There is a newer version of the record available.

Published May 18, 2026 | Version v1
Thesis Open

A Computational Analysis of the Style and Semiotics of AO3 Communities of Practice

  • 1. University of Chicago

Contributors

Description

While preexisting studies in digital and fan media have made note of several unique community practices relating to user-created metadata, community textual influence, and the relationship of fans to media canon, these findings have often limited their scopes specifically to that of 'fan' or 'original' work. This unintentionally constructs a dichotomy between fannish and non-fannish digital fiction, potentially implying a clean distinction between the two. Additionally, the common methodologies of studying these community practices tend towards a small-scale, qualitative approach, revealing a gap in the potential applications of computational and quantitative methods in revealing abstract or large-scale findings in this area of study. This thesis is focused specifically on the reality that 'fannishness' and 'originality' operates like the ends of a spectrum of identities that a piece of digital fiction can be attributed to. We employed a sociolinguistic-based framework focused on the role of user-created metadata on Archive of Our Own (AO3) as a way for authors to signal their perception of their work, and through it investigate how different user communities negotiate the fandom identity of different kinds of digital texts, such as original work, RPF, reader-insert fiction, etc. Using data from a corpus of about 18,000 works from the Marvel, K-Pop, and Original Work fandom tags spanning the years 2015-2025, we encoded and clustered the works of the corpus to reveal patterns in the terms used in metadata to identify works, as well as the narrative elements and writing styles employed in the works themselves. Our findings demonstrate the potential of computational methodologies in describing digital fiction communities, and indicate that metadata serves functionalities beyond its intended descriptive use, where it signals perceived belongingness (or lack thereof) of a work and its creator to certain communities on AO3. The practice of identifying a work with certain specific or generalized fandoms, or as an original work, also depends heavily on the relationship certain types of fandoms have to canonicity.

Files

Atticus Yang - A Computational Analysis of the Style and Semiotics of AO3 Communities of Practice.pdf

Files (844.2 MB)

Name Size Download all
custom python module used in the methodology
md5:d67dbf640e7f8855054aca9af85943ea
4.1 kB Download
thesis article
md5:ea7076bf6c1185e29bea5a210d5e6877
2.9 MB Preview Download
md5:79e9719e3ee56d14ff28640d0186fe52
835.5 MB Preview Download
scripts used to collect and interpret data
md5:9776305a6e3c289af0583851d9c39fe4
641.5 kB Preview Download
all graphs and visualizations made from the results
md5:ba8fc01710544d7ba3751b1aefe99085
5.1 MB Preview Download

Additional details

UChicago Information

Division(s)
Arts & Humanities Division
Department(s)
Master of Arts in Digital Studies of Language, Culture, and History