The Neil deGrasse Tyson Problem: Methods for Exploring Base Memes in Web Archives

Abstract

In this paper we introduce the concept of the “base meme” for characterizing unique information artifacts that are used to make derivative, new, and related memes. Base memes are antecedents to many versions of derivative memes that are published all across the web. While they can be created in meme template generator websites, their origins and diffusion can be difficult for researchers to verify. Despite the often ephemeral nature of memes that are shared via platforms, they can be fairly reliably found in web archive collections, such as the Internet Archive and the US Library of Congress’ Web Cultures Web Archive. In this paper, we first present the existing research on memes and discuss the challenges for researchers who study them (such as identification and language detection). We then describe the importance of web archives to social media research and building robust methods of inquiry for internet history. Using archived data from the Library of Congress’ Meme Generator Archive (N=57,652), we use descriptive analysis to calculate, measure, and describe this important public web archive of memes. Our results show that this collection has a variety of “base memes” that can be grouped with their related derivative memes (which we consider to be their related works). We use language detection software to identify a variety of languages present in the archived dataset of memes. We close by describing why approaching these metrics on “base meme” image macros alongside findings for derivative versions and the multiple languages present in web archives of social media allows researchers to study a diversity of voices, including linguistic diversity, distinctions in humor, and the variety of cultural expressions present in memes.