Within the recent popularity of Actor-Network Theory (ANT) in material culture studies, scholars tend to lose sight of its origin in ethnography of laboratory work. In particular, ANT studied how scientific facts are constructed and stabilized in laboratories so that they become universally accepted, seemingly platonic, categories. This paper returns to this initial insight, and links it to the long-standing issue of archaeological types. Analysis of the practices of production, consumption, and distribution of terra sigillata-Roman archaeology's most salient pottery type-shows how it became a category, how it was stabilized as such, and how this process imbued sigillata with specific agentic properties that allowed it to shape the range of possible actions in the past. By reframing platonic types as constructed categories, they can become active elements in our historical narratives.