Visualizing Ibn Tulun’s manuscript corpus: (1) From Palladio towards Gephi

It has been a while since I presented my first visualizations of Ibn Tuluns manuscript corpus with Palladio. Palladio is an easy-to-learn tool that works in your browser and thus is independent of your operating system. You just put your data in several simple tables and upload them there. Then you create the relevant connections between the tables in the interface. Basically, all you need is one table for the entities to be connected (nodes) and one for the connections between them (edges).

Also, you can add a second table with a different sets of nodes. In my case, I wanted to know which of Ibn Tulun’s works could be found in which manuscripts. As you can find in the first post, I added a second table for those (or actually the other way around). This table has further geographical information so you can plot the data on a world map and see where most of Ibn Tulun’s manuscripts or works are held today. The two variables give different results as one manuscript can contain up to more than thirty individual works.

In the meantime, however, the data sets uploaded in the prior posts have become somewhat outdated. I have discovered new manuscripts through archival visits and by simply cleaning up my data set. I also added more consistently information about when manuscript copies were produced and on which topics Ibn Tulun wrote the most works. Since I can be impatient and want to see results for the repetitive work that is data entry, I initially used Palladio again and I like the results. And this post will only deal with Palladio, its allure and shortcomings.  Number (2) will then move to Gephi.

The first image shows the most central themes in Ibn Tulun’s corpus. Note in the lower left center the big cluster of hadith related topics, including key words such as hadith, isnad, and so on. The bigger a circle is, the more works it is related to. One issue is, however, that this does not work the other way around: The light circles for works get bigger when you apply more keywords to them. That means that the better you know a work and the more effort you spend on defining it through keywords, the more important it appears in the graph.

Ibn Tulun’s works (light) in relation to topics covered (dark).

In a second step, I looked at the temporal distribution of copies of Ibn Tulun’s works. In order to estimate the diachronic relevance of an author, this is an important step and network visualization can actually be a helpful tool in making rhythms of reception apparent. The second graph shows how works (dark) relate to moments of production (light).

Ibn Tulun’s works (dark) by year of manuscript production (light).

The biggest cluster of works (top right) were produced by 950 AH, shortly before Ibn Tulun’s death. This is an oversimplification; I chose this date for most autographs that cannot be dated otherwise. The only exemptions are works not mentioned in Ibn Tulun’s work list. I assume that they were written even closer to his death in 953 AH.

In the lower half you see proof of an evenly distributed albeit low degree of reproduction between the 16th and 19th century. Finally, the cluster in the top left refers to more concentrated copying activities in the early 20th century. They took place exclusively in Egypt and Syria. There is a visible overlap between these four years as well as with the autograph cluster. It might be interesting to run this visualization again but with decades or centuries instead of individual years. It might also be interesting to include photographic reproductions of manuscripts. It would certainly emphasize the importance of that period for our current evaluation of the author.

Continuing in this line of thought, in what were later readers interested when referring to Ibn Tulun’s works? We cannot automatically assume that reading interests remained the same always and everywhere, unless we’d buy into tropes of an “unchanging Orient”. Let’s not do that.

Instead, one way to get an idea of diachronic developments in reading practices could be to look once more at the topics covered. This time, I related them to the type of manuscript: either autograph or copy.

Topics (dark) relating to type of manuscript (light)

We can see three clear clusters of topics emerging: those that are only covered in autographs, those covered only in copies, and those covered in both. The graph also gives away something about the number of manuscripts addressing certain topics: obviously, the most important topics can be found in the shared cluster. A next step could now be to identify those copies that deal with topics not found in the autograph or shared cluster. Whereas those clusters illustrate continuations of interest (through reproduction or preservation), this cluster could also show changes. To substantiate this hypothesis, we need to look at the times and places of the production of those copies.

This approach does, however, not account for the reproduction of individual works. For instance, the keyword “biography” to be found in the shared cluster refers to a rather significant part of Ibn Tulun’s corpus—overall some fifty works, some of which rank among the largest he has written (see e.g. here). If we only look at the three largest ones, al-Ghuraf al-ʿaliyya fī tarājim mutaʾakhkhirī al-ḥanafiyyaal-Tamattuʿ bi-l-iqrān bayn tarājim al-shuyūkh wa-l-aqrān, and its Dhayl Dhakhāʾir al-qaṣr fī tarājim nubalāʾ al-ʿaṣr, they show quite divergent patterns of transmission.

