Fischer/Trilcke: Introduction to the Digital Literary Studies – Kick-off Lecture

# Introduction to the Digital Literary Studies

📚     💻     🎮

[Frank Fischer](https://www.hse.ru/en/org/persons/182492735)¹ · [Peer Trilcke](https://www.uni-potsdam.de/lit-19-jhd/peertrilcke.html)²

¹Higher School of Economics, Moscow  ·  DARIAH-EU 🇪🇺 
²University of Potsdam  ·  Theodor-Fontane-Archiv

(Twitter: **[@umblaetterer](https://twitter.com/umblaetterer)**, **[@peertrilcke](https://twitter.com/peertrilcke)**)

This presentation: [**bit.ly/2Zl9MDL**](https://bit.ly/2Zl9MDL)

[Summer School "Debating Data"](https://www.uni-potsdam.de/de/isc/kurse/summerschool/dd.html) · University of Potsdam · 26 August 2019

## Study Group "Digital Literary Studies" · Schedule

- Session 1: 26.08., Monday – 11:15 to 12:15
 - Introduction to **Digital Literary Studies** & Introduction to DraCor (Drama Corpora Project – https://dracor.org/) and APIs (Frank)
 - Installing software (R, RStudio, Gephi)

- Session 2: 27.08., Tuesday – 11:15 to 12:15
 - Introduction to **Literary Network Analysis** (Peer)

- Session 3, 27.08., Tuesday – 14:30 to 16:30
 - Hands-on: Network Analysis (DraCor network data [or available network data from novels] + Gephi)

- Session 4, 28.08., Wednesday – 11:15 to 12:15
 - Introduction to **Stylometry** (Frank)

- Session 5, 28.08., Wednesday – 14:30 to 16:30
 - Hands-on: Stylometry with RStudio (data via DraCor API or other)

- Session 6, 30.08., Friday – 11:15 to 12:15
 - More hands-on, preparation for project presentations

## Today

- **I.** [Digital Humanities and Digital Literary Studies](#/1)
- **II.** [Distant Reading, a New Perspective on Literary History](#/2)
- **III.** [Data in the Humanities](#/3)
- **IV.** [Digital Corpora](#/4)
- **V.** [The DraCor API](#/5)
- **VI.** [Cultural Hacking](#/6)

---

# Chapter I

## Digital Humanities and Digital Literary Studies

### "What Is Digital Humanities?"

The website …

http://whatisdigitalhumanities.com/

… presents 817 definitions of the term 'Digital Humanities'.

## Digital Humanities According to Wikipedia

"Digital humanities (DH) is an area of scholarly activity at the intersection of computing or information technologies and the disciplines of the humanities."

Source: https://en.wikipedia.org/wiki/Digital_humanities

## Mapping the Digital Humanities (DH)

![Mapping DH (Patrick Sahle)](images/Patrick_Sahle_-_Mapping_DH_en.png)

Chart by Patrick Sahle, University of Wuppertal. – DH are "radically interdisciplinary" (Stephen Ramsay 2011).

## The DH Community

- team work rather than solitary work
- *projects* as basic units of research
- emphasis on building/crafting
- alternative teaching formats:
 - hackathons, workshops, summer/winter/spring schools
- virtual research environments
- social-network platforms not only as research objects, but as environments for collaboration and dissemination

---

# Chapter II

## Distant Reading, a New Perspective on Literary History

## Motivation: The Limits of Reading

"Life is so short! Even if you are a book eater and only need five days to read a book twice, you only manage to read 70 a year. For the 45 years of receptiveness, from age 15 to age 60, that makes only 3,150 books: they want to be wisely chosen!"

[**Arno Schmidt**](https://en.wikipedia.org/wiki/Arno_Schmidt): *Ich bin erst sechzig* (1955). In: Bargfelder Ausgabe, Werkgruppe I, Vol. 4. Zurich: Haffmans 1987, pp. 30. (My trans.)

## "Anecdotal Evidence"

"The literary scholar of the twenty-first century can no longer be content with anecdotal evidence, with random 'things' gathered from a few, even 'representative', texts. We must strive to understand these things we find interesting in the context of everything else, including a mass of possibly 'uninteresting' texts."

Matthew L. Jockers: Macroanalysis. Digital Methods & Literary History. University of Illinois Press, 2013, p. 8. ([Google Books](https://books.google.com/books?id=mPOdxQgpOSUC&pg=PA8))

## "The Great Unread"

- coined by Margret Cohen in "The Sentimental Education of the Novel" (1999), p. 23 ([see Google Books](https://books.google.com/books?id=tOgLRyVa4DYC&pg=PA23))
- many attempts in the past two decades to turn this diagnosis into a practice, cf. workshop titles like: "Text Mining the Great Unread: Data-Intensive Methods and Digital Tools for Analysis of Texts in the Humanities and Social Sciences" ([Aarhus University, 2016](https://kursuskatalog.au.dk/da/course/64216))

## Distant Reading

"[…] if you want to look beyond the canon […], close reading will not do it. It’s not designed to do it, it’s designed to do the opposite. […] we know how to read texts, now let’s learn how *not* to read them. **Distant reading**: where distance […] *is a condition of knowledge* […]."

