University of Virginia Library

The manuscript usually called "The Findern Miscellany" (Cambridge University Library Ff.i.6) has long been recognized as one of the most important vehicles for fifteenth-century poetry.[1] Often considered an example of the spread of London poetic taste into the hinterlands, it was copied in south Derby, almost certainly in a gentry household during the third quarter of the century. Recently, in a rich and carefully delineated study, Kate Harris has provided the most extensive annotation to date of how the manuscript was produced—an elaborate discussion which lays out in full the evidence of quiring, of the watermarks of the papers used, and of scribal stints.[2]

As a result of her intricate studies, Harris departs from what has in the past been received wisdom as to the major outlines of production of Ff.i.6. Most researchers have believed that the manuscript is essentially fascicular (only Edith Seaton thoroughly demurred), that it was produced as a sequence of separate booklets and only belatedly, probably at binding, gathered into its current form. Although there has been some discrepancy in various scholars' identification of booklets and their boundaries, their general agreement is clear. Harris, on the basis of a considerably more intense look at the evidence than that made by any past codicologist, rejects this view in favor

of the opinion that the manuscript essentially can be seen as reflecting a single act of production, shared among a great many hands.[3] In this article, I want to show that the evidence which Harris adduces instead directly supports the older view of booklet production.

Harris finds two principal ways to question that the volume was produced as a series of separate booklets. She looks for overlap among putative booklets of scribal stints or of materials, the thirteen separate paper-stocks used in producing the codex. Finding such overlaps, she assumes that they could not occur were the manuscript not produced as a single unit, and thus she rejects the fascicular theory.

But inherently, neither of these features, scribal overlap or reuse of paper-stocks is, in itself, telling. Although booklets in manuscripts may well be distinguished by the fact that separate scribes have written different chunks of the work or by the fact that different writing materials occur in separate blocks, the negatives of these two propositions are not necessarily true. That is, the presence of a single scribe in different portions of a manuscript does not, a priori, mean that he is not contributing to separate booklets, nor does the repetition of a paper-stock in different portions mean that all those portions are of a single origin and copied as a planned and unified whole.[4]

Moreover, the data which Harris has so attentively assembled render either of these conclusions rather unlikely. The scribal stints in the volume suggest a variety of copying practices, ranging from the carefully rational to the thoroughly whimsical. The use of paper-stocks appears similarly diverse: some quires contain only a single stock, often unique to the textual unit, whereas others have mixtures of stocks. In this context, all other issues aside, one might wish to adopt a flexible view of the modes of production.

The scribal copying of the manuscript palpably reflects a variety of situations. These Harris recognizes. Some items she believes might be thought of as "main texts," works always planned to occupy their folios in Ff.i.6. These are implicitly singled out by Harris's notation that other texts can be immediately classified as either "filler" or "added items," i.e. works written-in later on originally blank folios. Such a view suggests at least two levels of composition within the volume, one formal and one less so.

But even within formal stints distinctions can be drawn. Here I simply point at two extremes. The romance Sir Degrevant (item 27) was plainly copied from a split exemplar by two scribes (21 and 22, apparently women— Elizabeth Cotton and Elizabeth Francis). One copied into the end of quire G a substantial portion of the text, while the other copied the remainder into the head of quire H. The join in the two stints may not have been very accurately computed, since the last leaf of quire G has been cancelled, without

loss of material: the scribe of G probably did not receive enough copy to fill her quire completely. But in any event, as in several similar examples, the scribal procedure is planned and, generally speaking, carefully executed.[5]

As an immediate contrast, one might consider the two preceding texts, excerpts from Gower's Confessio Amantis (items 25 and 26). These were copied by five scribes total—two worked on both texts, three in company with them only on the second item. The copying proceeded by a succession of alternations which generally do not respect even folio boundaries; in some cases the stints are as short as seven or eight lines and thoroughly unmotivated by any normal detail of production. In this case, and others which approach it,[6] one is forced to assume highly unstructured procedures—either copying as a sort of social game, where the archetype and in-production codex were passed about in a gathering for successive additions; or copying by leaving archetype and in-production codex out (on a table, say) for chance additions by any interested members of the household.

