Bowker’s Memory Practices in the Sciences argues that scientific knowledge is inseparable from the media, archives, classifications and infrastructures through which it remembers its own past. Rather than treating memory as simple retention, Bowker shows that scientific memory is actively produced: traces are selected, formatted, indexed, standardised, lost and later reactivated through changing information systems. The book examines how, over roughly two centuries, information technology has converged with the production of scientific knowledge, moving across historical regimes such as geology, cybernetics and biodiversity databases. Its central insight is that what science can claim to know depends partly on how it stores, retrieves and organises evidence; the archive is therefore not neutral background but an epistemic machine. A key case is the digital database, which promises unprecedented preservation while also producing new forms of forgetting, because what cannot be encoded, classified or linked may disappear from future knowledge. Bowker’s argument is especially valuable for contemporary research because it reveals memory as infrastructural labour rather than passive accumulation. The conclusion is that scientific objectivity requires attention not only to facts, but to the systems that decide which facts remain available, comparable and reusable. In this sense, science is sustained by infrastructure as much as by theory.
