A black art: Ontology, data, and the Tower of Babel problem

Andrew J Iliadis, Purdue University


Computational ontologies are a new type of emerging scientific media (Smith, 2016) that process large quantities of heterogeneous data about portions of reality. Applied computational ontologies are used for semantically integrating (Heiler, 1995; Pileggi & Fernandez-Llatas, 2012) divergent data to represent reality and in so doing applied computational ontologies alter conceptions of materiality and produce new realities based on levels of informational granularity and abstraction (Floridi, 2011), resulting in a new type of informational ontology (Iliadis, 2013) the critical analysis of which requires new methods and frameworks. Currently, there is a lack of literature addressing the theoretical, social, and critical dimensions of such informational ontologies, applied computational ontologies, and the interdisciplinary communities of practice (Brown & Duguid, 1991; Wenger, 1998) that produce them. This dissertation fills a lacuna in communicative work in an emerging subfield of Science and Technology Studies (Latour & Woolgar, 1979) known as Critical Data Studies (boyd & Crawford, 2012; Dalton & Thatcher, 2014; Kitchin & Lauriault, 2014) by adopting a critical framework to analyze the systems of thought that inform applied computational ontology while offering insight into its realism-based methods and philosophical frameworks to gauge their ethical import. Since the early 1990s, computational ontologies have been used to organize massive amounts of heterogeneous data by individuating reality into computable parts, attributes, and relations. This dissertation provides a theory of computational ontologies as technologies of individuation (Simondon, 2005) that translate disparate data to produce informational cohesion. By technologies of individuation I mean engineered artifacts whose purpose is to partition portions of reality into computable informational objects. I argue that data are metastable entities and that computational ontologies restrain heterogeneous data via a process of translation to produce semantic interoperability. In this way, I show that computational ontologies effectively re-ontologize (Floridi, 2013) and produce reality and thus that have ethical consequences, specifically in terms of their application to social reality and social ontology (Searle, 2006). I use the Basic Formal Ontology (Arp, Smith, & Spear, 2015)—the world’s most widely used upper-level ontology—as a case study and analyze its methods and ensuing ethical issues concerning its social application in the Military Ontology before recommending an ethical framework. “Ontology” is a term that is used in philosophy and computer science in related but different ways—philosophical ontology typically concerns metaphysics while computational ontology typically concerns databases. This dissertation provides a critical history and theory of ontology and the interdisciplinary teams of researchers that came to adopt methods from philosophical ontology to build, persuade, and reason with applied computational ontology. Following a critical communication approach, I define applied computational ontology construction as a solution to a communication problem among scientists who seek to create semantic interoperability among data and argue that applied ontology is philosophical, informational in nature, and communicatively constituted (McPhee & Zaug, 2000). The primary aim is to explain how philosophy informs applied computational ontology while showing how such ontologies became instantiated in material organizations, how to study them, and describe their ethical implications.




Smith, Purdue University.

Subject Area

Metaphysics|Philosophy|Communication|Information science

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server