Because of such variations in the manufacture of the codex, it is difficult to view scribal behavior as so carefully controlled as it is in a professionally produced text. The reappearance of scribes may simply reflect the rather off-hand methods of production which the Finderns (or whatever family produced Ff.i.6) followed in creating their miscellany. In any event, the hands which recur in different sections of the manuscript provide no definite information that those sections are actually continuous, much less planned continuous, productions.[7]

The evidence of paper-stocks is similarly rather tenuous. In fact, leaving

aside one ambiguous case (quire A) which I will discuss below, only one stock is repeated in pieces of the volume usually considered separate booklets —stock 4 (which recurs in quires D, E, K, O, and P). But in fact all that this evidence may indicate is that the Finderns at one time had a substantial amount of stock 4 and utilized it repeatedly. Certainly, the thirteen stocks clearly present in the volume imply a paper supply highly mixed.

In this regard, the case of quire K is instructive: this ten-folio unit would, in the quarto format of the codex, have required a half-sheet. The watermarks indicate the quire is comprised of portions of three sheets, two of stock 4, one of stock 9; the preceding quire, and the three which follow, appear totally on stock 9. Here the stock 4 simply appears to be a remnant, perhaps including a half-sheet which otherwise would have been wastage, mixed with a stock in greater supply at the time of copying.

If the mixture of scribal hands and of paper-stocks does not provide convincing evidence of nonfascicularity, one may inquire whether other evidence exists which would suggest in some positive way that production has proceeded by booklets. Here, one can invoke a series of criteria associated with fascicularity and ask whether "The Findern Miscellany" provides any examples of these features. Following arguments I have made elsewhere,[8] the following preliminary questions would seem appropriate:

• 1. Does the manuscript contain any units which seem separate on the basis of bearing self-contained texts?
• 2. Do these units correspond to any independent set of quire signatures?
• 3. Do these units correspond to variations in the number of sheets per quire, particularly quires at the ends of such units being either deliberately long or deliberately short in order to contain the end of a text exactly?
• 4. Do these units end with leaves originally blank, even though these may now have later items added or may now be cut away as wastage?
• 5. Do these units correspond with variations of the paper-stocks on which the codex has been written?

The manuscript, on a contents basis, consists of eleven self-contained units. These are: quires A, B-C, D, E, F-G-H, I-K-L-M, N, O, and P, and two extraneous units which have been inserted into quires of separate manufacture, bb (ff. 22-28 within B) and e (ff. 64-67 within E).[9] In each of these cases texts run continously within the units, regularly crossing the boundary of the folio or (in larger units) the quire; but at the end of each such unit, there is no text runover.

But identification of textually self-contained units remains only a preliminary

step in marking off fascicles. One needs to confirm this textual evidence with further, codicological detail. Here some problems emerge, because Findern lacks a number of features which manuscripts more professionally conceived routinely include. Thus questions 2 and 3 offer virtually no help in identifying potential fascicles—the manuscript is devoid of actual signatures and shows such persistent variation in quire size as to allow no conclusions about whether final quires have been tailored to accommodate the conclusion of textual units.[10]

More helpfully, the manuscript offers a rich array of potential or actual concluding blank leaves. In some cases, these leaves have to be inferred, since they have either been cut out (presumably original blanks removed for some different writing purpose) or now bear texts (presumably added later, as a way of avoiding apparent wastage, in a hand not in evidence in immediately proximate portions). A simple listing of the occurrence of such leaves immediately suggests correspondences with boundaries of the textual units singled out above:

