Skip to content

Latest commit

 

History

History
193 lines (118 loc) · 9.4 KB

rise_talk.md

File metadata and controls

193 lines (118 loc) · 9.4 KB
Raw

Distant Reading and Computational Literary Studies

Internal lecture at RISE

  • Sphären der Digital Humanities
  • Digitale Literaturwissenschaft
  • Corpora
  • Stylometry
  • Text reuse
  • Data Viz (GIS, networks)
  • Distant reading

Sphären der Digital Humanities

Sahle (2013: 6)

  • Sahle's three realms: Referenzcurriculum for DH
  • Digital Humanities pur:

Die Digital Humanities liefern zwar grundlegende Lösungsansätze für Forschungsfragen der geisteswissenschaftlichen Fächer, sind aber nicht auf diese konkreten Fragen beschränkt, sondern fokussieren auf allgemeine Grundlagen und übertragbare Lösungen.

  • Digital [Fach X]:

Hier lassen sich die Digital Humanities als Summe von alten und neuen Disziplinen beschreiben, die selbst wieder durch digitale Medien und Methoden verändert worden sind.

  • [Fach X] mit Digital Humanities:

Digitale Methoden sind dann Teil traditioneller Fächer und erweitern den Methodenkanon und das Set der verfügbaren Werkzeuge.

Digitale Literaturwissenschaft

Digital Literary Studies, Literary Computing, Computational Literary Studies

Digital and/or statistical component; computational:

  • Authors and writing: Stylometry (how to recognize computationally the literaty style in a text); intertextuality, text-reuse (literary, historical, modern texts); sentiment analysis.
  • Space: literary space, geographical features in texts (GIS), literary/historical maps;
  • Visualization: data visualization; network analysis (literary characters, metadata, stylometry): visual and mathematical approaches (metrics);
  • Distant analysis: historical evolution of genre; literary canon.

☞ Statistics:

  • Descriptive Statistics (general view about content and tendencies).
  • Inferential Statistics (to generalize from sample data to population characteristics).
  • Machine learning (classification, ...)

Distant Reading

  • The unit of analysis is not only a few texts (close reading), but a large quantitative dataset.
  • To gain a wider perspective.
  • To focus on very broad units to reveal their overall interconnections, structures, patterns, etc.
  • Macroanalysis and microanalysis: text disappears as an interpretative unit.

(Jockers 2013. Macroanalysis: Digital Methods and Literary History)

Franco Moretti and the Stanford Literary Lab.

Of the novelties introduced by digitization in the study of literature, the size of the archive is probably the most dramatic: we used to work on a couple of hundred nineteenth-century novels, and now we can analyze thousands of them, tens of thousands, tomorrow hundreds of thousands. It’s a moment of euphoria, for quantitative literary history: like having a telescope that makes you see entirely new galaxies.

"most literary concepts are emphatically not designed to be quantified" (Moretti 2013: 114).

To be able to use digital quantitative methods, it is necessary to convert a literary concept into a quantifiable unit. Moretti's paper in detail

☞ protagonist

Graphs, Maps and Trees

The three elements are each ways of spatializing numerical data. Moretti (2007): A work that adds the visual component to the analysis of distant reading.

Distant vs close reading

  • “digital methods can bring us closer to literary texts” (Eve, 2022: 4)
  • "digital practices require validation on the micro level in order to scale” (Eve, 2022: 21)

Corpora creation

  • Literary corpus preparation (OCR, HCR, Copywriting ☞ common issues for DH).
  • Corpora Rationale: representativeness (balanced), authenticity, and size (mo’better).

Example ☞ Distant Reading for European Literary History (COST Action, 2017-2022).

1200 texts in 17 languages ☞ European Literary Text Collection (ELTeC)(Github and Zenodo).

Criteria (granularity ↓)

  • Level 0: Plain text
  • Level 1: Encoding (XML-TEI)
  • Level 2: Annotation (Lemmata and PoS)

XML validated with corresponding schemas; scripts for manipulation and transformation available (xslt, python):

  • XML/TEI has a common ground with Digital Scholarly Editions, with text collections: Deutsches Textarchiv, but more restricted XML/TEI, corpora rationale for research.

