Information and ontologies: Challenges in scaling knowledge for development

This article calls for a conceptual and empirical research agenda on ways in which policymakers and researchers can aggregate socioeconomic information shared by diverse communities without losing contextual information that is important for extracting meaning from the data. We describe the knowledge loss that occurs when information is aggregated across diverse ontologies into databases or archives relying on a single schema and use a series of illustrative examples demonstrate the significance of this information loss for policy design and implementation. While there are important gains from information aggregation across ontologies, the potential trade‐offs involved in creating large‐scale databases are significant. The differences between locally constituted ways of knowing and the organizing ontology used for larger scale databases affects the extent to which these collections, or “knowledge banks,” provide accurate guidance for policy and action. The article draws on insights from information science and social science to discuss two classes of socio‐technical approaches for overcoming information loss at the interface between ontologies: first, technology‐enabled efforts to soften ontological interfaces by making data open, unconstructed, and available and/or creating ontologies collaboratively and, second, organizational changes that reduce the need for information to cross interfaces, such as reconstructing knowledge platforms to be more interactive, thereby decentralizing decision‐making. The framing of the challenges involved in building large‐scale knowledge banks as a matter of ontology mismatch creates an opportunity for an interdisciplinary and analytically integrated research agenda to implement and test these potential approaches.