One of the challenges of digitizing a medieval collection lies in applying modern organizational standards to decidedly un-modern materials. Our aim is to create a digital version of each book that is both easily navigable and faithful to the original, but these two priorities occasionally come into conflict. One such point of conflict is the disarmingly mundane question of pagination.

For the staff who are at work capturing and processing the images, consistent page numbering is crucial to ensure that the images are kept in order and that no pages are missing from the digital book. In medieval manuscripts and incunabula, however, ‘consistent page numbering’ is a rare occurrence. Because centuries of scholarship rest on the esoteric and variable pagination techniques used by early scribes, printers, and librarians, we must seek to preserve the historic notations without sacrificing ease of navigation in the digital medium. This post discusses some of the issues at stake in this effort.

First, a brief review of bookbinding terminology may be helpful. Medieval books and manuscripts were assembled by folding larger sheets of parchment, vellum, or paper to produce leaves of the desired size. If each sheet was folded once, resulting in two leaves per original sheet, the book is called a folio; if each sheet was folded twice, resulting in four leaves, the book is called a quarto; three times and eight leaves, an octavo; and so on. All the leaves formed from a single original sheet are known collectively as a gathering or quire. If a book is held open, the page on the right side is called the recto; the page on the left side is called the verso. Together, they are called an opening.

The scribes and printers who produced these books didn’t number the individual pages. For centuries, however, manuscript curators have been numbering the recto of each leaf to facilitate study and conservation. This process is called folio numbering or foliation. Typically, the flyleaves of a book are foliated with roman numerals, and Arabic numerals are used once the text begins. The number given to each recto also applies to the verso on the other side of the leaf. Thus, the first leaf of text in a book will be foliated as 1 recto and 1 verso, or 1r and 1v. 1v and 2r will make up the first full opening of the text. Here, the 29th leaf of the Bodleian’s MS. Auct. T inf. 2.1 has been foliated in pencil, many centuries after the manuscript was written.

Many manuscripts contain errors in foliation, where the foliator has accidentally skipped a page or skipped or repeated a number. (In some of the Bodleian’s manuscripts, these errors have been noted on the back flyleaf.) It would be simple to override these errors now, as digitization workflow software makes it possible to number the page images automatically according to a given foliation or pagination scheme. In fact, however, it is necessary to preserve the existing foliation errors in our digital versions, so that researchers can easily find the images they need according to the folio numbers they have encountered in catalogues or scholarly texts. Imaging technicians and image processing staff will be working together to track foliation errors, and staff have also met with the Bodleian's manuscript and incunabula curators to find an acceptable way to annotate the errors within the image metadata, so that viewers can see these details in the digital versions. Where the manuscripts haven’t been foliated at all, digitization staff will add implicit foliation in brackets, so that it is still possible to navigate easily through the leaves.

Incunabula present another set of problems. Like manuscripts, they weren’t paginated or foliated at the time of creation, but many include printed or handwritten signatures. These are assigned to the first half of the leaves in each gathering, usually at the bottom of each recto; in an octavo volume, this would mean that there would be four signed leaves, then four unsigned leaves, in each gathering. Standard signatures usually use letters as well as numbers, with a letter or symbol assigned to each gathering and a number assigned to each leaf within that gathering. A screen capture of the folio listings for the Bodleian’s Douce 244 shows how difficult it can be to parse even a relatively straightforward signature numbering scheme (the signature of the current folio, i iiii, can be seen in the bottom right corner):

Signatures, which were included by early printers to help keep the leaves in order, have since been used by readers to identify pages within the text. However, in the early decades of printing, when the Bodleian’s incunabula were produced, the numbering was often done erratically. Letters and numbers were often skipped or repeated, and very often the gatherings weren’t signed or numbered at all.

Some incunabula have been subsequently foliated by hand, just as a manuscript might be. For example, the Bodleian’s Gutenberg Bible has folio numbers on every fifth leaf. However, the majority of the Bodleian’s incunabula haven’t been foliated, and scholars wishing to identify a particular leaf generally use the signatures. To facilitate this, curators producing catalogues of incunabula have worked to fill in the missing signatures and correct the erroneous ones. This is a laborious and demanding task, however, and it would be impossible to repeat it for each incunable being digitized for the Polonsky Project. It is possible that the signature information, along with other more detailed metadata, will be added to the item records at a later date.

In the meantime, to ensure easy navigation without placing unrealistic demands on digitization staff, we have been debating the advantages of adding implicit foliation — which would help to create consistency between manuscripts and incunabula on our site — versus simply numbering the images in sequence. (The Bodleian will display image numbers as well as any folio/page numbers for each digital object; for the Vatican Library’s collections, the image number will be encoded in the URL, but not displayed.) We have also been considering the question of quality control: if there are no numbers on the physical objects, it will be difficult to verify that the digital images are in the correct order. Imaging technicians will do their best to ensure that the images are in sequence, and metadata staff will do their best to pay attention to irregularities, but if there is one thing we can learn from a study of early foliation, it is that some human error may be inevitable. We will rely on the goodwill of scholars and researchers in alerting us to any errors they find.