  • Common ground with Corpus linguistics. Similar annotations using NLP: PoS, lemmatization, parsing (syntactic analysis); corpora rationale.

(Schöch 2022)

Stylometry

  • Digital Literary Stylistics [(Author attribution), (Sentiment Analysis), (Genre distinction)...]
  • Author attribution asumes that there is such a thing as each author's stylistic fingerprint, i.e. measurable (or detectable) unitary features of language.
  • Content words vs. function words.
  • Parameters:
    • features: MFW (More Frequent Words), PoS, ngrams
    • statistics: Cluster Analysis, Consensus trees
    • distances: delta, manhattan, etc.
    • classifiers.
  • Popularization thanks to stylo(), a package in R, but possible with other programming languages.

  • Unsupervised (no training data) / Supervised (training data) Machine Learning

Text reuse

  • Information retrieval: known vs. unknown (source in target; sources as targets)
  • Tracer, passim, textreuse
    • Featuring: words, n-grams
    • Selection: function words, ...
    • Linking: intern/extern
    • Scoring: % = reuse

  • Parallel visualization, collation,...

Data Viz

Literary GIS

Network analysis

Distant reading

Underwood (2019): Distant Horizons: Digital Evidence and Literary Change

  • Various hypotheses about the historical development of genres (gothic, science fiction and crime)

One of the central arguments of this book is that contemporary quantitative methods can be very good at representing perspectival problems and can give us leverage on that dimension of history.

  • Against the idea of science fiction as a product of market forces: it is indeed a textual construct, inherent in the work itself

  • Large corpus, 2 centuries span, HathiTrust Digital Library, annotated manually, supervised model.

Calvo Tello, José (2021): The Novel in the Spanish Silver Age: A Digital Analysis of Genre Using Machine Learning

  • new corpus of 358 Spanish novels (1880-1939), TEI/XML CoNSSA

Each text has been encoded in XML-TEI and enriched with several types of metadata (administrative, genre labels, literary information about the plot, etc.). Each file is also offered linguistically annotated with morphological, syntactic, semantic, and textual layers, including the difference between narrative direct speech passages.

  • Validation of metadata for genre (novel) and subgenre (historical novel, ...)

Conclusion

The journal seeks to expand the spectrum of computational methods for the analysis of literary texts and their (cultural, social, historical, performative) contexts with innovative methods appropriate to the subject. It provides a forum to address issues such as building literary corpora, identifying peculiarities of literary texts, domain adaptation of methods, operationalization of concepts, annotation of texts, evaluation of measures, interpretability and transparency of results, and reproducibility of research.

References

Calvo Tello, José (2021): The Novel in the Spanish Silver Age: A Digital Analysis of Genre Using Machine Learning, Bielefeld, transcript Verlag / Bielefeld University Press, https://doi.org/10.14361/9783839459256.

Eve, Martin Paul (2022): The digital humanities and literary studies, Oxford, Oxford University Press.

Moretti, Franco (2013): Distant reading, London; New York, Verso.

Moretti, Franco (2007): Graphs, Maps and Trees. Abstract Models for Literary History, 2.ª ed., London, Verso.

Sahle, Patrick (2013): “DH studieren! Auf dem Weg zu einem Kern- und Referenzcurriculum der Digital Humanities”, in DARIAH-DE Working Papers, Göttingen, DARIAH-DE, http://webdoc.sub.gwdg.de/pub/mon/dariah-de/dwp-2013-1.pdf.

Schöch, Christof (2022), "Do Sentences in Novels Get Shorter over the Course of the Nineteenth Century?", The Dragonfly's Gaze. Computational analysis of literary texts, Hypotheses, https://dragonfly.hypotheses.org/1152

Underwood, Ted (2019), Distant Horizons: Digital Evidence and Literary Change, Chicago, The University of Chicago Press.