Franco Moretti: *Conjectures on World Literature*. In: *New Left Review* 1 (2000).

## Distant Reading, or, …

- Scalable Reading (Martin Mueller)
- Macroanalysis (Matthew Jockers)
- …

## Close Reading

- emerged in the middle of the 20th century (New Criticism)
- thorough reading and interpretation of all nuances of a (literary) text, as opposed to putting all emphasis on historical, philological, biographical contexts
- example: Derrida dedicating > 50 pages to the word "yes" in Joyce's "Ulysses" (*Ulysses Gramophone: Hear Say Yes in Joyce*, 1984)
- **IMPORTANT:** distant reading does ***not*** stand in opposition to close reading, but is supposed *to complement* the shortcomings of it (especially regarding the extent of the object examined)

## Distant Reading: Example 1
### Moretti's Analysis of "Hamlet" (2011)

![Network of Hamlet](images/moretti-hamlet.gif)

(Source: [newleftreview.org](https://newleftreview.org/II/68/franco-moretti-network-theory-plot-analysis).)

## Distant Reading: Example 1
### "Mapping Shakespeare’s Tragedies"

![6 (out of 11) Shakespeare tragedies](images/grandjean-shakespeare-2015.png)

Martin Grandjean's network visualisation of 6 (out of 11) Shakespearean tragedies (Dec., 2015). Full poster and explanations [on Grandjean's website](http://www.martingrandjean.ch/network-visualization-shakespeare/).

## Distant Reading: Example 1
### "Distant-Reading Showcase"

![465 drama networks at a glance](images/distant-reading-showcase-poster.jpg)

"Distant-Reading Showcase", poster presented at #DHd2016 in Leipzig (9 Mar 2016). Download in full-res (28.88 MB) via Figshare. DOI: [10.6084/m9.figshare.3101203.v2](https://doi.org/10.6084/m9.figshare.3101203.v2).

## Distant Reading: Example 1
### Poster Detail

![Poster detail](images/poster-detail-1.jpg)

Zooming in to Goethe's pivotal Storm-and-Stress play "Götz von Berlichingen" (1773).

## Distant Reading: Example 2
### Stylometry

![stylometry, example](images/jannidis-lauer-dendrogram-kleist.jpg)

Taken from Jannidis/Lauer: [Burrow's Delta and Its Use in German Literary History](http://gerhardlauer.de/files/6013/9637/9784/Jannidis-Lauer_Burrowss-Delta.pdf) (2014).

## Distant Reading: Example 3
### Topic Modeling

![topic modeling, example](images/jockers-topic-modeling-moby-dick.jpg)

Topic Modeling (co-occurrence of meaningful words and their distribution in a text corpus) with "Moby Dick" as an example, against the backdrop of 3.345 other novels in Matthew Jockers' working corpus (*Macroanalysis*, p. 132).

## Distant Reading: Example 4
### "Between the Words"

![Alice, punctuation](images/alice-punctuation.jpg)

Nicholas Rougeux: [Between the Words. Exploring the punctuation in literary classics](https://www.c82.net/work/?id=347) (Jan., 2016).

### Quantitative Analyses

- Quantitative analyses are at the core of the Digital Humanities.
- This kind of analyses is an addition (not a replacement!) to traditional/qualitative research in the Humanities.
- The crucial part now is to find a formalisation: How do we turn culture into digits? And to what end?
- What is data in the Humanities?

---

# Chapter III

### Data in the Humanities

### What Is Data? (1/2)

- data = "a set of values of qualitative or quantitative variables" (Wikipedia)

![Mapping DH (Patrick Sahle)](images/data-types.gif)

Source: https://www.mathsisfun.com/data/data.html

### What Is Data? (2/2)

- in the context of scientific research (research data)
- examples: temperature, velocity of a car or train, surveys, crime rates, unemployment rates, …
- data is generated and collected by all kinds of organisations, institutions, companies, governments
- data in the humanities: books or corpora, files, lists, …
- raw data/primary data: unaltered, unsorted, uncategorised – data before starting to work with it

Is a novel or a poem data? (What could be measured?)

