Digital Humanities and Literary Studies

Thanks to the funding we received from the University of Potsdam (KoUP 1) and the Higher School of Economics (НУГ), we were able to organise two hackathons this year, one in September in Moscow, another one earlier this month at Fontane Archive in Potsdam. The latter concluded with a mini conference. The network analysis of literary texts remains the main business of our German-Russian research group. In 2017, though, we rebuilt our whole infrastructure so we’re able to look beyond network-analytical research questions and combine the network approach with other (quantitative) methods. Some of the scientific outcome of our efforts...

The network analysis of literary texts rests on a number of algorithmic foundations, which are often not sufficiently reflected in the field. In this regard, one problematic case is the existence of detached subgraphs. Here’s a classic example, the network of Goethe’s Faust, Part One (1808), visualised with our online tool ezlinavis (Faust being one of the examples you can select from the pull-down menu in the right upper corner): We can visually distinguish three subgraphs: the main graph revolving around Faust and Mephisto, which basically comprises the entire plot of the play, except for two detached single scenes: Vorspiel...

A couple of days ago, we presented a first version of our TEI-encoded Russian Drama Corpus (RusDraCor) at the CORPORA 2017 conference in St. Petersburg (slides). Our goal is to assemble hundreds of Russian plays from the 1740s (Sumarokov) up to the 1930s with authors like Gorky and Mayakovsky. Right in the middle, chronologically, our corpus features a number of plays by Gogol, one of which is “Театральный разъезд после представления новой комедии” (“Leaving the Theatre after the Presentation of a New Comedy”; full text at ilibrary.ru). We don’t concentrate so much on individual networks in our research, we’re more focusing...

We don’t know if you noticed, but the LINA research field (LIterary Network Analysis) has come up with pretty good PR videos lately. Look at this fancy Youtube clip produced by the “Nation, Genre & Gender” project at the University College Dublin (their project homepage is here). The NG+G project applies Social Network Analysis to Irish and British Fiction (1800–1922), their corpus involves 46 novels from 29 authors (according to the video they identified 9,630 unique fictional characters). And although the automated extraction of characters from novels has made progress in recent years (see, for example, Jannidis et al.’s paper...

Three weeks ago, we attended the annual Digital Humanities conference of the German-speaking countries (DHd2016), this time taking place at the University of Leipzig. We delivered two papers (more on them later) and a poster. And were really excited to be awarded the price for the best poster out of 78 poster submissions (listed in this PDF). I will try to quickly explain what we tried to do when creating our poster. But first and foremost, this is the poster we’re talking about, its full title goes as follows: “Distant-Reading Showcase: 200 Years of German Drama History at a Glance”....

This short article is a follow-up to our last posting, “The Birth & Death of German Playwrights”. Plotting the birth and death places of our 178 authors onto a map was bringing us closer to understanding the character of our corpus which – codenamed “Sydney” – contains 465 German-language plays. But it didn’t bring us close enough to understanding who the authors are. So let’s build a gallery with their portraits, a facebook of German playwrights, so to speak, and let’s do that automatically. We’re relying on Wikidata again and, for each author, extract a link to their principal image...

“If your metadata is good, it can help you in many ways,” mumbled Captain Obvious when we last met, and we couldn’t agree more. So let’s toy around with some metadata today to get a better impression of what our corpus of roughly half a thousand German-language theatre plays actually contains. You surely have seen the piece in Science, “A Network Framework of Cultural History”, and the corresponding lifetime-curve videos. Max Schich et al. set out to visualise “intellectual mobility” based on “spatiotemporal birth and death information (…) of more than 150,000 notable individuals”. That’s a lot of people, and...

Some of you will have seen our distant-reading showcase poster, this one (hi-res version on figshare): These are the character networks of 465 German-language dramas from 1731 (left upper corner) to 1929 (bottom right) at one glance. You can see how networks are changing over time, the first network explosions occurring with Klopstock’s “Hermanns Schlacht” (1769) and Goethe’s “Götz von Berlichingen” (1773): second row, fifth and second from the right. The network of Klopstock’s piece can be studied in detail here, the Goethe one here. All 465 network graphs can be accessed in a folder on GitHub. Character-Centric Data Visualisations...

As described in a previous post, our DLINA intermediary format stores structural data extracted from the full-text TEI files of the TextGrid Repository as well as various metadata, including the author’s name and date of origin of a play (and its publication and/or premiere date). In addition, the DLINA format also stores specific title information, three in total: the main title of a play, its subtitle (if available) and a genre title (only if a genre can be derived from the official subtitle of a play). To give an example, the first piece of our Sydney corpus, Gottsched’s “Der sterbende...

That’s right, we transcribed the talk we gave at the DH2015 in Sydney, on 2 July 2015, entitled “Digital Network Analysis of Dramatic Texts”. Please note that our grammar might appear a bit jetlagged here and there. ;) We were the last group to speak in a very interesting network-analysis centric session chaired by Glenn Roe. If you take a veeery close look (hehe) at this panorama pic, you will recognise us setting up the room together with the other speakers, Elisa Beshero-Bondar and Ryan Heuser (big hello there!): Since we used reveal.js as presentation framework, we can easily reference...