• 1. At the end of bb, f.28v was originally blank and item 11 written in later (see n. 7 above); this feature suggests that bb forms a fascicle.
• 2. At the end of C, f. 43 is a stub and f. 44 has been completely excised; this feature suggests that B-C at least form a fascicle.
• 3. At the end of D, f. 59v was originally blank (it now has a legal note not included in the contents list and clearly later) and f. 60 has been excised; these features suggest that D forms a fascicle.
• 4. At the end of G, an unnumbered folio between ff. 99 and 100 has been excised. This feature has been discussed above: it reflects, not a fascicle boundary, but the effort at regularizing the mid-text border between two scribal stints.[11]
• 5. At the end of H, ff. 114 and 115 have been excised; this feature suggests that F-G-H form a fascicle.
• 6. At the end of L, an unnumbered folio between ff. 136 and 137 has been excised; the text does run over, in the same scribal hand, and there is no text loss. This feature, not at a putative booklet division, will require further explanation.
• 7. At the end of M, ff. 140-42 have been excised (f. 140 is a stub); this feature suggests that I-K-L-M form a fascicle.
• 8. At the end of N, ff. 162r and f. 164v are blank; f. 163 is a stub. In spite of the texts on ff. 162v and 164r and various other obscurities in the production of this monster quire, these features suggest that N forms a fascicle.[12]

---

67




• 9. At the end of O, f. 178v is blank (f. 178r, now with item 61, may originally have been) and f. 179 has been excised; these features suggest that O forms a fascicle. (By extension, the concluding quire P must also be a fascicle.)

This evidence singles out eight units (B-C, bb, D, F-G-H, I-K-L-M, N, O, and P) as very likely fascicles, making a total of nine when joined with the second independent inserted quire e.

Further confirmation of these units comes from the evidence of the paper-stocks. As I have already suggested, the reuse of stocks in different booklets does not necessarily prove that production proceeded continuously. However, paper-stocks can offer positive evidence for the presence of booklets: units produced on single paper-stocks or on unique paper-stocks possess an integrity which marks them off from other portions of the codex. Such units occur with frequency in Ff.i.6:

• 1. A-B-C share stock 1; this feature brings into question whether simply the textual independence of A is sufficient to claim for it the status of a separate fascicle.
• 2. bb, inserted in B (stock 1), is written on the unique stocks 2 and 3; this feature confirms the evidence of quiring in suggesting that bb is an independent fascicle.
• 3. e, inserted in E (stock 4), is written on the unique stock 5; this feature confirms the evidence of quiring in suggesting that e is an independent fascicle.
• 4. F-G have in common stock 7 (F also includes a unique sheet of stock 6). H contains the separate unique stock 8, but given the shift of scribal stints at the boundary of quires F and G noted above, this variation simply reflects the paper available to a scribe who may not even have copied in physical proximity to other producers of the manuscript. But the common and unique stocks suggest a separate fascicle.
• 5. I-K-L-M are almost entirely on the unique stock 9 (with two sheets in K of stock 4, as noted above); this feature suggests that these quires form a fascicle.
• 6. The complex quire N (see n. 11 above) contains three unique stocks, numbers 10-12; this feature suggests that N forms a fascicle.
• 7. O is mixed, part on stock 4 but part on the unique stock 13; this unique stock suggests that O forms a fascicle.

This evidence confirms a number of the units already singled out above— at least B-C (ignoring for a moment the status of A), bb, e, F-G-H, I-K-L-M, N, O. Only quires D and P, on the ubiquitous stock 4, are not again confirmed by this evidence; certainly one should take the seven units above as definitely produced as separate fascicles, and D and P as probably so.

One is left with a few problem areas, some soluble, some not. Two units, the quires A and E, are textually self-sufficient but confirmed as fascicles by no other evidence; some further investigation of the possibilities here seems in order. And the end of quire L contains a cancelled leaf which also requires discussion.

Quire A remains thoroughly ambiguous. If one were to insist upon the evidence of textual disposition as paramount, it would certainly comprise a fascicle; on the other hand, the paper evidence suggests it belongs with B-C in a single unit. The latter view seems to me the more compelling; the quire parallels F-G in providing extensive selections from Gower at the head of a textual unit. But even this evidence is not entirely unambiguous: the independent D also opens with a Gower selection.

