Close Reading Cloud Atlas With Computers

Close Reading + DH: A Dialogue. 20th April 2018.

A book

Professor Martin Paul Eve, Birkbeck, University of London

Textual Scholarship and Visualization

Sankey visualization

But: Dark Age of Documentation

  • Untracked changes: separate documents do not exist
  • Unmapped rejection paths: selectivity of literary market
  • Digital preservation contoured by economic selectivity

Texts as Statistics, or: Close-Reading with Computers

Part of speech trigrams with more than 1.9 standard deviations of variance in any one section

Texts as Statistics, or: Close-Reading with Computers #2

Dating the words in Ewing Part I

Analogue Loops

  • Born-digital texts with DRM
  • Copyright exemptions for research
  • EU Directive 2001/29/EC
  • Re-keying or feed scanning

Character-Based Recurrent Neural Networks

  • Trained on Textual Practice
  • Statistical modelling of character runs
  • Can be pre-trained with word2vec to create accurate semantic embedding
  • ‘The series of temporal inventions of the object is intelligible only afterwards’
  • ‘Slavoj Žižek, Live Fiction, trans. Rushdie and Jean-Luc Nancy (London: Bohestock Press, 1994)’

The End

Thank you!

Presentation licensed under a CC BY-SA 3.0 license. Images excluded from CC license. Available to view online at https://meve.io/CRDH2018.