After creating our corpus and extracting the structural data that are of interest to us it’s time to run some statistics. As it is with statistical data, they can evoke manifold interpretations and sometimes have the inclination to speak in riddles. We will certainly need a few more months to make sense of all the values we computed and collected. Nevertheless, we’re prepared to offer at least some observations and insights already, all of which is still very much a work in progress. Our statistical analyses are quite rudimentary for the time being, more complex calculations will follow. However, some...

The DLINA zwischenformat we recently introduced also stores amounts of speech acts, words, lines, chars. Truth be told, we will always have to cope with some erroneous and inaccurate markup contained in the TextGrid Repository TEI files here and there, but now we can roughly specify how many speech acts are executed by each character, how many words are uttered by each of them, and the amount of letters used by everybody. These values were elevated from all dramas of our Sydney corpus, i.e., 465 dramas written or published inbetween 1731 and 1929. A complete list of all 9,913 characters...

Introduction After the structural data have been extracted and put into the DLINA zwischenformat, manual intervention is often necessary to improve the data quality and correct errors in the source data. Especially the TextGrid data proved to be quite problematic due to OCR errors and false tagging. Some of the “external” problems we encountered are (that is, problems not inherent to the text per se but introduced through automated or manual conversion to a computer-readable format and creating the markup): no or insufficient structural data encoded, OCR errors in a <speaker> names (strings), stage directions interpreted as part of a...

Our research interest focuses primarily on structural aspects of dramatic texts. The structural data is extracted from the 465 dramatic texts that constitute our Sydney corpus and then screened and edited before it can be evaluated statistically with regard to literary history. The structural abstraction is provided by a PHP script that processes the TEI files, collects all the data needed for our purpose and puts it in our own zwischenformat (roughly translates as ‘intermediary format’, the DLINA data format we developed for this project and announced in our previous post). The script and what it produces, our zwischenformat, represent...

Our working corpus is based on the 666 dramas extracted from the TextGrid Repository (the not-so simple extraction process was described by Frank and Mathias in an earlier post). This blog post will describe the criteria for selecting 465 dramas from said repository to represent our working corpus. The version number 15.07 is referring to ‘July 2015’ as we’re going to present our results at the DH2015 conference on July 2, 2015. Further versions of the DLINA Corpus will receive according versioning numbers. As the imminent reason for needing a reliable corpus with clean data is the upcoming conference in Sydney,...

As we underlined before, we can’t stop celebrating the fact that there are so many literary corpora on the web today. Just a fortnight ago, Martin Müller released the Shakespeare His Contemporares (SHC) collection, a corpus of early English modern drama, encoded in TEI Simple. We will definitely look into this corpus at a later point, but today we will again be bothering you with the depths of the TextGrid Repository. No worries, today’s blog entry won’t be as excessive as the one we published yesterday. ;) If you’re trying to work with corpora you didn’t create yourself, you will...

How Many Dramatic Pieces Are Contained in the TextGrid Repository? Simple question, seemingly. Before we try to answer it, a little heads-up: This blog post is ridiculously long. It can be regarded a proof-of-concept of what Mareike König recently said at the “Wissensspeicher” conference in Düsseldorf in the beginning of March: “Blogs have no space constraints.” (In this video, 17:45 mins. in.) True that! So here we go: Corpus building is a crucial task of many Digital Humanities projects and it is great to see a number of new corpora appear on a fairly regular basis. Many of these text...

Ok, time for some Digital Humanities fun facts! We had another meeting today and, as always, were working our way through the vast TextGrid Repository. Since we’re only interested in the dramatic texts contained in the corpus, we had to find a way to automatically extract these kinds of texts which isn’t as easy as it sounds. Anyway, we finally managed to do so and also wrote a small (well …) 30.000-character piece on the subject which is to appear later. For the time being, the extracted dramas can be found as single XML files here on our GitHub. When we...

Met today to work on our stuff for Sydney. Office panorama: Wanted to include a Sydney screenshot from International Karate (spirit of 1986!), but a link to the screenshot will do.

In a few days, March 12/13, we’re taking part at a conference at Bayerische Akademie der Wissenschaften, Computer-based analysis of drama and its uses for literary criticism and historiography: The CfP is here. The program can be found here (PDF). Our presentation will be held on Thursday, 12 March 2015, 17:15, in German: Digitale Netzwerkanalyse dramatischer Texte. Update: The conference can be relived on Twitter: #CompDrama15.

We’re going to present our first set of results at the annual DHd conference, this year held in Graz, Austria: See the conference website and program. Our presentation slot is Wednesday, 25 Feb 2015, 9:00–10:30. The slides for our presentation are here (PDF; 1.41 MB). Conference hashtag (beware, tweets in abundance!): #dhd2015.

In the last couple of weeks, Frank Fischer (GCDH), Mathias Göbel (SUB), Dario Kampkaspar (HAB Wolfenbüttel) and me sat down to reshape the whole project, “Network Analysis of Dramatic Texts”. We reworked and corrected a bunch of theatre pieces from the TextGrid Repository and added them to our corpus, compiled some new statistics and generated new visualisations: Our first round of results will be presented at two upcoming conferences, in Graz and Munich. See you there!