The first work, which focuses on Hanafis, was copied only once between the 16th and 17th centuries and is nowadays held in Istanbul (the autograph volumes are held in Cairo and London). It is possible that this copy was deliberately made for the Istanbuli context, whereas the other two works were reproduced in a decidedly Damascene framework. The second one only survives in an excerpt from early 17th-century Damascus (now held in Berlin). The final one exists in a fragmentary autograph and three copies made between the 18th and 20th century (held in Cairo and Gotha), thereby indicating different rhythms of reception. The final work indeed might have provided then influential Damascene families such as the Shuwaykis with social capital.

But I digressed. Let’s get back to Palladio and network visualization more general. Palladio shows itself as an easy tool to play around with your data. I think the graphs above make that clear enough. Yet, it has its limitations. Miriam Posner says that more clearly in her introduction to the topic than I can. Palladio does not allow you to use color schemes for your edges and nodes; it does not show the weight of the edges either. And finally, it gives you no tools for further analysis: which nodes are central in connecting clusters to each other and which have more connections to on another than others?

In order to answer these questions, we have to use a more powerful tool, which could be either Gephi or Cytoscape. But both tools require to enter data in a different way wich reduces the bipartite network described here to a monopartite network. In other words, we have to turn either manuscripts or works from nodes to edge attributes. And that will have to remain for a future post.

An ijāza for Ibn Mufliḥ

Akmal al-Dīn, a scion of the prominent Ḥanbalī Ibn Mufliḥ family, was mentioned here several times before and was probably the most visible of Ibn Ṭūlūn’s students in terms of engagement with his writings. Ibn Mufliḥ copied them, annotated and rubricated them, and also added material to some of them.

Among the evidence of their student-teacher relationship survives an ijāza Ibn Ṭūlūn penned for Ibn Mufliḥ after the latter’s discussion of some chains of transmission he had received from him (for the transcripts, see github). This document survives in the Istanbul manuscript MS Laleli 3747, fols. 192b-193a. Unfortunately, I have no further information about the other contents of this manuscript except that it did not contain any other writings by Ibn Ṭūlūn.

The two folios that I have seen carry a number of textual items. Two are directly concerned with acts of transmission. The ijāza proper documents a session on 9 Shawwāl 941 in the Umayyad Mosque and covers about one and a half pages. It is followed by a shorter ‘update’ of about three lines, which testifies to Ibn Mufliḥ’s discussion of an introductory work (muqaddima) on logic by one al-Sāghūjī about two years later (25 Rabīʿ II 943).

In addition, there are three items to be considered for the history of the document’s own transmission. On the originally empty recto of the first folio (192b) Ibn Mufliḥ added title information:

اجازة كاتبه اكمل 
من الامام محمد بن طولون 
رحمه الله نعم
To its left, a different hand added a note stating the date of Ibn Ṭūlūn’s death (13[?] Jum. I 953) and the place of his grave (near the Cave of Abel on the slopes of the Qāsyūn). (I leave aside here a note in pencil that gives Ibn Ṭūlūn’s authoritative name and must be more recent). It is clear from the wording of the title that Ibn Mufliḥ added it after Ibn Ṭūlūn’s death and thus at least a decade after the transmission. Finally, below the end of the text a waqf stamp can be found testifying to the movement of the manuscript to Istanbul. I read the inscription as follows.
Library Stamp, Süleymaniyye Kütüphanesi, MS Laleli 3747, fol. 193a

هذا وقف سلطان الزمان / الغازي سلطان سليم خان / ابن السلطان مصطفى خان / غفى عنهما الرحمان / ١٢١٤

The two names mentioned here refer to Ottoman sultans of the 18th to early 19th centuries. The stipulator of this book endowment was Selīm III, known as a reformer and patron of the arts. As his father, the mentioned Muṣṭafā III, he was buried in the Laleli Mosque in Istanbul, which Muṣṭafā had constructed and where Selīm’s book endowment remained until its transfer to the Süleymaniyye Library. The date on the stamp indicates the addition to the endowment of this manuscript occurred in 1799.

The trajectory emerging from this information suggests that Ibn Mufliḥ acquired the ijāza at some point following the second iteration of transmission or, more probable, following Ibn Ṭūlūn’s death. Whether he took it with him to Istanbul or whether Selīm III bought the manuscript from Damascus cannot be ascertained without an analysis of the entire codex. The identification of the other scribe on fol. 192a might also help to pinpoint the transfer.

