11 7. Progress of copying and construction of the manuscript
This section gathers together features that were identified in the description of each quire. They relate to the physical make-up of the quire, the text or both. These codicological observations illustrate how the scribe organised the material he had to copy, but the data present the reader with problems of interpretation. Dd resembles a draft in progress. The manuscript shows in several instances hesitations and interruptions, not only from his daily routine but also between the receipt of one or more stints of text.
The watermark pattern that one would expect from a quarto format is disturbed in three places and two of them appear in quire 5. Evidence, on fol. 105v of the transfer of a paraph mark, suggests that the scribe may have cut and rearranged folios at that point in the quire. The other manuscripts often display a different ordinatio in this sequence which shows that scribes were uncertain as to how the tales in this central part of the poem should be arranged. The same evidence was not found for the other paper bifolia used to copy L17 and the beginning of SQ. It is, however, possible that in this quire the scribe experimented with the sequence CL-L15-ME-L17-SQ. As already noted, it is difficult to determine what text could have been there for numerous folios are missing. However, for a mirror transfer of a paraph mark to be present, the replacement had to take place after the paraph was inserted. Paraph marks were normally added at a later stage in the preparation of a manuscript for completion.
The signatures of each quire are unusual. The first quire does not have any that are visible, the second uses lower case letters, the third Arabic numbers, the fourth capital letters and from the fifth onwards lower case letters with roman numerals. According to Ker (1969, pp. ix-x), it became common after the thirteenth century to indicate quire numbering with lower case letters and roman numerals. This arrangement in Dd presents problems of interpretation because either the scribe started copying the manuscript from quire 5 or he started using this technique later after the first four quires were copied. Both these explanations are possible, but it is likely that the scribe started numbering the quires from quire 5. He may have thought that quire 5 was problematic so he decided numbering was important to keep folios in order. Consequently, he could not sign the folios in the former quires with b, c etc before a, or confusion would have arisen. So he had to devise another method to indicate the former quires.
Although tales and links do not coincide with the end or the beginning of a quire, the way in which the scribe inserted them is significant. At times, the scribe starts a textual part at the top of the verso without leaving any space for its incipit. The diagrammatic representation of Dd highlights these parts in lilac that refer to L2, CL, PD, SH and L30, which are inserted, respectively, in quires 2, 4, 6, 7 and 8. In the other four quires the scribe started on the recto of the folio in the same way, with the incipit and explicit either in the header or on the preceding folio: MI L7 L15 L21 (indicated by blue). The parts starting on the recto seem to be either insertion of text in a planned gap (MI, L7 and L21) or replacement of folio (L15 ME). In Dd only quire 9 does not have any textual parts inserted in either way.
Numerous gaps are present in the manuscript, which are marked in yellow. In three instances (quires 4, 6 and 7) a gap is left in the folio that precedes those tales starting on the verso, suggesting the difficulty that the scribe had in allocating space for his material. This shows that either the space planned for the textual part to be copied was too large for what arrived or space was left for something else to come later or the text had to be rearranged. The gaps after CO, SU, SQ and between PH-L21 and PD-SH can also be found in Hg. Some of these gaps coincide with the fragments that traditionally scholars have thought may reflect stages in the circulation of the poem by Chaucer. Partridge (1993) suggests that the gap after CO and SQ may be Chaucerian. It seems, however, that in Dd the gaps relate to the ordering of the different parts.
