated ‘prologe’ and ‘ploughmannys tale’ is in keeping with the diverse sources used in the manuscript’s copying and with the overall effect of tale disorder noticed by Manly and Rickert.⁹⁶ It is typical, too, of the resourcefulness which often characterises early practices of bibliographical completing. The Miracle of the Virgin was an ideologically appropriate text with which to supplement the Christ Church manuscript, but it also provided a bibliographically convenient fix for the second scribe who copied it. The same hand is responsible for corrections, filled-in lines, instructions, and signes de renvoi to help the book’s reader make sense of leaves disordered during production.⁹⁷ That scribe was responsible, in other words, for completing and perfecting the manuscript, and the addition of the Miracle of the Virgin, in addition to being doctrinally suitable for the Winchester manuscript, also solved the immediate bibliographical problem of an unsightly gap at the end of the Squire’s Tale.

In the following century, a very different sort of Plowman tradition – one with an anti-Catholic strain – would come to be tacked on to Chaucer’s works by stationers, editors, and readers alike. The title page of the anti-clerical prose satire Jack Upland (c. 1536) would claim that it was ‘compyled by the famous Geoffrey Chaucer’, and John Foxe enthusiastically endorsed this attribution when he reprinted the text and promoted Chaucer as a Wycliffite reformer in his 1570 Actes and Monuments.⁹⁸ The most pervasive textual incarnation of the idea of Chaucer’s proto-Protestantism, however, was the anonymous Plowman’s Tale, a fifteenth-century allegorical debate between a Pelican and a greedy Gryphon (or Griffin) who represent Christ and the Catholic Church respectively.⁹⁹ During the sixteenth century, this anonymous work would become hitched to Chaucer’s name and eventually to his works. No early manuscript survives, but the text was first printed in the 1530s by Thomas Godfray, who also printed the first collected edition of Chaucer’s Workes around the same time.¹⁰⁰ From 1542, the association of the Plowman’s Tale with the Canterbury Tales would be fortified by its inclusion in printed copies of the Workes. Here, the sixteenth-century Prologue introduces the Plowman as a participant and tale-teller on the ‘pylgremage’ who is enjoined by the Host to ‘tell us some holy thynge’.¹⁰¹ What follows is a rather one-sided debate dominated by the invective of the Pelican. Originally attached to the end of the Canterbury Tales (following the Parson’s Tale) in 1542, the text was transferred to a more secure position between the Manciple’s Tale and Parson’s Tale from c. 1550.¹⁰²

As with Jack Upland, the attribution of the Plowman’s Tale to Chaucer was further promulgated by Foxe’s extremely popular martyrology. In 1647 the Gloucestershire preacher John Trapp could remark that, ‘M. Fox tels us, that by the reading of Chaucers books, some were brought to the knowledge of the truth’.¹⁰³ Speght avowed that the Plowman’s Tale was ‘made no doubt by Chaucer with the rest of the Tales’, while his fellow antiquary Francis Thynne recounted that Cardinal Wolsey had suppressed an anti-clerical text called the ‘pilgrymes tale’ from his father’s edition while the Plowman’s Tale was ‘with muche ado permitted to passe with the reste’.¹⁰⁴ In 1606, it was published in quarto format by Samuel Macham and Matthew Cooke, with an attribution to an ennobled Chaucer on the title page: ‘Written by Sir Geffrey Chaucer, Knight, amongst his Canterburie tales: and now set out apart from the rest’.¹⁰⁵ This edition of the Plowman’s Tale, dense with explanatory marginal glosses, begins with ‘A description of the Plowman’ from the General Prologue, then presents a note on the place of the tale within the Canterbury collection:

In the former editions of Chawcer. This Tale is made the last, but in the latter, set out by M. Spights aduise, and commendable paine, it is the last sauing the Parsons Tale, I doubt not but this change is warranted by some olde coppies written

For the majority of early modern readers, then, the Plowman’s Tale was a genuine Chaucerian work – one whose authenticity was affirmed in authoritative printed books like those of Foxe and Speght. Problematically, however, the Plowman’s Tale was conspicuously absent from the manuscript record. If it were a genuine Chaucerian work, as commentators like Foxe insisted, and as the editions also attested, then its authenticity should be verified by ‘olde coppies written’ as the 1606 edition assumed – that is, by evidence of circulation with the rest of the Tales or attribution to Chaucer in early manuscript copies. Speght’s reassurance to his readers of the tale’s genuineness stops short of such a claim; as it turns out, his belief that it was authentically Chaucer’s was based on rather shaky proof: ‘For I haue seene it in written hand in Iohn Stowes Librarie in a booke of such antiquitie, as seemeth to haue been written neare to Chaucers time’.¹⁰⁶ The nature of this antique ‘booke’, whether a collection of Chaucer or not, goes unspecified, and the facts about what Speght has ‘seene’ – an evidently old book which merely ‘seemeth’ to date from Chaucer’s lifetime – are put into service of the bolder conclusion that the tale was ‘made no doubt by Chaucer’ and belongs ‘with the rest of the Tales’. In 1570, Foxe had served up an ingenious reason to explain the paucity of manuscript evidence for the authenticity of the Plowman’s Tale: the tale had, naturally, been suppressed. Given its ‘playnely tolde’ exposure of the Catholic Church, he argued, it is ‘therfore no great maruell, if that narration was exempted out of the copies of Chaucers workes: which notwithstandyng now is restored agayne, and is extant, for euery man to read that is disposed’.¹⁰⁷ The absence of the Plowman’s Tale from the early manuscript copies of the Canterbury Tales could therefore be accounted for by its heterodoxy. This sixteenth-century myth that Chaucer wrote the tale (only for it to be suppressed) persisted for centuries, aided by the continued commercial success of Foxe’s and Speght’s books.

