Witches, Pizzas, and the Politics of Storytelling: Toward a Generative Model of Social Stories

Guest lecture by Professor Timothy R. Tangherlini, UCLA

In the lead up to the US presidential election, two competing narratives of political scandal took hold: Bridgegate and Pizzagate. The former was based on an actual political cover-up while the latter was a fictive narrative that cast Hillary Clinton’s campaign as party to a child sex ring.

The structure of the Pizzagate narrative reveals an important aspects of the dynamics of storytelling across social networks. As such, these internet rumors – "fake news" in contemporary parlance – allow us to model the folkloric genre of legend (sagn) and its hyperactive sibling, rumor, at very large scale.

In this presentation, I present a generative model of legend, using three target domains – Danish legends about witchcraft, Internet stories about vaccination hesitancy, and the Pizzagate “fake news” stories. I show how relatively straightforward computational methods can be used to discover the domain, extract the main actants and relationships, and derive a generative narrative framework that supports storytelling. The mutually constitutive relationship between a rapidly stabilizing narrative framework and the performance level instantiation of individual storytelling performances helps explain the observed phenomena of stability and variation. This approach has great potential for both archival research and contemporaneous research on the ideology of storytelling in politically fraught environments.

The day after the guest lecture Timothy Tangherlini is a keynote speaker at the DIGHUMLAB one-day conference