### Data/File Formats

BMP – JPEG/JPG – PNG – TIF/TIFF – GIF

PPT/PPTX – XLS/XLSX – DOC/DOCX – RTF – ODT

TXT – CSV – JSON – HTML – XML – PDF – …

Proprietary vs. Open Standards

### Data in the Humanities

- "a digital, selectively constructed, machine-actionable abstraction representing some aspects of a given object of humanistic inquiry" ([Schöch 2013](http://journalofdigitalhumanities.org/2-3/big-smart-clean-messy-data-in-the-humanities/))
- two main types of data in the humanities: big data vs. smart data

### Big Data

- 3V paradigm:
 - volume (amount of data)
 - velocity (speed of data, in-/output, real time)
 - variety (text, images, audio, video)
- "messy and implicit, relatively large in volume, and varied in form" (Schöch 2013)

### Big Data in Action

- Google Ngram Viewer: https://books.google.com/ngrams
- released in 2010
- Erez Aiden, Jean-Baptiste Michel: *Uncharted: Big Data as a Lens on Human Culture*. Riverhead Books, 2013.

### Smart Data

- usually contains markup, annotations, metadata
- "involves human agency and demands time"
- "semi-structured or structured, clean and explicit, as well as relatively small in volume and of limited heterogeneity" (Schöch 2013)

### Linked Open Data

- part of the Semantic Web
- identified by URI (Uniform Resource Identifier), conveniently available via HTTP
- RDF (Resource Description Framework): subject – predicate – object
- direct querying via SPARQL endpoints
- example: Wikidata ([wikidata.org](https://www.wikidata.org/wiki/Wikidata:Main_Page))

### Querying Wikipedia (Via Wikidata)

Go to https://w.wiki/Ahr4 and hit "Run".

Results?

Now let's refine our query.

Go to https://w.wiki/Ahr7 and hit "Run" again.

Results?

### "Charting Culture"

- Maximilian Schich et al.: A network framework of cultural history. *Science*, Vol. 345, Issue 6196, pp. 558–562. DOI: 10.1126/science.1240064.
- "Schich, an art historian at the University of Texas at Dallas, and his colleagues used the Google-owned [and now extinct] knowledge base, [Freebase](https://en.wikipedia.org/wiki/Freebase), to find 120,000 individuals who were notable enough in their life-times that the dates and locations of their births and deaths were recorded."
- … and what did they do with it?

*Charting culture*, YouTube ("nature video" channel), 31 July 2014 
(births are blue, deaths are red)

---

# Chapter IV

## Digital Corpora

## Digital Editions

- ongoing digitisation efforts (university libraries, companies like Google)
 - scans, OCR (optical character recognition)
- in the Digital Humanities, a digital edition almost automatically involves the acronym TEI
 - Text Incoding Initiative: responsible for TEI Guidelines ([tei-c.org/Guidelines/](http://www.tei-c.org/Guidelines/))
 - basically developing/prodividing a stable, versatile XML schema (~ 550 elements at the moment)

### Example 1

```xml
<div type="sonnet">
 <lg type="quatrain">
 <l>Les amoureux fervents et les savants austères</l>
 <l>Aiment également, dans leur mûre saison,</l>
 <l>Les chats puissants et doux, orgueil de la maison,</l>
 <l>Qui comme eux sont frileux et comme eux sédentaires.</l>
 </lg>
 <lg type="quatrain">
 <l>Amis de la science et de la volupté</l>
 <l>Ils cherchent le silence et l'horreur des ténèbres ;</l>
 <l>L'Érèbe les eût pris pour ses coursiers funèbres,</l>
 <l>S'ils pouvaient au servage incliner leur fierté.</l>
 </lg>
 <lg type="tercet">
 <l>Ils prennent en songeant les nobles attitudes</l>
 <l>Des grands sphinx allongés au fond des solitudes,</l>
 <l>Qui semblent s'endormir dans un rêve sans fin ;</l>
 </lg>
 <lg type="tercet">
 <l>Leurs reins féconds sont pleins d'étincelles magiques,</l>
 <l>Et des parcelles d'or, ainsi qu'un sable fin,</l>
 <l>Étoilent vaguement leurs prunelles mystiques.</l>
 </lg>
</div>
```

Charles Baudelaire's sonnet *Les chats* in TEI format.

### Example 2

```xml
<sp who="#FAU">
 <speaker>
 <hi rendition="#g">Faust</hi>.
 </speaker><lb/>
 
 Habe nun, ach! Philosophie,<lb/>
 Juristerey und Medicin,<lb/>
 Und leider auch Theologie!<lb/>
 Durchaus studirt, mit heißem Bemühn.<lb/>
 Da steh' ich nun, ich armer Thor!</lb>
 Und bin ſo klug als wie zuvor;</lb>
 […]
 
</sp>
```

First six lines of Goethe's tragedy "Faust" (1808) in TEI.