So pervasive was this story, in fact, that the absence of the tale from manuscript and printed collections of Chaucer was starkly obvious to some readers. A seventeenth-century hand that might belong to the antiquary John Barkham observed in a Canterbury Tales manuscript that the collection was missing only the Plowman’s Tale and that ‘if it were Chaucers, it was left out of his Canterbury Tales, for the tartnes against the Popish clergie’.¹⁰⁸ A copy of the 1532 Thynne edition held in Glasgow likewise contains marginalia which record a reader wondering about the missing Plowman’s Tale in the seventeenth century (see Figure 3.5). Amidst the printed table of contents, the annotator observed that ‘The Tale of the Ploughman / The Pelican & Griffin omitted’ and has squeezed in a precise cross-reference to John Foxe’s ‘Actes & monumentes fol. 56. colum. 1 – volume last edit… Printed anno 1641: Chaucer comended’.¹⁰⁹ In a 1709 commentary on Chaucer, the antiquary Thomas Hearne remarked similarly: ‘Now the Plough-man’s Tale having given more offence than all the rest of Chaucer’s Works, perhaps that is the reason why it appears in so few MSS’.¹¹⁰ A few years later, in 1715, the young Thomas Martin bound his copy of the 1606 Plowman’s Tale with a 1591 Spenser quarto and reinscribed what he believed to be the authorship of each text on his improvised title page: ‘Chaucers Plow-mans Tale: with some of Spencers Works’.¹¹¹

Figure 3.5 An early modern reader’s note on the omission of the Plowman’s Tale and a reference to Foxe’s Actes and Monuments in a copy of Thynne’s 1532 edition.

University of Glasgow Archives and Special Collections, Bs.2.17 (STC 5068; sig. A3^v).

In one copy of Thynne’s 1532 Workes, now held at the Harry Ransom Center in Austin, Texas, the Plowman’s Tale also constituted a conspicuous absence. In this copy, the text has been transcribed by a sixteenth-century hand onto sixteen paper leaves and bound into the book at the end, following the Parson’s Tale.¹¹² There is some scholarly disagreement about whether this manuscript copy was transcribed from a lost printed exemplar, a lost manuscript, or from one of several editions printed in the second quarter of the sixteenth century.¹¹³ While its textual origins are uncertain, the text copied into the Texas Thynne remains significant because it represents a codicological affirmation of a widespread belief that would be ‘decisively rejected’ only centuries later: that the polemical Plowman’s Tale was written by Chaucer.¹¹⁴ Its retroactive addition to the book is an act of textual supplementation, but it is also an attempt to correct a historical record which would suppress heterodox works. One modern commentator has expressed ‘bewilderment that so unsophisticated a piece could ever have been seriously thought of as Chaucer’s’.¹¹⁵ But readers in the sixteenth century, who believed Chaucer was sympathetic to the reformist beliefs of the Lollard John Wyclif, had good reasons for finding the national poet’s voice in the Plowman’s Tale. Although the text circulated independently and in smaller-format editions, early modern commentators, as we have seen, judged its legitimacy in part on bibliographical grounds. Its proximity to ‘the rest of the Tales’ in printed books was (they assumed) a token of Chaucer’s authorship, of genuineness, and of some rare but incontrovertible manuscript evidence which had been semi-successfully suppressed. Meanwhile, books in which the Plowman’s Tale was absent furnished proof of its radically righteous message and a spur to see it reinstated, as in the Texas copy. However incongruous its attribution to Chaucer may appear with the benefit of hindsight, the Plowman’s Tale was an appropriate text with which to supplement a copy of Thynne’s 1532 Workes in the sixteenth century, for that edition was the last to exclude the Plowman’s Tale for nearly two and a half centuries, until Tyrwhitt oversaw its removal from the print canon in 1775.