On several occasions the scribe placed a small almost invisible prick to indicate beginning or end of certain textual parts, namely L1, RE and ML (right corner) and L14, which is an indication of planning. The scribe uses this technique rarely either because he was uncertain as to what to do in these parts or he knew that these parts were coming and their length, but he did not have them immediately available. In many tales the text of Dd has lines crammed in at the end of the verso and initial lines are written in a slightly lighter ink on the recto. In the quires it is not always easy to tell whether the scribe was casting off the text in some way or simply started to copy a tale from beginning to end. He must have carried on some sort of calculation while planning where the text should go. It is possible that, for example, he was working on a folio and its conjugates when he had available all the text he had to copy. This type of organisation implies casting off the text such as writing the first two lines at the beginning or at the end of each page that made up the quarto format. It is difficult to detect this type of evidence everywhere in the manuscript. Only in MI, ME and MO are these characteristics prominent. It is difficult to explain why this had happened. The reason could be the insertion of the tales in an already planned gap as the evidence attests. In quire 1 the text is inserted with a fewer number of lines per page and this may indicate that the scribe probably had more space that he needed and spread the text across the folios he had available.
The text is copied in different shades of inks, as explained in some detail in the codicological description of the quires, highlighting on many occasions how difficult it can be to describe the colour of the ink. The matter is even more complex when I try to represent these differences diagrammatically. Manly and Rickert (1940, 1, p. 24) note: ‘Changes of ink are often, like changes of hand, of no significance, but often such changes mark important events in the assembling of the texts and the order in which the parts of the MS were written’. It seems that sometimes changes of ink in Dd may confirm the irregularities highlighted by the distribution of the watermarks. For example, there is a clear change in ink between the last folios in MI, where the text is copied with a light shade of brown, and L2, copied with a darker shade. The same could be observed for SU-CL and for MO-L30.
The question arises whether those parts that are diagrammatically represented with the same colour of ink were copied continuously. This is again a question of interpretation. The difference between the brown and the light brown is difficult to explain, with the exception perhaps of those instances in which the difference can be explained textually. However, the use of the very dark brown is intriguing. It is used to copy quire 1, the last folios of quire 5 and SQ in quire 6. It is finally employed to copy the sequence TM, after fol. 161r, to MO and NU. It is also used to correct SU, inserting two missing lines toward the end of the tale. Assuming that this ink is used at the same time, there are two possible explanations as to when it is used. It could be the first ink used in Dd or it could be the last one. The first hypothesis is appropriate if the scribe started to copy those parts of the manuscript in which he accepts there are fewer problems of organisation, i.e. he knows how to organise these parts. However, he must use this ink after he copies SU and this supports the second hypothesis. This is an ink employed in Dd when the manuscript is heading towards completion.
In other instances, the change of ink does not relate to the organisation of the text or it is difficult to interpret this information. Manly and Rickert (1940, 1, p. 24) again suggests: ‘It is very probable that inks which are now strikingly different looked much alike originally … inks of the same shade of color may be very different in composition and may be affected very differently in the course of time by moisture, heat, and the chemistry of the air’. Owen (1980) observes that change of ink in El can give information about the text and its order. However, he warns, using Hg as an example, that sometimes a different ink could be a consequence of ‘the absorptive capacity of certain areas of the vellum … rather than the quality of the ink itself’ (Owen 1980, p. 14). He concludes his note inviting caution when a scholar is dealing with colour of ink in manuscripts at least until a more accurate terminology is put forward. It is, therefore, important to point out that the evidence of the ink has to be combined with other internal features to support any argument. On the contrary, there are instances, in which the scribe does not have any doubts about how to organise his text. On these occasions, there is only one shade of ink and no uncertainties in the watermark sequence and placing of incipit and explicit. This is evident in the sequence from ML to SU. The variation in the colour of the inks is such that it indicates that the manuscript was not copied progressively from beginning to end.
All the textual parts in Dd have been rubricated, the only exception being the end of TT and the beginning of L28. The use of incipits and explicits relates to the organisation of the text and to the ordinatio of the work (Parkes 1991). The question of the incipits and explicits in the manuscripts of the Canterbury Tales needs a more detailed study than, as far as I am aware, has yet been carried out. However, there are a few observations that could be made concerning Dd.
