Exploring Homestuck

Main | XML Homestuck | Character Statistics | Profanity | Topic Modeling

A young man stands in the first page of a webcomic. It just so happens that this webcomic, Homestuck, is an innovative multimedia internet adventure. Though it was over three years ago this comic was given life, it is only today it will be subject to computational research in the humanities!

This is a project in progress. Two amateur researchers will explore the ways that Andrew Hussie, the author of the webcomic Homestuck, distinguishes the stylistic voice of various characters. An author often consciously varies his writing style in order to create realistic characters. After all, no two humans speak alike. An author can do this in various ways, and we have compiled a pile of variables in speech that we will identify in the text of the webcomic:

• Capitalization habits
• Sentence length
• Grammatical errors per utterance (such as subject/verb disagreement, run-on sentences, omitted subject)
• Passive versus active sentence structures
• Ratio of commands to questions to declarative sentences
• Topics of speech (Football? Horses?)
• Profanity
• Catch-phrases
• Sophistication. We may measure this by length of words and variety of words.
• Use of metaphors and similes. This would require a human, not a computer, reading the text and marking it up.
• Use of intensifiers (Really, so, hella)
• Use of emoticons
• Amount of rapping in speech
• Punctuation style (Repeated points and marks??? Lack of punctuation?)

Our topic has been inspired by the field of Stylometrics. Furthermore,computers will help us analyze text in ways that would be practically impossible without them.

Send fanmail to the researchers:

Kate Pigliacelli: klp77@pitt.edu

Mary Zuzack: mvz2@pitt.edu