Writing of this tale, Brendan O’Connell has observed that there can be ‘far-reaching implications when an apocryphal work is incorporated into the canon, and in particular when a new tale is incorporated into a framed story collection such as the Canterbury Tales’.¹¹⁶ Some of these implications, as he argues, were thematic; others took the form of the material and bibliographical interventions into copies of Chaucer’s works that I have been describing, as medieval and early modern books alike were evaluated and perfected to reflect the new realities of the accepted canon. The variously supplied explanations for its absence posit that the Plowman’s Tale could belong ‘with the rest of the Tales’ (Speght), that it was ‘exempted out’ of the manuscripts (Foxe), that it may have been deliberately ‘left out’ of the collection (Barkham), and that its offensiveness caused it to be removed from ‘all the rest’ of the works (Hearne). Foxe went further to add that this omission ‘now is restored agayne’ in the printed copies. To be ‘brought to the knowledge of the truth’, as the preacher Trapp hoped they would be, readers needed access to the entire, unredacted book. Such remarks reveal that what was at stake in the early modern establishment of Chaucer’s canon was not only a matter of philological and scholarly investigation, but of religious truth itself. In the context of this widespread cultural narrative about the censorship and later reintroduction of the Plowman’s Tale, it is easy to understand why an early modern reader might supplement their copy of the Canterbury Tales with the tale of the Pelican and the Gryphon, or remark upon its absence. Moreover, the fact that other Plowman literature was similarly crowbarred into surviving copies of Chaucer’s works shows the cultural persistence of the narrative promoted by Foxe and the Chaucer editors in print.

The Trinity College manuscript with the inserted lyrics in its initial quire also contains another sixteenth-century supplement in the same hand. At the end of the original codex, after the conclusion of the Parson’s Tale and the Retraction, the later hand has supplied the anonymous alliterative poem Pierce the Ploughman’s Creed on twelve paper leaves.¹¹⁷ The poem, an anti-fraternal satire influenced by Langland’s Piers Plowman, dates from the late fourteenth century, although the earliest surviving complete copy is an edition published in 1553.¹¹⁸ Two early modern manuscript copies of the poem, of which the Trinity text is one, also survive. Unlike the short poems copied from Thynne’s edition at the beginning of this manuscript, this version of Pierce the Ploughman’s Creed appears to have been copied from an authoritative manuscript which has since been lost. This reader, who Skeat characterises as ‘a scrupulous and painstaking antiquary, who carefully put down what he saw before him’, was influenced by Chaucer’s contemporary reputation for Wycliffite views and by the attribution to him of the Plowman’s Tale.¹¹⁹ A title for the work was written above the top line on the first page, likely in the same hand, but it is now illegible due to cropping.¹²⁰ It is therefore not certain that the copyist thought Pierce the Ploughman’s Creed to be written by Chaucer, but its inclusion in a copy of the Canterbury Tales shows that they thought it an appropriate fit.

While the sixteenth-century compiler of the new additions remains unidentified, the book’s place in Archbishop Matthew Parker’s orbit via his son John points to the reformist motivations which may lie behind this choice of supplement. The Parkerian red crayon used to paginate the manuscript, including the leaves of the belatedly added Pierce the Ploughman’s Creed, suggests that they were in place – and perhaps even appended – during the book’s time in the Parker milieu. As is well established, the antiquarian pursuits of the Archbishop and his circle were animated by a desire to assert ancient precedent for English Protestantism. According to John Foxe, Parker wished to prove that the new religion ‘is no new reformation of thinges lately begonne, which were not before, but rather a reduction of the Church to the Pristine state of olde conformitie, which once it had’.¹²¹ Forni, Gillespie, Cook, and others have described the processes by which the works of Chaucer, a paragon of English learning and literary authority, provided a convenient vessel for conveying the antiquity of the new religious way. Apocryphal texts and tales were key to that mission and the printed canon of Chaucer was duly made to accommodate such works.¹²² The presence in the Trinity manuscript of Pierce the Ploughman’s Creed – a text deemed ‘the most obviously Lollard member of the Piers tradition’ – appears a comparably politic addition.¹²³ The supplement demonstrates the gravitational pull of Chaucer’s name and authority, as well as the perceived extent and capaciousness of his literary works. This text, unlike the Plowman’s Tale and Jack Upland, was not included in any printed collection of Chaucer’s Workes, nor was it attributed to him in any extant edition. But while there is no explicit link to a printed source behind this particular supplement, it remains impossible to discount the background involvement of the influential printed tradition in establishing and extending Chaucer’s reputation as a writer of stories about Plowmen and setting the readerly expectation that his books should contain such tales.¹²⁴