Back to the ijāza itself. It stands out among Ibn Ṭūlūn’s documents on transmission in that it starts out with panegyrics whose length is comparable to those of many of his writings. The Hamdallah covers three-and-a-half lines. It is separated from the following by a baʿd which leads into another seven-and-a-half lines of elaboration of Ibn Mufliḥ’s name and pedigree (as well as Ibn Ṭūlūn’s connections to members of that family).

Following this the presentation itself is lauded and finally we learn the subject of the session: the Ḥanbalī legal compendium (mukhtaṣar) of “the shaykh, the imām Abū al-Qāsim ʿUmar b. al-Ḥusayn al-Kharaqī” (d. 334 in Damascus; Onomasticon Arabicum, note ID: 12269). Until today, the importance of this work is overshadowed by that of Muwaffaq al-Dīn al-Maqdisī’s al-Mughnī ʿalā mukhtaṣar al-Kharaqī. Muwaffaq al-Dīn and his brother Abū ʿUmar, founder of the ʿUmariyya Madrasa, also have a part in the chain of transmission given by Ibn Ṭūlūn (fol. 192b).

The rest of the ijāza is concerned with the transmission through several chains (musalsal). One goes through Ḥanbalīs or, more specifically, through those residing in Ṣāliḥiyya or belonging to the Maqdisī family. Another goes through Shafiʿīs and Egyptians. And it obviously these chains which were the reason for documentation. This ijāza was one of riwāya, not teaching. Garrett Davidson describes this expansion of hadith transmission into other fields (Davidson 2014, 209):

The transmission of books through chains of transmission according to the protocols of hadith, has its roots in the idea that, like a hadith, it was only though a chain of transmission of trustworthy transmitters that the attribution of a text to an author could be reliably established. The chains of transmission it would be stated, “are the genealogies of books,” without a chain of transmission, to follow the analogy, a book was like a fatherless child cut off from sources of legitimacy.

This stands in stark contrast to the second, much shorter note following the ijāza proper. It seems to have been written much more hastily and is bare of visual aids to navigate the text—in the ijāza red ink is used to indicate new sections. Even the black ink is not as even. More importantly, this note documents only the subject, place and date of the act of transmission but omits information on its chain.

One possible reason is that this was session on a “muqaddima … fī ʿilm al-manṭaq” was in fact part of a more formal education for jurists. Therefore, its main concern was to record Ibn Mufliḥ’s successful exam on this text. The juxtaposition of both ijāzas could explain the terseness of this record, since all the necessary background information on Ibn Mufliḥ is given on fol. 192b. But the difference in what is recorded in both cases is striking.

Thoughts after Workshop: Textual Analysis Using Stylometry (AUB, 24-25 April)

On 24 and 25 April 2018, Najla Jarkas organised a workshop on “Textual Analysis Using Stylometry” at the american University of Beirut. The workshop was hosted by David Wrisley and given by Maciej Eder, an Associate Professor at the Institute of Polish Students at the Pedagogical University of Krakow, Poland, and and at the Institute of Polish Language at the Polish Academy of Sciences.

For those who are not familiar with the term stylometry (as was I before the workshop), it is a method of textual analysis that does not focus on the big and exceptional words but on the frequency of every-day terms in writing. By measuring frequencies and distribution, it has become a major method for everything from authorship attribution (note the case of the identity of the pseudonym Elisabeth Ferrante) and proximity to proving plagiarism.

Within the workshop, we took first steps in using the software R on a preassembled textual corpus of 19th-century English literature. We used Dr. Eder’s stylometry package for R which also can be found on his github page. You also find there two corpora to try out what it can offer you. Following these first steps I tried to apply stylometry to my own small corpus of teaching certificates getting mixed results. This might be in the nature of these texts, being epitomised by the most common three-word combinations being “Muhammad bin Muhammad” and “bin Muhammad bin”. Within the small sample, there was also little proximity to be found between the different texts.

Yet, as digital methods so often do, they make you think about your own methods and conditions of research. And R seems promising when you think of the possibilities in analysing premodern Arabic texts. As historical texts in particular were often to a large degree the product of recompilation of earlier sources, would it not be cool to find out where an author most probably copied from another one? This is basically what Sarah Savant’s Kitab Project promises to achieve.