Quire E, like quire D, is on stock 4, and it lacks any evidence of being tailored at the current end of the quire. But it does contain one peculiar feature—an originally blank leaf in mid-quire, f. 70; this leaf now bears an inventory not included in the contents list and clearly later. Moreover, the preceding text, Index of Middle English Verse number 2279 (item 23), also appears to be a later addition, on the blank lower portion of f. 69v beneath the conclusion of Chaucer's Venus (item 22). This confirmation of items looks suspiciously like the conclusion of a fascicular quire. In fact, it is perfectly possible that quire E originally ended at f. 70, since this leaf with its conjoint f. 71 is the exact center of the quire; as occasionally happens,[13] the quire may have been refolded so that its original first leaf became the first leaf of its second half.

This possibility is enhanced by the contents of the second half of the quire. The six leaves contain a single work, Hoccleve's Epistle of Cupid (item 24). Inclusion of this work at the head of the quire would follow a tendency frequent in Findern (and indeed central to the production of fascicular manuscripts generally)—the opening of a fascicle with a substantial and important work.[14]

Moreover, the appearance of this work at the head of the quire might also explain the current truncation of the text. The Findern copy of The Epistle of Cupid, related to the misarranged copies of the so-called "Oxford group," lacks twenty stanzas at the close.[15] Assuming that the work originally opened the quire, the codicological unit would have had to be at least a 16, not the current 12. The missing stanzas would then have occupied parts of the two central bifolia, now lost, and the quire would have run ff. 71-76+four lost leaves+ff. 61-63, 68-70. Loss of leaves would have been facilitated by the refolding of the quire; the resulting form would have been the extant 12 (ff. 61-63, 68-76) followed by a loose quired pair of bifolia. Especially since it contained a blank leaf, whose excision could have led to the loss of its conjoint as well, such a small extra piece of text might have been susceptible to being misplaced or suppressed as a fragment of uncertain origin.[16]

The refolding of E, if it occurred, may have been a late decision associated with the binding of the fascicles into a codex. It may have seemed appropriate at that time to place the succession of short texts which conclude quire D beside the similar short texts at the end of quire E. If one accepts that such a refolding occurred, then there is evidence for E as an independent fascicle, and the last four leaves, had they survived, would have appeared a short, fascicle-ending quire.

However, if one does not accept the idea of a refolded quire, the situation is somewhat more murky. If the Hoccleve poem was a complete copy, E must have been followed by a small additional quire (at least a 4, partially blank). In this case, it would have clearly comprised a fascicle separate from D. But if The Epistle was always a fragment, there is no reason not to believe that E formed a single fascicle with the adjacent quire D, with which it shares a paper stock. But if D and E belong together, then given the tailoring of the conclusion of D, their original order must be reversed: E would lead into the concluding D, rather than the inverse.[17]

The cancel at the end of quire L reveals some of the extemporaneous decisions by which "The Findern Miscellany" achieved its current shape. The fascicle, which now comprises quires I-K-L-M, originally consisted only of the first three of these units, ff. 117-36 (plus the leaf now excised). Virtually the entirety of this sequence of quires (down to f. 134v) was taken

up with the copying of La Belle Dame sans Mercy (item 30). At this point the quires were presumably set aside as complete; shortly thereafter, someone excised the old final folio, considering it a useless piece of waste at the end of a completed fascicle. However, this vandalism was followed by a decision to extend the textual unit, a decision of some promptness, since the unique stock 9 continues into the newly made quire M. In effect, L ends with a superseded fascicle boundary, now obscured by the addition of new texts (items 31-39; item 32 crosses the quire boundary). In addition, quire M, originally a 6, might be construed as a short, fascicle-ending unit.

Thus the physical evidence Harris so carefully assembles makes clear that the production of Ff.i.6 proceeded by fascicles. The exact number of these separate units cannot be determined absolutely. There were at least nine such and perhaps as many as eleven: minimally A-B-C, bb, E-D, e, F-G-H, I-K-L-M, N, O, P; at most A and an expanded E were also of separate genesis. This complicated shape, with its frequent and tantalizing remnants of projected modes of production later cancelled, seems typical of that informal creation of literary artifacts which one associates with the aristocratic menage, not the stationer.