In the Trinity copy, the added Pierce the Ploughman’s Creed begins on what is now fol. 317^r, on the facing page to the conclusion of the Parson’s Tale and the Retraction on fol. 316^v. The latter pair of works has an entangled relation to the Plowman tradition associated with Chaucer, and modern critics have pointed out that the placement of each in early editions generates a series of interpretative ambiguities. The position of the Plowman’s Tale at the end of the Tales in 1542 might be a marker of that pilgrim’s moral triumph over the Parson, parts of whose own contribution are labelled a ‘Canterbury tale’, or trifle, by printed marginal glosses.¹²⁵ But that statement of approval for the heterodox is somewhat undone by the new position of the Plowman’s Tale before the Parson’s in the c. 1550 and subsequent early modern editions, where it is the more orthodox pilgrim who has the final word. Similarly, the deployment of the word ‘fable’ by each of the two tellers leaves the question of their moral authority open to readerly interpretation, as Ensley has observed.¹²⁶ Several of these problematics concerning the tale order may be extended to the text of Pierce the Ploughman’s Creed in the Trinity manuscript. Appearing directly after the conclusion of the Parson’s Tale and the Retraction, the anti-fraternal supplement works as a counterweight to the Catholic piety of the Parson’s prose treatise. Yet the added satire spoken by Piers the Ploughman only rebalances, rather than replaces, the more orthodox texts which claim the authority of Chaucer himself.¹²⁷ In the Trinity copy, the Parson’s Tale is unmarked by any dissenting marginalia on the part of its later owner. This tolerant approach to Middle English devotional texts and Catholic doctrine is reminiscent of the ‘dispassionate objectivity’ exhibited by Parker’s associate Stephan Batman in his annotations on Piers Plowman and other medieval religious texts.¹²⁸ The architect of the satirical supplement did not reject the Parson’s Tale or Retraction, but opted instead to reframe and complete the tale collection with a genre of text already associated with Chaucer. It is even possible that the annotator made the supplement on the basis of a mix-up between Pierce the Ploughman’s Creed and the assumedly Chaucerian Plowman’s Tale, having no copy of the latter to hand.¹²⁹

In the snippets of commentary that survive from early commentators and readers, there are nonetheless glimmers of hesitation about the genuineness of the Plowman’s Tale and other anti-fraternal plowman works as Chaucer’s: in Speght’s too-emphatic phrase ‘no doubt’, in the ‘ifs’ of Foxe and Barkham, and in the ‘perhaps’ that qualifies Hearne’s statement about its offensiveness.¹³⁰ In the absence of physical evidence from the manuscripts, the authenticity of this cluster of texts became a matter to be untangled. A published record of another copy (now lost) of The Vision of Pierce Plowman and Pierce the Ploughman’s Creed (1561) supplies further evidence of the confusion these texts wrought.¹³¹ The reader of this copy appears to have used their acquired knowledge of Chaucer’s canon to make some deductions about the authorship of Piers Plowman and Pierce the Ploughman’s Creed. In a note dated 1577, they observed that ‘Mention is made of Peerce Plowghman’s Creede, in Chawcers tale off the Plowman’, and ‘I deeme Chawcer to be the author [of the Creed]’. On the other hand, the note continues, ‘I thinke hit not to be on and the same yt made both [the Creed and Piers Plowman]’. The reader’s conclusions are drawn from a range of first- and second-hand evidence: the note cites John Bale’s assertion that the Piers Plowman poet was a Wycliffite named ‘Robert’ Langland; the Latinate language of that poem; perceived inconsistencies in the chronology of the two texts; and the apparently unshakeable belief about the authorship of ‘Chawcers tale off the Plowman’. It has been suggested by Simon Horobin that these annotations ‘fit closely with Batman’s recorded interest’ in Piers Plowman; however, the loss of the copy precludes any palaeographical confirmation that they are his.¹³² There is nonetheless some insight to be derived from this inscription, independent of the annotator’s identity. What interests me most about this unknown reader is their interest in authorship and their reasoned triangulation of the three texts according to knowledge they had gathered from printed sources or observed from the texts themselves. For this learned reader, the question of what Chaucer wrote was far from settled. Thus the deluge of anti-clerical Plowman literature in circulation during the sixteenth and seventeenth centuries gave rise to some understandable confusion amongst Chaucer’s early modern readers. Quibbles and commentary on the authenticity of the Plowman’s Tale, voiced by authorities like Foxe and Speght as well as by diligent and curious readers, show a series of arrestingly similar attempts to circumscribe the limits of Chaucer’s canon.

The evidence from extant copies and surviving marginal notes – in which early readers supplemented Chaucer’s print and manuscript works with anti-clerical material or simply remarked on the absence or authorship of the Plowman’s Tale or topically related works – constitutes a material testament to the pervasiveness of the proto-Protestant Chaucer established in print. Such evidence illustrates, too, the widespread belief that the Canterbury Tales was an incomplete, fragmentary, or censored work, and the willingness of readers to ‘restore agayne’ in their own books the parts of the canon that they thought wanting. In 1570, Foxe had characterised Chaucer as gifted with a special foresight – he ‘saw in Religion as much almost, as euen we do now, and vttereth in his workes no lesse’. His anti-Catholic views had successfully evaded censorship under Henry V111’s Act for the Advancement of True Religion (1543) and now, proclaims Foxe, ‘Chaucers woorkes bee all printed in one volume, and therefore knowen to all men’.¹³³ The Plowman’s Tale and the Testament of Love – works which Foxe believed confirmed Chaucer’s Wycliffite views – already comprised part of that ‘one volume’ in the early sixteenth century, and Jack Upland would follow in 1602. But the same was not true of the fifteenth-century manuscripts of the Canterbury Tales, which, of course, were produced under different circumstances and which (for the most part) did not purport to contain these texts. Early modern Protestant readers thus sought to bridge a seemingly censored and partial textual tradition with the clear-eyed religious beliefs of their own day – to augment the older books with what was suppressed in the past and what is ‘knowen’ to Foxe and his contemporaries ‘now’. The rehabilitation of the poet for the Reformers’ cause thus finds its material correlative in the supplementation of his books with tales of Chaucer’s Plowman.