But what would we need to enable the machine to help us with our texts? Clearly, it depends on what you work on: the period, the genre, and the fame of “your” authors all matter for how fast you can accumulate a relevant corpus. And it is generally better for you if you work with edited sources only, since chances are that you find a copy of those on shamela or another grey website. Even better, you might find them on the github page of Maxim Romanov’s project Open Arabic (current URL / new URL). This is by far the largest and best curated repository of Arabic texts you can hope for at the present moment.

Still, it suffers from what I’d like to call the “edition bias”, which distorts our perspective of Arabic literary of, more so, writerly heritage in much the same way as the better-known survival bias and the more recent “digitisation bias” do. It tends to over-represent works of the Abbasid period, for instance, over those of the Ottoman Period. This is not Maximov’s fault, I have to emphasise! It is simply the nature of modern academic tradition which, in turn, also has built upon the modern and traditional Arabic traditions in valuing and hierarchising literary works and other writings according to often implicit standards.

In other words, despite its many merits, I urge you to not to take the numbers it gives you for included works from different centuries as statistically sound representations of the ups-and-downs of actual literary production. I have before spoken about the size of Ibn Tulun’s written corpus: the surviving works alone range in the hundreds. So, what do we find on Open Arabic?

Ibn Tulun works in Open Arabic

Six works, only one of which has a considerable size! Even Ibn Abd al-Hadi (see IbnMibradHanbali), whose works were rarely copied in manuscript, has fourteen entries! One reason for the selection of works by both authors is a continuing interest in the Arab world in their religious and ethical writings. Other works, some of which have been introduced on this blog before, are currently not regarded as edition worthy (you see, there is the bias).

So, one main obstacle in applying stylometry to my own project is the lack of a sufficient corpus. As Dr. Eder explained, however, you would not need to transcribe everything in your corpus. A few passages per book might be enough to get you started. And it might be interesting to run these analyses not only when your corpus has been finalised but regularly in the process.

But I was wondering also whether we can actually use this method for cases like teaching certificates and other documentary sources. Even though Florian Sobieroj has shown in an article, that these texts never assumed a uniform format, we cannot get around the long chains of names connected into long chains of transmitters and interrupted by the likewise long titles of books. There is not much verbal action happening in those—nor in much of hadith literature.

In turn, this is connected to the question which parts of such texts were actually conceived of by their authors. Which parts of such texts could we use as the standard by which to evaluate the authorship of other passages or works? The basmalla? The tamhid? The justification of the work? The chains of transmission? The interpretation of their source material? This is a tricky question and I have no answer for it. And this comes before even problems connected to machine-readability are taken into account. These include the issue of the assimilation of the “wa” to the following word, the distinction search functions make between words with or without article, plural forms, let alone letters that become visible in writing only when the letter is present.

So, where is the silver lining of all those obstacles I keep ranting over? I would say that all of us can be the silver lining. We just need to share our own transcripts with each other and the Open Arabic (soon Open ITI) project in particular, thereby enlarging our shared corpus of machine-readable texts. These can even be partial ones like my contents list of Ibn Tulun’s biographical dictionary al-Ghuraf al-ʿaliyya. And the benefits of sharing your data go way beyond the applicability of stylometric analysis. It is perhaps the best entry point to understand better the opportunities digital humanities offer for our research and its dissemination. Or perhaps I am just idealistic in this matter…

Digital Perspectives on hypotheses

Earlier this year, I attended the second instance of the Digital Humanities Institute Beirut (#DHIB2017). It was the second of its kind in the Arab World and, thanks to the generous support by AMICAL it offered participation free of charge (also it included nice gimmicks such as a reusable water bottle which I have since lost, a high-quality tote bag, and a notebook with its own sticky tags and notes).

The DHIB featured, among other workshops, two sessions on markup languages: multimarkdown and TEI XML. Since I am still struggling to understand how to integrate either of those into my own research and publication process, I will not dwell on those – mostly failed – efforts here.

Now, I kind of forgot about this entry for some time and some of it might seem dated by now. Yet, as the first ever digital humanities internship at the Orient-Institut Beirut draws to a close, bringing us that much closer to a publication of our Arabic text editions in a workable html format (besides the classic print and open access PDFs derived from it), it might regain some of its earlier timeliness.

Instead of summarizing the DHIB, it might be a better idea to point out those blogs on hypotheses which are much further in their mastery of digital possibilities and thus offer advise to others who want to stride down this road.

The first one is foxglove which gives you an overview of French DH initiatives. It also introduces workflows for TEI related edition projects for Latin script texts.

