common-sense ontologies, crowdwork, ideology, information infrastructure, learning algorithms, ontology engineering, performativity hegemony, power, science and technology studies, semantic computing
This chapter researches practices of building a “meaning-centered” semantic infrastructure and aims to unfold some of the political and socio-material choices that are currently made in the data-driven restructuring of networked systems. It examines practices of ontology engineering−a core technique in semantic computing. Two research trajectories shape the field: (1) the aim to create not-yet known knowledge by inducing complex and non-trivial relationships from large datasets and (2) accomplishing this by combining human and machine intelligence. Semantic infrastructures and ontologies are fabricated in semi-automated processes by learning algorithms and the labor of ontology engineers, domain experts and crowdworkers. I argue that they co-emerge with a set of epistemic and economic practices that require careful scrutiny of their multi-layered and entangled political implications. Based on the example of developing common-sense ontologies, the chapter shows how different modalities of power interlock in these practices and thus (re-)constitute phenomena of difference and structural inequality.