The means by which the Plowman’s Tale and other anti-clerical literature became tethered to Chaucer’s oeuvre is remarkable in itself, for it demonstrates the authority and persistence of narratives about the poet and his canon that circulated in popular printed volumes, and the agency of readers in adapting other books to reflect, interrogate, or at least engage with those narratives. More extraordinary still is the fact that this anti-clerical polemic is only one of two very different types of text which masqueraded in the guise of a Plowman poem written by Chaucer. The codicological complexity of the Canterbury Tales, which in the fifteenth century saw a Marian devotion enlisted to serve as an orthodox ‘ploughmannys tale’ in the blank space in one copy, also left room for a radically different interpretation of Chaucer’s Plowman and his ideological stance in the sixteenth century. Forni has suggested that his putative authorship of the acerbic Plowman’s Tale enhanced Chaucer’s reputation amongst his readers in Protestant England.¹³⁴ Books like the Glasgow copy of Thynne and Barkham’s manuscript support that suggestion, showing that seventeenth-century readers wondered about the absence of the Plowman’s Tale in copies of Chaucer. The Texas Thynne and Trinity manuscript, meanwhile, affirm the desirability of Plowman literature within a conception of the Chaucer canon shaped by contemporary printed books, and the willingness of readers to supplement older copies of Chaucer in that image. In these ways, the surviving medieval manuscript books render vivid the early modern remaking of Chaucer in line with persistent narratives about him which circulated in print.

3.4 Locating Chaucer’s Retraction

Religious ground was also at stake in the reception history of another frequently supplemented text, Chaucer’s Retraction, in which the author seeks divine mercy for having written sinful works, revoking these and expressing gratitude for his moral and devout writings. While the Plowman’s Tale is a spurious work added to the canon as part of the attempt to ‘Lollardize’ Chaucer in the early modern period, the Retraction seems to be genuine, but was probably excised on account of its orthodox piety. This, at least, is the scholarly consensus today – but not so in the late medieval and early modern periods.¹³⁵ The examination of evidence from the fifteenth century reveals that the Retraction appears in a ‘slight majority’ of the surviving complete manuscripts and was included by Caxton in his first and second editions of the Canterbury Tales (though not by all of the early printers who succeeded him).¹³⁶ William Thynne’s exclusion of the text from his 1532 edition proved to be decisive, for the Retraction would not again be printed with Chaucer’s works until Urry’s edition. Some therefore believed that the Retraction was a spurious addition to the canon. In the sixteenth and seventeenth centuries, it was the Plowman’s polemic that was able to secure a place in the print canon, and the Retraction that was overlooked. While Thynne’s refusal to print Chaucer’s final revocation of worldly vanity in 1532 may have stemmed from causes that were more accidental than ideological,¹³⁷ its occlusion in this first collected edition would come to be extremely convenient for the version of the poet put forth in later editions and in Foxe’s Actes and Monuments. For the most zealous followers of the new religion, the penitent Chaucer who ultimately revoked some of his most celebrated works but not his ‘bookes of legendes of seintes, and omelies, and moralitee, and devocioun’ was scarcely compatible with the author of Jack Upland and the Plowman’s Tale.¹³⁸ The author of the Retraction was a poor fit with the proto-Protestant Chaucer who Speght made famous for having beaten a Franciscan friar in Fleet Street. The antiquary Thomas Hearne preserved the sentiment about this mutual opposition between different parts of the canon in a diary entry written in 1709:

I believe the Revocation annex’d to the Parson’s Tale in some Copies of Chaucer not to be genuine, but made by the Monks, who were strangely exasperated for the Freedom he took, especially in the Plow-Man’s Tale of exposing their Pride, Loosness and Debauchery.¹³⁹

The idea that the Retraction was a late interpolation into the canon and introduced at the expense of the Plowman’s Tale was popularised in the 1721 Works, where it was included even though the lately deceased editor Urry, following the opinion of his friend Hearne, had doubted its genuineness, guessing that ‘the Scriveners were prohibited transcribing [the Plowman’s Tale] and injoyn’d to subscribe an Instrument at the end of the Canterbury Tales, call’d his Retraction’.¹⁴⁰

