Codicological Visualization of a Syriac Manuscript with VisCodex

A data model for modelling and visualizing the structure of books in codex formats has existed since the middle of the past decade. Its name is VisColl.

VisColl is free, available for everyone to use, and the code for its current version (v1.0) is hosted by the Kislak Center for Special Collections at the University of Pennsylvania on GitHub. It was developed in 2013 at the Schoenberg Institute for Manuscript Studies (SIMS) by a team working within the then called Manuscript Collation Project. The VisColl project is led by Dot Porter (SIMS, University of Pennsylvania), who in 2010 conceived of the idea of a new tool aiming to show the structure of the original codex and digital images of each page, and Alberto Campagnolo (University of Udine), who wrote the VisColl alpha code in 2013.

The project also collaborated with the University of Toronto Libraries and the Old Books New Science lab to develop a browser-based modelling tool. An exaustive introductive video tutorial (‘Introduction to VisColl and manuscript collation’) has been recently uploaded by Dot Porter on YouTube (see below). From this video we learn that a beta version of VisColl, based on a new model that will enable users to generate, through a new tool, visualizations of very complex manuscript structures, is currently under development, and that it is expected to be launched soon. Moreover, we learn that the new collation modeler (it will be called VCEditor) of VisColl v2.0 will use as its starting point VisCodex (GitHub code), which is a visual collation web application implemented upon VisColl’s data model.

‘Introduction to VisColl and manuscript collation’ YouTube Video.

The new model will be able to describe manuscripts that read from right to left, and generate diagrams (unlike the current ones, flipped 180 degrees over the longitudinal axe) that would fully reflect the features of artifacts belonging to such important manuscript cultures as the Syriac and Arabic ones. Wonderful! So, waiting for VCEditor, let’s get acquainted with the web application VisCodex.

As an example to practice on, I’ve chosen Syriaque 377 from the Bibliothèque nationale de France (BnF), a 272-folio manuscript, copied by an unknown copyist in the middle of the nineteenth century (1854/55 C.E.) on a three crecents watermarked Italian paper (the so-called ‘tre lune’ paper). The manuscript consists of a collection of twenty-six texts from the West Syriac tradition (Serto script), mostly homilies. I have chosen this manuscript for two main reasons. Firstly, because it has a very simple and regular quire structure, being made as it is of 26 quires of 10 folios (quinions) plus a final quire of 12 folios (a senion). And, secondly, because of the fact that the copy, thanks to its “youth”, came to us without physical lacunae and in its original binding (the opposite case is a problem of which Syriac codicology scholars are well aware).

A description in plain words of the manuscript’s quire structure is given in Briquel-Chatonnet’s Catalogue.[1] The catalog also reports the folios (f. 267v and the endleaves) left empty, a duplication error (number 245 occurs twice) in the foliation, and the presence of decorated quire signatures. In addition, at a “distanced codicological examination” of the manuscript, made through its digital surrogate, available in a low-resolution form through the BnF’s digital library, Gallica, I could notice the presence of catchwords (with their correspondents over the following folio; likely supplied as a guide to the binder to arrange the quires) on each verso of all the folios, a fact which confirmed to me that no folio is actually missing.

VisCodex is an easy to use tool and its graphical interface is user-friendly. After registration and login, you access the front-end window, a dashboard that is destined to be populated by the list of your collation projects as their number grows.


Step 1: Adding a new project and project details

You can start to a project through the ‘NEW‘ button. Three possibilities are given (I chose the first):

  • Create a new collation from scratch.
  • Import a collation from VisColl XML or JSON.
  • Clone one of your existing collations.

Next, you need to fill out a three-point form specifying the following ‘Object Details’:

  • ‘Object Title.’
  • ‘Manuscript Shelfmark.’
  • ‘Manuscript Date.’

Entering something is mandatory for the first two fields. The field ‘Object Title’ should be thought of more as the title of a project than the name of an (historical) item. If you forget by chance to fill a field (a red message warns you that a “Project Title is required”).

So, having escaped – even if only partially – from an intriguing problem related to referring to a ‘title’ when it comes to this type of Syriac manuscripts (Multiple-Text Manuscripts, or MTMs [2] ) which do not have a general title, I have filled the form with a string. In that string, I indicated the type of project and specified its object by an alternative name by which the item (i.e. the manuscript) is known (a name composed of the name of the city in which the repository is located, of the repository’s short name, and of the manuscript shelfmark).

Step 2: Modeling the collation and generating the project

At this point you need to decide if you wish to start with:

  • An empty project (as you intend to build the collation model quire-by-quire), or
  • Pre-populate your collation with quires and leaves by using a formula (you start from here, typically, when the manuscript shows a more or less regular structure as regards the overall number of leaves of each quire). With that formula you can select the number of quires and the number of leaves, moreover specifying if the leaves conjoin or not.

The number and the type (conjoined or unconjoined) of leaves will be the same for each quire thus generated by clicking the “ADD” button, but you can edit the structure formula multiple times, changing the features of every single quire.

In the case of MS. Syriaque 377 I decided to first create a group of 25 quinions, to which I added, in a later moment, another quinion (Quire 1) and the final senion (Quire 27). By selecting a close quire and then adding to it, by using the ‘ADD’ button, a new group of quires is created (here, a group made of #1 quire, placed above the Quire 2 and below the Quire 26).

