The AdArchive project is a collaborative experiment that considers how feminist principles can shape scholarly and practical engagement with Linked Open Data (LOD) and the Semantic Web. The goal of the AdArchive project is to better understand what periodical researcher Felicity Tayler describes as “the social, economic, and ideological bonds” (Tayler 2022, 275) of feminist publishing communities. The data foregrounded in the project are the contents of advertisements published in American feminist-identified social movement periodicals in the late 1970s and early 1980s. The project represents these advertisements using RDF triples in a dataset that is interoperable with other datasets and queryable on the larger Semantic Web. Methodologically, the AdArchive project is animated by questions about how feminist orientations can shape Digital Humanities methods and practices. We respond to the #transformDH movement’s efforts to centre intersectional feminist approaches (Bailey et al), we take up the FAIR data principles to make data Findable, Accessible, Interoperable, and Reusable (Wilkison et al), and we engage with the rapid development of the tools and technologies of the Semantic Web. Given the centrality of LOD to the AdArchive experiment, we are particularly concerned with the impact that a feminist orientation to data authoring has on AdArchive, and how this orientation might inform future LOD projects.
AdArchive is a core project affiliated with the Linked Infrastructure for Networked Cultural Scholarship (LINCS), which “aims to mobilize Canadian scholarship through the creation, dissemination and use of linked open data” (Brown and Martin). LINCS is committed to creating interfaces that mediate among multiple perspectives while ensuring compatibility across datasets (Brown and Martin). It brings together a range of diverse source data into its core ontology, CIDOC-CRM, while encouraging the use of particular and nuanced vocabularies to provide space for contextualized and situated knowledge (Brown and Martin, Haraway). As a project that is funded to leverage existing digital research infrastructures, LINCS deploys an instance of the ResearchSpace platform, a generalized platform for representing Linked Data. Yet the complexity of the technology stack presents significant challenges for making all the data usable for research (Brown and Martin).
This paper reflects on the impact that a feminist approach to linked data authoring has on the remediated AdArchive data represented in the LINCS ResearchSpace Platform. We ask: how does our feminist data authoring praxis play out in its representation on this generalized platform? What are the strengths and pitfalls of deploying existing or generalized digital research platforms for feminist projects? Can these platforms be deployed in ways that support feminist research, or must feminist projects construct alternative platforms?
Over the past three decades, feminist scholars working in the field of digital humanities and data sciences have increasingly argued for the importance of a feminist approach to the collection, analysis, and management of data. Early foundational projects by Martha Nell Smith (Emily Dickinson’s Correspondence), Susan Brown (Orlando: Women’s Writing in the British Isles), and Julia Flanders (Women Writers Online) applied feminist methods to the creation of early text encoding initiatives, asking important questions about editorial practises, literary history, and women’s writing in the context of new online environments. More recently, in their influential collection, Bodies of Information , Elizabeth Losh and Jacqueline Wernimont provide strong evidence for the ongoing and impactful role that feminist approaches have had on DH scholarship, including its influence on the objects, methods, and praxis taken up within the field. D’Ignacio and Klein’s Data Feminism centres intersectional feminism as a framework that ought to inform the methodologies and ethical imperatives practiced by data science researchers. Intersectional feminisms are a vital resource for data scientists and digital humanists whose aims include reinvention of the field and responsibility to justice in our research and scholarly practice (Webb et al, D'Ignazio and Klein).
AdArchive’s practice of remediating advertisement data to the machine-readable format of LOD has been an experiment in feminist praxis. While the networked logics of LOD lend themselves to epistemologies that promote a decentralization of knowledge and a rejection of hierarchy, they are not inherently feminist (Corn and Patrick 2019; Sadler and Bourg 2015). As feminist researchers, we do not claim that data can be made feminist by applying specific ontologies, or by beginning with specifically feminist material. Instead, we maintain that it is a feminist approach or orientation to data-authoring that determines what makes a project feminist (Smith Elford and Meagher 2023). Feminism, in other words, is a praxis, not a static quality of things, objects, or artifacts. Throughout the life of the project, we have developed and applied feminist principles to guide our work (Smith Elford and Meagher 2023). These principles have shaped our selection of LOD vocabularies. They have influenced our decision to embrace plurality and messiness in an effort to avoid the potentially limiting nature of authority files and the hierarchical inclinations of structured data. All of our data has now been described in machine-readable triples, mapped to the CIDOC-CRM ontological structure, and reconciled against authorities. Preliminary AdArchive data has been ingested in the LINCS instance of the ResearchSpace Knowledge Map, and stage one of our dataset will be available on that platform in Spring 2024. On the occasion of this significant milestone in the AdArchive project, we are now in a position to reflect on the impact that a feminist orientation made in this LOD project.
We will share our insights on the project’s methodological successes, failures, and reflect on the opportunities uncovered for future feminist projects in LOD, particularly those that draw on platforms adapted for distinct projects and diverse audiences.