Tracing the revival of the Retraction in eighteenth-century print, Dane has suggested that ‘[t]he Chaucer canon is a question of what belongs in a printed edition of Chaucer’.¹⁴¹ The varying answers to that question provided by individual editors over the centuries resulted in a canon that was frequently subject to change. Chaucer’s Retraction had an irregular presence in the pre-1532 editions, and was then out of print until 1721. Having been produced by a committee of editors following Urry’s death in 1715, even the 1721 edition had an ambivalent approach to the text’s authenticity.¹⁴² Urry wanted to exclude it, while his collaborators William and Timothy Thomas probably believed it to be a genuine work of Chaucer. Dane suggests that the black letter type in which the Retraction is printed in 1721, and which Urry had intended to mark textual authenticity in the edition, would actually become ‘the mark of something spurious’.¹⁴³ As with the Plowman’s Tale, the result of this long history of variability in print was doubt and confusion on the part of readers about the place of the Retraction, not only within Chaucer’s printed books, but in the early manuscripts and in the canon itself.

Three surviving volumes show readers across different centuries reckoning with the presence or absence of the Retraction in early copies of the Canterbury Tales. The first is an incunable, a copy of Richard Pynson’s 1492 edition held at the John Rylands Library.¹⁴⁴ Pynson based the text of the Tales on Caxton’s 1483 edition, but the 1492 volume is distinct for being the only early edition to exclude the Retraction. In passing over the Retraction, Pynson deviated from the standard set by his Caxton copy text, and would evidently have a change of heart by 1526, when he restored it in his second edition of the Tales. The Rylands copy of Pynson’s first edition bears the marks of this patchy publication history. On the recto of the original blank leaf following the Parson’s Tale, a fifteenth-century hand has copied out the Retraction in its entirety.¹⁴⁵ This amounts to twenty-eight lines, which are written in a neat and heavily abbreviated script. The Retraction was added to this book by (or for) Robert Saham, a chaplain from Bury St Edmunds who also personalised the ending of Chaucer’s prayer by adapting it into his own voice: ‘Amen quod Saham’ follows the text apparently copied from Caxton’s second edition.¹⁴⁶ With its solemn notes of penitence and prayer, the Retraction serves as an appropriate supplement to a cleric’s copy of Chaucer, a readerly move that invites a quick dismissal of the ‘worldly vanytees’ contained in the preceding collection in favour of its ‘other bokes as of legendys of seyntes and omelyes moralite and devocion’. This completion of Saham’s book with the Retraction, that is, achieves a necessary spiritual ending for his copy of the Canterbury Tales. At the same time, aesthetic considerations are also present in this programme of supplementation, for the book has been embellished with rubricated initials in several places for which space had been left in the printing.¹⁴⁷ The last of these rubricated letters, a capital letter N, appears not on a printed page but in the first word of the transcribed Retraction, suggesting that the book was rubricated in the same period that the text was added.¹⁴⁸

About a century after Saham’s copy of Pynson was supplemented with the Retraction, Joseph Holland also had the same text added to his manuscript copy of Chaucer’s collected works.¹⁴⁹ In the interim, the Reformation had indelibly altered the devotional context in which Chaucer might be read, and although his reputation for anti-clerical commentary kept him in favour, the religious orthodoxy which had made the Retraction so desirable, perhaps even necessary, in the chaplain Saham’s book at the beginning of the century had become uncomfortable by its close. The transcription in Holland’s Gg manuscript reflects this anxiety, and shows that the text of the Retraction in Caxton’s second edition was tellingly adapted for its new context in what was now a post-Reformation manuscript. Thus while the text in Caxton (and in Saham’s book) sees the speaker praising Christ, Mary, and the saints – ‘our lord Jhesu Crist and hys blessyd moder and alle the sayntes of heuen’ – only Christ is retained in the version copied into Gg.¹⁵⁰ This was no accidental omission by Gg’s early modern scribe, for the third-person plural pronouns which in Caxton refer to Christ and the holy intercessors – ‘hem’ (Middle English ‘them’) and ‘they’ – also become masculine singular in Holland’s copy – ‘hem’ (now serving as ‘him’) and ‘he’. As Wolfe has pointed out, the privileged place of the Retraction at the end of Chaucer’s collected works invites a reading of that text as ‘a general work’ which comments seriously on the poet’s literary legacy; it ‘may well be the one place we hear the “real” Geoffrey Chaucer speaking to us’.¹⁵¹ For Holland, the Retraction was significant enough to be included in his improvement of Gg, yet even that locus of Chaucerian literary authority could be modified and adapted for new religious ends which could ultimately ‘outweigh the textual authority of an exemplar’ found in Caxton’s print.¹⁵² The versions of the Retraction copied from Caxton and inserted into other copies of Chaucer by Saham and Holland demonstrate markedly different forms of readerly adaptation and supplementation which rely on print for their models if not for their ultimate forms. Both books show that their owners actively sought out what had become a rare text for the purpose of completing their own copies.