### Example 3

![Digital evolution of the text/corpus](images/tolstoy-digital.jpg)

Tolstoy Digital. Digital evolution of the text/corpus (source: [tolstoy.ru](http://tolstoy.ru/projects/tolstoy-digital/)).

### Full-Text Sources (with and without markup)

* Project Gutenberg
* Wikisource
* Deutsches Textarchiv
* zeno.org
* lib.ru
* rvb.ru
* ilibrary.ru
* feb-web.ru
* …

---

# Chapter V

## The DraCor API¹

¹ API: Application Programming Interface

### Drama Corpora Project (DraCor)

![DraCor frontpage)](images/dracor-frontpage.png)

https://dracor.org/

## DraCor Technology Stack

![DraCor technology stack](images/dracor-drawio.svg)

All repos are open source: https://github.com/dracor-org

## DraCor API (1/2)

- provides metadata, bespoke excerpts of the actual text (stage directions, spoken text per character, etc.) and various metrics
- e.g., network data for all plays in CSV and GEXF format
- live documentation via Swagger: https://dracor.org/documentation/api/

## DraCor API (2/2): Example Queries

https://dracor.org/api/corpora/ger/play/goethe-faust-eine-tragoedie/spoken-text?gender=FEMALE

Spoken text by female characters in Goethe’s „Faust“.

https://dracor.org/api/corpora/shake/play/hamlet/metrics

Basic network metrics for Shakeseare’s „Hamlet“.

https://dracor.org/api/corpora/rus/play/chekhov-vishnevyi-sad/stage-directions

All stage directions of Chekhov’s „Cherry Orchard“.

https://dracor.org/api/corpora/rus/metadata

Metadata for all plays of the Russian corpus.

## Simple Use Case
#### Number of characters per play in chronological order (1/3)

```R
library(data.table)
library(ggplot2)
rusdracor <- fread("https://dracor.org/api/corpora/rus/metadata.csv")
ggplot(rusdracor[], aes(x = year, y = numOfSpeakers)) + geom_point()
```

Very simple R script.

#### Number of characters per play in chronological order (2/3)

![number of characters per play](images/num-of-speakers-rusdracor.png)

Output in RStudio.

#### Number of characters per play in chronological order (3/3)

![number of characters per play (LibreOffice Calc)](images/num-of-speakers-rusdracor-libreoffice-calc.jpg)

Output in LibreOffice Calc (Microsoft Excel is similar).

## Ready-Made Files

(Preparation for the Coming Two Days)

* text files stripped from markup plus network data:
  * RusDraCor: https://owncloud.gwdg.de/index.php/s/0rJVhnD6pfZDjCf (179 plays)
  * GerDraCor: https://owncloud.gwdg.de/index.php/s/4gIu6Kz1fpbPmx0 (473 plays)
* Corpus of German-Language Fiction (txt): https://doi.org/10.6084/m9.figshare.4524680.v1 (2,735 files)

---

# Chapter VI

## Cultural Hacking

### Last Chapter for Today (1/4)

- "hacking culture": DH is about developing a specific kind of creativity
- answering (research) questions with data
- learning how to get from (research) question → data → numbers/visualisations → interpretation
- especially the first and last step are what makes a digital humanist

### Last Chapter for Today (2/4)

- for illustration: we want to know how much the German National Library weighs on books (initially to learn more about the extent of the printed cultural heritage)
- how do you go about this:
 - find suitable data
 - harvest/extract/convert data
 - calculate answer using appropriate technology
 - interpretation

### Last Chapter for Today (3/4)

![German National Library, pic from Wikimedia Commons](images/dnb-german-library-leipzig-2008.jpg)
German National Library; Leipzig, Saxony (source: [Wikimedia Commons](https://commons.wikimedia.org/wiki/File:Deutsche_Buecherei_%28German_Library%29_2008-Feb.jpg))

- they publicly offer a [data set](https://www.dnb.de/metadatendienste) with all their catalogue data (RDF format, > 21 GB)

### Last Chapter for Today (4/4)

- extracted data and fed it into Elasticsearch (so we can do all sorts of things with the data in real-time)
- found extent info for 6 Mio. entities (books), totaling around 1 billion pages (average: 166 pages/book)
- assuming that one page weighs 2,5 grams, multiply with number of pages and divide by two gives us: 1.25 mio. kilograms = 1,250 tons of pages (not counting book covers, etc.)
- account of this 4-hour hackathon: ["Empirical Data on Over-Length Books"](http://weltliteratur.net/Empirical-Data-on-Over-Length-Books/)

---

## Preparations

* install R, RStudio, Gephi

See you tomorrow!