Next, you can set two ‘Project Options’:

  • Automatically generate folio numbers (foliation) or page numbers (pagination), and selecting, eventually, a starting number different from #1, and
  • Select a starting texture (if Hair or Flash) of the support (parchment).

MS. Syriaque 377 displays a foliation in Arabic Western numerals that has to be considered not-original. As to the Syriac manuscripts, in fact, an absence of leaf signatures is recorded. [3] However, the visualization of the foliation, too, could help in detecting relationships between the texts as they are arranged within the manuscript, and some of the physical aspects of the manuscript itself. Therefore, I set the starting number for Quire 2, that is #9, and deleted the default starting texture. I added the foliation for Quire 1 and Quire 27 in a later moment, and after fixing the numbering for Quire 25 where the same number has been written twice.

By clicking on the ‘FINISH‘ button your new project will be generated. Then, click on ‘Name’ ranged the ‘List View’. Within the box appearing on the right, you can OPEN (or even DELETE) the project.

Step 3: Opening the project and editing the collation model

Once opened, you can finally view the diagrams of the quire collation (VISUAL MODE). You can also view the collation in a table (TABULAR MODE). Editing the model is possible in both modes. Batch editing is possible by using the CTRL and SHIFT keys.

By selecting a quire and clicking on it, you can do a lot of things, such as:

  • Add the sewing, by drawing a line that goes through the middle of each quire, or
  • Change the default title of a quire (adding, for example, in that same space, some information on the number and type of the folios).

Through the NOTE manager, it is possible to create a set of leaf level annotations that can be attached to ‘Sides’ (folios recto/verso), to ‘Leaves’, or to ‘Groups’, and can also be viewed in the diagrams. I have created, for example, a note named ‘Quire signature’ to mark those folios (the first recto and the last verso folio of each quire) where the quire signatures occur (here, in the screenshot, you see a mark on Quire 1 (f. 8v) and on Quire 2 (f. 9r)). Quire signatures in BnF Syriaque 377 are made specifically in Syriac letters, placed at the middle of the bottom margin, and they appear with various, slightly different, decorations. It might be good to create, instead, individual notes for each quire (e.g. for Quire 1: ‘Quire signature: letter olaph, middle bottom margin, with decoration (4 crosses with 4 points, opposite colors)’).

Furthermore, VisCodex gives you also the possibility to export:

  • The collation diagram in .png format (EXPORT), and also
  • The collation data in .json and .xml formats.

It is possible also to get a shareable URL (SHARE menu) to your finished collation’s project.

What I found extremely useful in this tool (used here together with a distant look at the manuscript) is the possibility to see through the collation’s visualization. Only by reading the description, and unless I had drawn at least a sketch of the gathering structure by hand, in fact, I could hardly have understood that the first two leaves of the first quire and the last two of the last one function, indeed, as endleaves. From this simple observation, made possible thanks to this visualization, and moreover recalling that the text ends with some ease on the recto of the quire’s tenth folio, one could also create a hypothesis on the copyist’s choice to make a slight exception to the scheme based on quinions that he used throughout the manuscript. This could have been made in the specific case, not – or, at least, not only -, as has been assumed, for adapting the size of the last quire to the amount of text he still had to copy so as not to waste paper [4], but, more likely, for the sake of symmetry, and thus giving even better protection to the text.

The first and the last endleaves seem to have been used as pastedowns to the boards. The former would appear to be detached from its respective quire, while the last likely not (the real condition of the ending folios beeing very difficult to grasp by means of a B/W surrogate). But these are the kind of assertions that, thankfully, you can’t yet make with certainty without examining the physical manuscript.

Cover image: Bnf Syriaque 402, Tail. (Source / BnF).


For VisColl project bibliography, see

On Syriac Manuscript tradition and Syriac codicology, see the General introduction (3.12: “Syriac manuscripts”) and chap. 1.11 (“Syriac codycology”) in the Comparative Oriental Manuscript Studies (COMSt) publication Comparative Oriental Manuscript Studies: an Introduction, edited by: Alessandro Bausi (General editor), Pier Giorgio Borbone, Françoise Briquel-Chatonnet, Paola Buzi, Jost Gippert, Caroline Macé, Marilena Maniaci, Zisis Melissakis, Laura E. Parodi, Witold Witakowski. Hamburg, 2015.


[1] Françoise Briquel-Chatonnet, Manuscrits syriaques de la Bibliothèque nationale de France (nos 356-435 entrés depuis 1911), de la Bibliothèque Méjanes d’Aix-en-Provence, de la Bibliothèque municipale de Lyon et de la Bibliothèque nationale et universitaire de Strasbourg. Catalogue (Paris: Bibliothèque nationale de France, 1997): 68 (N. 377).

[2] On the ‘multiple-text manuscript’ concept, see Alessandro Bausi, “Composite and Multiple-Text Manuscripts: The Ethiopian Evidence”, in One-Volume Libraries – Composite and Multiple-Text Manuscripts (Berlin: De Gruyter, 2016): 79–121.

[3] cf. Françoise Briquel-Chatonnet, ”Cahiers et signatures dans les manuscrits syriaques. Remarques sur les manuscrits de la Bibliothèque Nationale de France”, in Recherches de codicologie comparée. La composition du codex au Moyen Âge, en Orient et en Occident (Paris: Presses de l’École Normale Supérieure, 1998): 162 (note 2).

[4] Ibid.: 162.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s