I’m excited to announce a new research project dubbed “Unfolding the Novel” (which is a play on both “paper” and “protein” folding). In collaboration with colleagues from the Stanford Literary Lab and Arizona State University and in partnership with researchers of the Book Genome project of BookLamp.com we have begun work that traces stylistic and thematic change across 300 years of fiction, from 1700-2000! Today UNL posted a news release announcing the partnership and some of our goals.
The primary goal of the project is to map major stylistic and thematic trends over 300 years of creative literature. To facilitate this work, BookLamp is providing access to a large store of metadata pertaining to mostly 20th and 21st century works of fiction. This data will be combined with similar data we have already compiled from the 19th century and new data we are curating now from the 18th century. The research team will not access the actual books but will explore at the macroscale in ways that are similar to what one can do with the data provided to researchers at the Google Ngrams project. A major difference, however, is that the data in the “Unfolding” project is highly curated, limited to fiction in English, and enriched with additional metadata including information about both gender and genre distribution.
Our initial data set consists of token frequency information that has been aggregated across one or more global metadata facets including but not limited to publication year, author gender, and book genre. Such data includes, for example, a table containing the year-to-year mean relative frequencies of the most common words in the corpus (e.g the relative frequencies of the words “the, a, an, of, and” etc).
I’ll be reporting on the project here as things progress, but for now, it’s back to the drudgery of the text mines. . .