The main structure of the database is a list of observations about three aspects of the manuscript. The first of these is the codicological unit (for the sake of brevity referred to as Codex), that is, the unit of writing that is a historically coherent unit of production. We distinguish codicological units from codices, because different units can be bound together in one codex at later times, or be split up over several codices. However: it was not our goal to do in depth research into the division into units of the codices we described. So, whenever these data of a division of a codex into multiple units was given to us, for example through descriptions in articles or catalogues, we recorded it in the database. But whenever these data were not available in modern scholarship, we did not start a thorough analysis. The second is the unit (or units) of content in the codicological unit (for the sake of brevity we use Text as our label): the identification of titles, authors, textual genre and date. The third is the units of marginal activity in the manuscript: Margin.
Here, again, we need a disclaimer. While it is our intention to distinguish different layers of marginal activity and describe them as separate units, this proved too time consuming for the purpose of the database. When we were able to make out clear units, for example because they were chronologically wide apart, we described them as such, but when the layering was too difficult to analyse, we described them as one layer, with additional remarks in order to point this out.
At the back-end of the database, a very full record of observations of each codicological unit was filled in. However, not all of the observation fields can be accessed from the front page of the database. For example: the record allowed for entering the collation of a manuscript, but this is not visible as a filter in the front page. In the descriptions below, the fields that can be used as filters are marked with an asterisk (*), to distinguish them from those that cannot.
For the Codex, we have the place*, shelfmark and (if a single codex consists of more codicological units, bound together at a later stage) page range. We give information about the date*, origin* and provenance* of the Codex. We identified the types of script used in the Codex, the number of hands and the language(s) used. We describe the measurements of the Codex, the collation, the page lay-out. The amount of marginal space* is a field that can be used as a filter. Please be ware, however, that this percentage gives you the marginal space on the page in the present situation. This may not always be the same as in the original situation: manuscripts were often trimmed in order to be rebound.
If specific persons are linked to the manuscript, as scribes, patrons, annotators and/or owners, we recorded that information in the database. These can be accessed separately via the tab Persons & Places*. In this part of the database, names* can be combined with roles of involvement* (author, annotator, patron, donor). Furthermore, we analysed the quantity of marginal activity in a Codex in three ways: the ‘annotated pages percentage’* for the first few quires of the Codex, ‘the empty pages percentage’* for the whole Codex, and the ‘most annotated page percentage’*.
The ‘annotated pages percentage’*: With this percentage we express the ratio between annotated and non-annotated pages in the first 40 or so pages of the manuscript. So, if 10 out of the first 40 pages have annotations (and even the smallest annotations count), then the ‘annotated pages percentage’ is 25%. Of only 4 of them have any kind of annotation, then the percentage is 10%. If they are all annotated, then the percentage is 100%.
The ‘empty pages percentage’*: With this percentage we look beyond the first few quires, and express the number of pages with no annotations whatsoever (again, even the smallest annotations count) in the whole codicological unit under consideration. Thus: if, in a manuscript of 200 pages, 10 pages are unannotated, then the ‘empty pages percentage’ is 5%. If 100 pages are empty, then it is 50%, and if there no annotations at all in the manuscript, then the percentage is 100%.
The ‘most annotated page percentage’*: this percentage expresses the ratio between blank marginal space and marginal space filled with annotations on the page which has the most annotations in the manuscript. In the lay-out section, we give the percentage of space which is not covered by the text block of the main text: the ‘marginal space percentage’. In the ‘most annotated space percentage’ we give, by estimate, the percentage of marginal space which is taken up by annotations. Thus, if the margins are completely covered with annotations (as, for example, in Codex Paris, BnF, Latin 7505, fol. 25r), this percentage is very high: up to 80 or 90%. If there are only a few glosses scattered in the margin of even the page which is most densely glossed, the percentage is very low (3 or 2%). We manually measured a number of samples, and used these to give estimates for our other cases.
In Text, we give information about the contents of the Codex: author(s)*, title(s)*, language, state of preservation*. We also assigned one or more genres* to each text (hagiography, history, poetry, liberal arts, etcetera) and a time frame (date*): ancient, late-antique, medieval.
In Margin, one or more layers of annotation are described that are found in the Codex. If the layers were easily discernible, we decided to describe them as separate units. Often, however, this turned out to be impossible within the time span that we had to collect our data. So, even if the layers are in principle distinguished and described separately, you’ll find that in practice we did not stick to that promise, and compromised by describing several similar layers as one.
For the margin units, we recorded the types* of annotations and the specific phenomena* that we encountered in the margins. In types* you will find the following list:
- marginal annotations
- interlinear annotations
Marginal annotations added to a text in the margin, for example to clarify, to add information, or discuss issues. Interlinear annotations are pieces of writing attached to a text in the space between the lines, for example to suggest synonyms or give translations (glosses) or to clarify grammatical issues. Corrections actively interfere with a main text in a manuscript: they delete, for example by scratching out, crossing out or expunction. Commentary, in our terminology, is a ‘set’ body of annotations copied into a manuscript in the margin and/or interlinear space. It contrasts with ‘ad-hoc’ annotations or spontaneous annotations, which can be expected to be individual notes made by students or readers on the spot. Attachments is the term we used for pieces of writing or drawing added in a manuscript without having a clear relation to the main texts in it. E.g. a chant added to a homiliary. Often (but not always) Probations are in the same category; these are pen-trials or doodles, added to manuscript margins and flyleaves.
In specific phenomena you will find the following list:
- construe marks
- correction sign
- critical sign
- excerption marks
- nota signs
- quotation signs
- reading cues
- source references
- technical signs
- tie marks
- Tironian notes
Most of these speak for themselves, but some of them may need some explanation. More on the terminology is found in our list of terminology, but we shall give you a quick tour here too, just to give you an idea of what kind of observations you will be offered here. We use the term reading cues for signs, letters, words or numbers which guide the reader through the structure of a text. E.g. running titles, marginal keywords, section numbers. Technical signs and/or critical signs are symbols used to express doubt or approval. They facilitate both textual criticism and content criticism. They guide the interpretation of a reader. E.g. obelus (-), asterisk (x), chi-rho symbol. Source references are names and/or titles, in full or abbreviated, added in the margin to give the source of a quotation. Quotation marks are signs added in the margin to mark a quotation (for example s- or ss-shaped signs). Nota signs are NT- or N-abbreviations, signifying a noteworthy place in a text. Selection signs are signs added by a master to instruct the copyists working for him which parts of a text should be copied into a new compilation, and in which order. Construe marks are small graphic signs which guide the reader through the grammatical structure of a text. Neums are pieces of musical notation. Tironian notes are symbols from the set of Tironian shorthand symbols. They are used, for example, to save space, or to add inobstrusive markings for the copyists or students of a text (usque hic). Tie marks are graphic signs or lettershaped signs which are used to tie the pieces of text in the margin to the appropriate passages in the main text (signes de renvoie). These types and phenomena are illustrated with some examples in this prezi presentation.
For each of these types and specific phenomena we made remarks about the frequency of their use (abundant, many, some, few, single), and we included remarks about them (for example to describe a specific shape, or refer to specific pages).
Furthermore, we supplied data about the dating* of the annotations, their language*, their script type*, their possible origin and the name of the possible annotator*, if these data were available to us.