Can data-capture be a tool for feminist historiography? Can contemporary frameworks for understanding networks—actor-network theory, linked open data standards—help to shift our understanding of cultural production and literary history? This paper argues that data capture modelled on the principles of the “ego-network” is rich in its potential to address persistent problems in the recovery of non-canonical literary history. An ego-network is a data representation of the way that individuals are “embedded in local social structures” (Hanneman). Women’s cultural production has frequently been that of secondary cultural labours like editorship. Literary scholarship has struggled with appraising this secondary labour since at least 1986, with the publication of Shari Benstock’s foundational
Women of the Left Bank, in which she points out that
women were frequently the “midwives of modernism,” editors and caregivers performing who supported the construction singular male author. Likewise, scholars like Jack Stillinger (1991) have tried to break apart the “myth of solitary genius” which perpetuate the burial of these secondary labours. Despite the efforts of traditional scholarship to unsettle these myths and valorize the literary labours performed largely by women, the field has not yet succeeded in turning critical attention away from canonical, singular authors and towards what I call a practice of
distributed authorship
. I contend that this is a problem of methodology, and that the careful creation of network graphs to represent distributed authorship may assist in correcting this persistent literary historical sticking point.

This paper emerges from the early stages of “Modernism, Feminism, and the Ego-Network,” a postdoctoral research project with the major linked open data modelling project, Linked Modernisms. It concentrates on the archival collections of British activist, author, anthologist, and editor, Nancy Cunard. Cunard’s
archives at the Harry Ransom Center (University of Texas at Austin) represent a who’s-who of literary and historical figures from the modernist period, paper remnants of a professional and personal network. Multiplicity and polyvocality are the hallmarks of her oeuvre, and her texts demonstrate networked connections amongst modernist writers, ideas, and events. In an echo of the scholarly trend at large, periodic attempts to recover Cunard’s work and legacy (Chisholm, 1979; Marcus, 1995; Moynagh, 2002; Gordon, 2007) have not taken root, despite her deep connectedness. In DH, feminist digital scholarship has revealed the way that histories of literature and histories of DH have been obscured in the wake of canonical digital archival projects (Mandell, Earhart), and so the problems of archival recovery affect print and digital scholarship alike. This paper will present visualizations and theoretical concerns that emerge from the on-going building and modelling of a prototype of an ego-network for feminist archival recovery.

This project takes up a relatively simple example as a prototype for feminist data collection: Cunard’s conventional anthology
Poems for France (1944), published in the last years of the Second World War as a tribute to newly fallen and occupied France. The process of its publication is well represented by the archival collection, in which Cunard has meticulously preserved received correspondence in response to calls for contributions in the
Times Literary Supplement and individual solicitations. Cunard clearly drew upon the breadth of her literary network, as letters from well-known figures like T.S. Eliot, Cecil Day-Lewis, and Vita Sackville-West show up, whether or not they ultimately contributed poems to be anthologized. However, from the point of view of conventional literary studies, this collection offers little in the way of telling examples or golden anecdotes. Cunard has little to no editorial hand over the text of the submitted poems, many of which had been published elsewhere. She rejects few contributions, and those who decline her invitation are brief and polite. Studies of editorship in the early-twentieth century have by and large looked to the depth of involvement of individual editors like Ezra Pound in the writing of authors like Hilda Doolittle or T.S. Eliot or concentrated on authors’ own processes of revision (Sullivan 2013). Even in these studies of editorship, an impulse towards individual authorship persists. The work of the cultural contributor is contained in the perceivable strong hand of the individual, and collaboration is reduced to direct editorial intervention amongst canonical figures. The version of cultural contribution that Cunard undertakes in this anthology must be read differently to be read as an instance of cultural contribution.

My methodological approach to modelling this anthology has been to build a small dataset of the poems and letters relating to the collection. The dataset is currently in the form of a relational database. I begin from the position that Cunard represents the central node in an ego-network, and that the anthology can represent the immediate social structure in which she is embedded. I made a few key decisions in modelling my data. First, I have expanded the dataset’s definition of
publication. Inclusion in the published anthology and publication elsewhere are recorded as an instance of a poem’s publication. In addition, a poem’s inclusion in manuscript form in the Collection, whether that manuscript is received from a contributor or transcribed by Cunard is also recorded as an instance of publication. This decision aims to give similar weight to the work of solicitation and curation as it does to the instance of publication. Second, I created unique rows for each mention of a poem or of the work as a whole in the letters held in the Collection. This additional data has allowed me to sketch the shape of individual relationships across a social network that emerges in the anthology tethered to discussions of the anthology. The current dataset contains over 600 data points relating to one small anthology. This dataset is an ego-network in the sense that it take a single node in a social network—Cunard—as its tether. Rather than expanding to the whole network of modernist literary production, it is interested in the relationships between people and publications (in the expanded sense) in context of the event of the anthology.

In the current phase of the project, I am cleaning and refining my data model. I am incorporating name authority files provided to me by the HRC and maintained by standards like VIAF. I am also incorporating node type descriptions drawn from Linked Open Data ontology developed by Linked Modernisms. This phase of the project aligns with what Laura Mandell calls “guerilla coding”—in which projects that attempt to make a cultural critical interventions into technologies and standards must also make themselves legible to those same existing, problematic technologies and standards. As “Modernism, Feminism, and the Ego-Network” emerges from Linked Modernisms, it is already in conversation with the dynamics of canon creation in digital space. LiMo, already more comprehensive in its scope than many canonical DH archival projects, has made admirable attempts to redress a persistent scholarly bias against women’s cultural contributions by partnering with the Orlando project to address the equitable representation of women’s histories. But equitable representation is only one part of the on-going project of re-evaluating and re-narrating women’s historical experiences. A feminist digital humanities approach also requires that we examine the data structures in which we perform this work. In the “guerilla coding” phase of this project, I hope that my ego-network may shift the way that major digital projects construct whole networks.

The lessons in feminist data capture and modelling that emerge from this prototype dataset are laying the groundwork for modelling data in relation to the work and communities of women’s cultural production. The immediate next step in this project will be to expand the data model refined in this prototype to the study of Cunard’s other works. As the data model argues that inclusion in an archives has equal weight to traditional publication, it leaves open the possibility of treating private documents like scrapbooks as a work of cultural production. This is particularly fitting for Cunard’s private work: Cunard owned a printing press, and frequently mocked up her scrapbooks to look like volumes for publication, blurring the line between private and public work. Cunard, of course, is not the only cultural producer whose study might benefit from the creation of an ego-network, and later stages of this project will experiment with building overlapping ego-networks in line with the models developed in the prototype.