In contrast, Freakonometrics is rather concerned with optical character recognition and machine-readability of texts (e.g. PDFs). In general, it is rather about text analysis than text enrichment but it also provides perspectives on the usefulness of markup languages.


Personally, I find the third one, Himanis, the most interesting. Himanis is short for “HIstorical MANuscript Indexing for user-controlled Search“. It speaks to me mostly because it brings together Digital Humanities perspectives and an understanding of books as objects. The notion of books as objects becomes especially important when thinking about their organization in book cases or shelves. How they deal with that in their TEI based edition of royal charters, they explain here.

Obviously, hypotheses has much more to offer which I will not address here at length. And for those who, like me, read French at a snail’s pace, there are also all-English sites. For instance, digilex covers the creation of a digital dictionary of spoken German with TEI and XSLT.

But one of my favorites is certainly The Recipes Project. They again start out from a distinct interest in manuscripts and old texts. In addition to sharing their insights about encoding these texts in a digital format, the blog also provides useful descriptions of manuscript collections, digitization, historical trajectories of books, teaching with manuscripts, and flabbergasting bits and pieces from the history of strawberries. Even though it deals exclusively with European texts, everyone should check it out.


Ibn Ṭūlūn’s Autograph Corpus: Online Catalogues

Making sense of a manuscript corpus as large as that of Ibn Ṭūlūn can be tiresome. In particular, since his autographs are today dispersed in libraries in Syria, Lebanon, Egypt, Turkey, Germany, England, Ireland, the Netherlands and the United States, and the relevant institutions offer very diverse degrees of information about the manuscripts in their catalogues. Yet, these catalogues are obviously a central source in establishing any manuscript corpus.

Things get especially irksome when one wants to make sense of the collections of the Dār al-Kutub (for guidance to its collections, see this review) and the manuscript center of the Arab League (GAD) in Cairo. Access to microfilms at Dār al-Kutub is comparatively easy if rather slow (due to the limit of three items one may order each day); access to the physical manuscripts themselves, however, requires good connections and a lengthy stay. One might circumvent the Dār al-Kutub altogether by turning to the GAD, which offers an online catalogue to browse the selection even beforehand and allows to collect digital copies within the same day.

The GAD offers access to the (partial) holdings of several important manuscript collections of the Arab World. Yet, therein lies a danger as well as new possibilities. The catalogue was apparently created on the basis of the individual library catalogues but their holdings were not compared to each other. What I mean by that: Several collections hold microfilms of manuscripts belonging to another institution. In the GAD catalogue, the same material artifact may thus appear several times. In some cases, doublets might be visible through an allusion to the original collection in the “subject” section but often enough it is not.

How Digitization Has Changed the Cataloging of Islamic Books

This is, however, not a new problem. With the spread of microfilm and microfiche in the early 20th century, collectors and scholars alike made ample use of these technologies, and, among others, the Taymūriyya Library (today in Cairo) and the Arab Academy in Damascus supplemented their manuscript collections by acquiring microfilms from other institutions around the world.

These microfilms received their own call numbers, being treated somewhat similarly to manuscripts physically present in the respective collections. Even though this process is clearly important to the constitution of the corpora which ground our understanding of Arabic literature and book culture, to my knowledge no study has yet been dedicated to this important chapter in the transition from a living manuscript to a print culture (and hence to digitization). Yet, in the face of the ubiquity of the medium, we need a new “codicology” for  microfilms.

To conclude with just one basic example, the integrity of a text in a physical manuscript is partly assured by the use of catchwords on the verso page of each folio. The text is bound to the materiality of the two-sided leaf of paper. In contrast, microfilm images present a surface to the viewer, which usually covers the verso page together with the recto of the following leaf. In order to assure that the following image actually refers to the following double page in the physical manuscript, it would require a catchword on the recto page.

While this might seem a theoretical issue at first sight, the microfilm of the Ibn Ṭūlūn autograph manuscript Majāmīʿ Taymūr 759 is only one case where the addition of such catchwords would be really necessary. For about the first half of the files I bought from Dār al-Kutub, the slides only seem to be in the opposite order (fol. 100 comes before fol. 99 before fol. 98 etc.). Yet, around the middle, the confusion increases even more and it is painstaking to find out which image belongs to which work. Finally, about half of the texts ascribed to this manuscript by the latest print catalogue is completely missing from the microfilm.