Beyond Search: Literary Studies and the Digital Library

Martin Wattenberg, founding manager of IBM's Visual Communication
Lab, joined the workshop this month as another researcher in our
ongoing experiments with genre classification. Martin is the founder
of Many Eyes, the website that allows users to upload data and
produce graphic representations for others to view and comment on.
He and his colleague Fernanda Viégas have become increasingly interested in visualizing text are now able to produce "word trees" and "phrase
nets". These show the most commonly used words in texts, and also

Michael Witmore from UW-Madison joined our February meeting to
present the classification work he has recently completed on a sample from our corpus. Primarily a Renaissance scholar, Witmore became interested in the research program Docuscope when he found it could use data-mining and factor analysis to classify Shakespeare’s canon into the different genres.

This month we heard from two graduate students on how work done with Beyond Search is feeding into their theses. Sarah Allison presented on a chapter-in-progress on characterization in Charles Dickens. Adverbial phrases often modify character speech in Dickens (particularly his early work), interpreting the character's speech as well as reporting how it is delivered.

This week Aaron Stanton introduced us to the the data-mining capabilities of his company, BookLamp. Having identified 50 variables in prose, Aaron created software that measures novelistic style by recognizing variations in action, description, dialog, density and pacing. Readers, writers and publishers currently use it to identify stylistically similar novels, thus improving their chances of commissioning, writing or buying the fiction they're looking for. Aaron is currently running BookLamp's software on our corpus, and will return to the workshop next year to discuss the results.

Franco Moretti led our first meeting of the academic year with an introduction to the three major components of quantitative methods of analysis: a new object of study, falsification and stability, all highlighted by the initial results of our adjectives project.

We are in the process of configuring and updating the Beyond Search Workshop website. In the next few days we should have everything in place. In the meantime, note that this year's workshop will meet from 11-1, in the 4th floor conference room of Margaret Jacks Hall, on the following Mondays:

October 13
Nov 3
Dec 1
Jan 12
Feb 2
Mar 2
April 6
May 4
June 1

Jockers's project, which he presented at the 2007 Modern Language Association conference as "Beyond Boston: Georeferencing Irish-American Literature," uses Google Earth to plot where and when Irish-American literary activity took place across the United States between 1769 and 2002. He began rethinking the map of Irish-American literature long before Google Earth gave him a new way to visualize it.