The return to print of the Retraction in Urry’s edition is marked by another instance of supplementation in the same approximate period. A copy of the Canterbury Tales, BL, MS Egerton 2726 has had a paper leaf containing the end of the truncated Retraction appended where the original vellum leaf has been lost (fol. 271). It has been suggested that the hand may be that of the eighteenth-century antiquary William Thomas.¹⁵³ As was observed, William and his brother Timothy were contributors to the 1721 edition and they likely believed, unlike Urry himself, that the Retraction was a genuine Chaucer text. Dane has demonstrated that the Thomas brothers had only partial access to and an imperfect understanding of the textual tradition of Chaucer’s works in manuscript as well as print. The resulting gaps in their understanding of these historical books allowed them ‘to imagine their contents however they wished’.¹⁵⁴ If the hand belongs to William, the supplied conclusion to the Retraction provides additional evidence for his belief in the text’s authenticity as a Chaucerian piece, and for the codicological interventions that accompanied the brothers’ reconstituting of the poet’s corpus.¹⁵⁵

The furnishing of the Retraction in Egerton, especially if it is the work of William Thomas, should be considered, too, in the context of another supplement made to the same manuscript. This latter takes the form of eight parchment leaves inserted in the eighteenth century, this time by Timothy Thomas.¹⁵⁶ The leaves contain the spurious Tale of Gamelyn and have been inserted in the midst of the Cook’s Tale of Perkyn Revelour.¹⁵⁷ According to notes made by Timothy, he copied the text from the manuscript now called Bodl. MS Laud Misc. 600 (then MS Laud K.50) while Egerton was on loan from the Earl of Carnarvon during the preparation of the 1721 edition.¹⁵⁸ A headnote to Gamelyn in the edition indicates that Urry himself had considered the tale to be genuine and wondered why no previous editor had previously printed it, given its presence in many manuscripts.¹⁵⁹ He reasoned that perhaps they had never encountered these manuscripts, or perhaps they simply doubted its genuineness. He was ultimately in favour of its inclusion: ‘But because I find it in so many MSS, I have no doubt of it, and therefore make it publick, and call it the Fifth Tale’.¹⁶⁰ Timothy’s supplementation of the Egerton manuscript with the Tale of Gamelyn therefore coincides with a second wave of its inclusion into the definitive Chaucer print canon. The apocryphal Gamelyn achieved such canonicity in 1721, but it would only maintain this status until the publication of Tyrwhitt’s edition, which excised it once again, and for good.¹⁶¹ Given that the editions that came before and after 1721 ruled against the inclusion of Gamelyn, the legacy of that book on the tale’s canonicity might appear slight. This should not detract from the considerable periods in which Gamelyn was considered canonical, for about fifty years following 1721, and for much of the fifteenth century, when it was accepted as a second tale for Chaucer’s Cook in at least twenty-five manuscripts that survive today.¹⁶²

Was the addition of Gamelyn to Egerton an act of improvement carried out at the request of the book’s owner while it was on loan? Whatever its immediate motivations, Timothy’s perfecting of the manuscript with a missing text which he believed to be canonical suggests an attempt to put right the manuscript record itself – to bring it into line with the textual state of ‘so many MSS’ that he had examined. The Thomas brothers emerge as the final actors in this history of supplementation and it is fitting that they appear to have been engaged with improving these manuscript books at the same time that they were editing Chaucer. Their retrofitting of the Egerton manuscript is a material reshaping of the medieval book to align with the editorial practice of their own historical moment and with a Chaucerian canon whose limits were accustomed to being redefined in print.

3.5 Chaucerian Compilations

The material supplements traced in this chapter enact, in codicological form, responses to the canon which are often overlooked in favour of more straightforward textual or literary evidence of Chaucer’s reception, such as that detectable in the work of early modern authors who used the poet’s work to inform their own creations. But that evidence – much of it indebted to the influence of successive generations of editors – is matched by a rich record of readerly engagement with the same ideas and their offshoots: that Chaucer was a poet of fin amour, that he condemned Criseyde to a wretched death, that he assigned his Plowman an anti-clerical tale, and that the Retraction was a later monkish forgery.¹⁶³ Both medieval readers of manuscripts and their early modern counterparts exploited the book’s seemingly limitless capacity to be annotated, supplemented, and expanded to particular ends. Sometimes, a codex might be updated to include new texts that are apparently unrelated to its prior contents, as in the case of a manuscript of the Canterbury Tales copied in the second half of the fifteenth century by John Brode and now at the Rylands Library in Manchester.¹⁶⁴ Brode went on to add further short texts, such as verses on the death of Edward IV, a religious lament, and articles on Christ’s passion, to blank leaves in the beginning of the book.¹⁶⁵ In her recent study of medieval manuscripts in the longue durée, Elaine Treharne writes that the addition of drawings or texts to blank space is ‘not at all rare’, and invokes the ‘tens of thousands of manuscripts where incomplete, partial, or abbreviated notes, comments, drawings, and literary snippets are written into space’.¹⁶⁶ Despite this proliferation, the practice should not be dismissed as quotidian or commonplace, and Treharne goes on to argue compellingly for the meaning inherent to any intervention in a manuscript.

