All philosophy is ethnographic

7 min readDec 2, 2023
Photo by Pixabay:

If you ever get the chance to read Origins of Totalitarianism one of the first things that strikes you is how ethnographic Arendt’s writing is. Her core philosophy is deeply rooted in her lived experiences, an exile forced out by hate. Thomas Hobbes’ writing is rooted in the same exiled experience, though he is diametrically opposite to Arendt in preferring authoritarian order over liberal freedom. Plato, Descartes, Wittgenstein, Foucault, Anscombe, indeed any philosopher you care to name root their philosophy in their personal perception of the world, it is inescapable. For every attempt at objectivity ultimately it is the subjectivity of their map of the world that directs them. I find this fascinating, not at all a problem, rather something you must click into in order to appreciate their writing.

My research is rooted in attempting to square rights and perception, hence my joy at reading Wittgenstein and Anscombe. In the age of AI and machine learning I believe it is imperative that we interrogate the assumptions we make about philosophy, especially moral philosophy, as the objectivity we assume is baked into those algorithms is the ethnographic experiences of the coders. Indeed, in the last decade there has been an upswell of concern over whose perspective we encode into the normative understanding of the world. Meredith Whitaker, Catherine D’Ignazio, and Laruen Klein amongst others have called for a deeper appreciation for who controls data, with Cathy O’Neil bringing this message to the masses in Weapons of Math Destruction.

It is not enough to be mindful of others’ perspectives, if we are coding them into systems we need to explicitly acknowledge the ethnographies present. Roland Barthes stressed the need to kill the author and project your own understanding onto a text, yet if you only rely on your own subjectivity you fail to appreciate the wider context any given text or information sits in. All facts depend on your own subjectivity to interpret and contextualise, something which machine learning relies on coders’ own perception in mapping out the world. This matters, because unlike a newspaper, film, or television show it is practically impossible to trace all the threads that lead to an output.

Plato’s cave was imagined in a time when actual caves were still in use for hermits and dwellings. Much of the…