Digital Humanities |
Character Networks in Superhero Comics
Using data from the Marvel and DC Fandom Wikis, I develop new methods for the analysis of character networks. Many existing character network analysis methods were developed for well-defined narratives, for example dramatic texts. These mostly have a clear beginning and end, and are only seldom intertextually connected to other texts on a diegetic level.
Superhero comics, however, are multi-layered, intertextual narratives: While storylines in a comic series are easily identifiable, there are overarching stories existing through the interconnectedness of these narratives: you could only read the series Miss Marvel or read her larger, more embedded storyline in the multitude of team books and events that feature Miss Marvel in some, sometimes minor, capacity.
Furthermore, these narratives have no end, as any story could be picked up at a later point and be changed. So, even though there are riches of data available with regards to pop culture narratives, there is a lack in terms of methods for delineating units of observations in these vast texts. Furthermore, most existing approaches operationalize narratives as static networks, while pop culture narratives such as superhero comics are highly dynamic.
To be able to use existing metrics for the analysis of character networks, I use community detection methods to find temporal clusters of characters at various levels of granularity. These clusters show the character configuration of the storyworlds underlying a specific narrative. Existing methods for static and dynamic character networks can then be applied.
As a case study, I apply my approach to the Marvel and DC comics universes of the 1980s and 1990s to compare narrative patterns at various levels of narrative abstraction, as well as the multi-layered trajectories of certain popular characters.