This book demonstrates that the identification of patterns is one means of recovering the meaning behind historical acts of reading, and the present chapter has located a series of such patterns in the early modern supplements made to Chaucer manuscripts by their readers. The juxtaposition of manuscript and printed Chaucerian books in this chapter has turned up striking parallels in the ways certain texts moved within and across volumes. Cycles of textual attachment and detachment are evident, for instance, in the use of Chaucerian lyrics as fillers in fifteenth-century anthologies, in Thynne’s editions (and subsequent ones), and in manuscripts supplemented by Holland and the Trinity annotator. Ways of reading also repeat themselves in the persistent yet ambiguous pairing of Troilus with Henryson’s Testament in sixteenth-century print and manuscripts, in the repeated assignation of a voice to Chaucer’s Plowman, and in the desire – of fifteenth-century scribes and eighteenth-century editors alike – to furnish a satisfying supplement to the aborted Cook’s Tale. These patterns and echoes in a text’s reception point to certain enduring readings of Chaucer across manuscript and print. From these histories, the book emerges as modular, changeable, and capable of being adapted to the ends desired by its readers or required by historical circumstance. What for Jeffrey Todd Knight is true of the printed editions – that they possessed ‘a flexibility in poetic content that permitted inclusions, annexations, and other forms of textual intervention by publishers’ – is also demonstrably true of the manuscripts.¹⁶⁷ These fifteenth-century volumes and their afterlives embody Knight’s notion of the ‘custom-made corpus’ and demonstrate some of the entanglements and continuities possible between medieval and early modern habits of book use. Such exchanges moved across different media as well as across different temporal horizons. The textual and codicological supplements, alterations, and transformations that are this chapter’s chief focus took place in medieval manuscripts, but they were carried out in light of versions of Chaucer and his canon which circulated in contemporary printed books and were often directly extracted or informed by those volumes.

This enlargement of physical copies to accommodate different texts which early modern readers believed belonged in the Chaucer canon reveals a mode of reading guided by a spirit of renovation.¹⁶⁸ In this context, supplementing old books with desired texts copied from print was a logical part of updating and improving prized volumes. It went hand in hand with the careful handwriting often used in copying and the frequent choice of parchment as a writing support. As with the repairs and continuations previously discussed, such choices signal that these books were cherished for their age and cultural importance. From erasure marks, for instance, it seems that Holland had all of the pages of Gg cleaned to remove existing marginalia that had accumulated over nearly two centuries, pursuing a bibliographical ideal of unadulterated authenticity even as he added post-medieval and non-Chaucerian material to the same book.¹⁶⁹ For these readers, to supplement a book with new leaves and text was fully compatible with the guiding desire to preserve their old copies of Chaucer.

For all their mobility, the Chaucerian texts copied from print were assembled into formations which were meaningful to their compilers and copyists, or out of which meaning could be later constructed. The author, as Gillespie has compellingly argued, was one increasingly prominent site of meaning for the printed collections of Chaucer’s works, but the makers of those books could only suggest, rather than contain, the forms and readings later imposed upon them. Chaucer-the-author was the primary selling point for the Workes but he was also, to some extent, an abstraction – ‘a category grand enough, and convenient enough, to accommodate the writings of other medieval authors’.¹⁷⁰ As this chapter has shown, he could accommodate other identities too: amongst them, a moralist, love poet, Wycliffite, or repentant author writing from his deathbed. It is not always clear from the volumes studied here whether Chaucer-the-author was an organising principle around which texts were grouped, or to what extent his name simply functioned as a magnetic pole that attracted texts which appeared generically, linguistically, or historically compatible. In this too, the printed editions retain some of the ambiguity present in their manuscript antecedents and reflect it back onto the newly supplemented copies. This elastic conception of Chaucerian authorship is in keeping with the material contingency of the pre-modern bibliographical culture that Knight describes – one characterised by ‘insertions and other forms of intrusion and compilation’, and one in which individual texts could move across multiple constellations of meaning.¹⁷¹ When it appeared in 1532, the first collected edition of Chaucer garnered praise from printer Thomas Berthelet, who emphasised its variety and capacity to convey ‘many other’ of Chaucer’s works ‘that neuer were before imprinted, & those that very fewe men knewe, and fewer hadde them’.¹⁷² With each successive edition, greater numbers of readers ‘knewe’ and ‘hadde’ access to a growing body of texts associated with Chaucer’s name, and were able to assign new, and newly relevant, meanings to them. The acts of supplementation considered here show readers appraising lyric poems, the Testament of Cresseid, the diverse tales of different Plowmen, the Retraction, and Gamelyn for their suitability in the books discussed. The conclusions they reached about Chaucer’s authorship of these works are sometimes indeterminate, but their interventions in older copies are precious evidence of early modern literary taste and judgement on the matter of what belonged in a Chaucerian book.