Technology

Technology is the made world of tools, systems and infrastructures, considered not as neutral instruments but as arrangements that shape what is possible, and for whom. The instrumental view, on which a tool is a means that leaves our ends untouched, has been resisted from several directions: Heidegger’s account of modern technology as a way the world is revealed, Winner’s argument that artefacts can have politics, and Ihde’s postphenomenology, which studies how instruments mediate perception and action. Star and Ruhleder added that infrastructure tends to be invisible until it breaks.

Two features of technology carry most of the weight in my work. The first is architecture. Whether a system treats people justly often depends less on its technique than on its structure, and in particular on whether, at the point of decision, anyone is present who can see that a rule is failing the case in front of them. A clerk who has known a widow for a decade can honour her pension claim though her papers are lost. A biometric gate cannot, because manifest entitlement is not among the things it can read. Bovens and Zouridis traced this shift from street-level to system-level bureaucracy; my claim is that, for algorithmic welfare systems, the same architecture also produces the standing of those it governs.

The second is that instruments are sources of knowledge in their own right. A barometer does not testify, yet a competent reader learns the pressure from it. The classical epistemology of instruments presupposes transparent calibration: we know what a thermometer is calibrated to, by what procedure, and how it fails. Language models lack this, and the deeper trouble is not the user’s access to the calibration but the calibration itself (see instrument reading).

A calculator is not deflated by the observation that it is not a mathematician; it is correctly described. Much of the philosophy of technology I care about consists in describing things correctly. An accurate description does its own corrective work.