The text of the rubrics indicates the beginning and the end of the tales and their prologues. The rubrics guide the reader through the text giving the name of the pilgrim who is going to tell the next tale. They organise the text in a sequence of prologues and tales, indicating clearly when a tale does not have a prologue (PH and NU), and announce the sequence of the tale if too large a gap is left between one tale and the next (PD-SH ‘Here bigynneth the Shipmans tale, next folwyng the Pardoner’). Rubrics are also used to explain that a tale is unfinished because this is as much as Chaucer has made (SQ). This type of indication within rubrics reflects the observation made by Parkes concerning the function of the subdivisions of a text from a reader’s point of view. Their position, though, may indicate something that could relate to the scribe at work.
In every quire for each part I have pointed out the position in which the rubrics appear. Most rubrics are written in a planned gap (RE, L3, CO, ML, WBP, L10, FR, L11, SU, CL /tale/, ME/tale/, L17, SQ, PH, L21, PR, L28, TM, L29, MO, NU, L33 and CY). Some are written in the header (MI, L2, L7, 2nd prologue of PD, SH and NP). Others are included in the footer of the folio or just above its ruling line (CL, ME, MO/end) and one only (L1) is inserted in the margin. In Dd, the rubrics inserted in headers and footers occur later after the textual part is already in situ, for instance, L2, MI, CL, SH and NP. This practice also suggests that the scribe does not know which textual part is going to precede what he is copying. In other instances he inserts a rubric wherever he finds a space. The scribe may not be aware that all parts have to be rubricated. The insertion of the rubrics in headers and footers could also indicate that it was done at a second stage, to standardise the overall ordinatio of the manuscript. This is not an indication of a sloppy scribe.
Did Chaucer choose his own rubrics? Do any of the Dd rubrics reflect Chaucerian intention? This is difficult to say. According to Parkes (1991) an author may or may not have dealt with the ordinatio of his/her work, but it seems clear that a scribe would have imposed an ordinatio on the text whether or not the author had done any work. The difference in the style and composition of incipits and explicits in the manuscripts of the Canterbury Tales does not seem to shed light on this. The Dd scribe chose his own method of indicating the text in the rubrics, although he followed a basic pattern found also in Hg.
Another characteristic of the rubrics in Dd is that they are in two different languages and written with different scripts. The Latin ones are used in the central quires and start at the end of CO and beginning of L7 and terminate at SQ indicating the beginning of the tale. They are all written in a type of anglicana hand whose minims are regularly from 2 to 3 mm. The English rubrics introduce and end all the other textual parts and are written in bastard anglicana (notice the enlargement of the feet of the minims), which can vary from 1.5 to 3 mm. The fact that these two languages are used in two separate parts of the manuscript may indicate that in one instance the scribe may have found the rubrics in his exemplar and then he supplied the rest himself, or that he copied them as he found them. The last hypothesis is difficult to argue, as the scribe in both the English section and the Latin one is uncertain as to where the rubrics should go. There is not as yet enough information on the rubrics of the Canterbury Tales to indicate what happened in Dd. Some manuscripts have the incipits and explicits completely in English (El, Hg), others only in Latin (London, British Library MS. Lansdowne 851) and others have both. This matter may depend on several possibilities such as scribal preference, earlier arrangement of the text, Chaucerian attribution of pilgrims to their tale or simply transmission of the text, but one single position appears not to be easy to take.
These remarks point out some of the problems that the material presents. They could be related to the difficulty that the scribe encountered in ordering a text that had not as yet reached a stable order or with the piecemeal reception of his exemplar/s. The scribe is clearly aware of what is happening in the text. He may or may not have all the text available for copying, but he carries out careful planning, leaving gaps if necessary to receive other text later. It looks as if the scribe is receiving his texts in parts and has to insert the parts according to an order for which he knew only roughly the length of each textual part. It is evident that Dd is not copied from beginning to end. The scribe is not working from a complete manuscript and making a straightforward copy, and this may imply a close connection between 'the supplier of exemplars and the copyists', as Stubbs (2000) suggests for Hg. It is, therefore, possible that the Dd scribe is working in close relationship with other London scribes, who are compiling and organising the Canterbury Tales.