A group of digital humanities researchers at Stanford has analyzed the text of about 5,000 British novels published between 1700 and 1900 in order to create an emotional map of London based on how geographic locations were described in the books.
The researchers’ method involved first mapping references to geographic locations using a technique called Named Entity Recognition to label words and phrases found in the books to a map. They then took the 200 words surrounding the geographic references to come up with 15,000 passages that were then given to volunteer taggers to read and assess for positive, negative, and neutral emotions. The taggers’ identifications were then corroborated by codings from graduate students and a sentiment analysis computer program. They took the most extreme emotions – fear and happiness – and mapped those to their locations.
The study bears some interesting results:
- It reveals the discrepancy between reality and fiction. While the population of London expanded from 600,000 in 1700 to 4.5 million by 1900, the geographic sprawl of the population was not reflected in fiction until 1850.
- It suggests how the authors and, by extension, their readers thought about particular places. Most readers at the time were middle and upper class, and the study reveals a strong bias for negative emotions being associated with impoverished areas and positive emotions’ association with wealthy areas.
- Fear was often associated with unnamed places (“a dark alley”), while happiness was tied more with specific place names frequented by the upper class (“the Savoy”).
The study notes that its results were not entirely unexpected, but they mark a revealing way to quantify and map emotions in literature.
Lauren Baker, LIS 653-01, Spring 2017