Our research explores the potential benefits of network science and social network analysis for understanding histories of literary influence and affiliation. Although topic- modeling and data-mining of texts have dominated recent digital humanities work in the field of literature, we adopt a sociological approach that looks at explicit ties between writers and texts as recorded in publication histories. Specifically, we look at early- twentieth-century poetic networks in America, Japan, and China, which are characterized by the continuous grouping and regrouping of poets around small journals (“little magazines”) according to mutual acquaintance and/or common aesthetic purpose. Given this structural feature, these real-world complex networks offer a valuable data source for analyzing how patterns of ties inform individual artistic production, how they correlate to various metadata (e.g., birthplace, education, poetic style, political ideology), and how they do so differently in different national settings. In the American and Japanese cases, we are working on an order of magnitude of hundreds of thousands of poems, many thousands of poets, and several hundred journals that span from the early teens to the early 1940s.
While we are still in the process of assembling our data, preliminary results suggest that the application of network science to such humanities-style data can help us rethink the dual articulation of cultural and social structures on a macro scale and also challenge received categories of aesthetic categorization that rely on close readings of just a small subset of texts or poets. As we will discuss, our data also has potential implications for network science because of its inherently bipartite, and weighted, structure. Furthermore, because we are concerned not only with where poets published, but when and how often they did so, our data raises questions about the evolution of transitivity in networks as a way to assess the resiliency of poetic affiliation.