Steps to an Ecology of Algorithms

Abstract

Anthropological expeditions seeking out algorithms frequently return empty-handed. They are confronted with the challenge of the object: what to study when studying algorithms? In this article, I draw together a number of literatures to outline one possible answer to the question of how to study algorithms in social science. I argue that what we should study are algorithmic ecologies. I sketch five modalities of algorithmic ecologies and review concomitant literatures: (a) imaginaries, (b) infrastructures, (c) interfaces, (d) identities, and (e) investments and interests. The speculative propositions offered here are that algorithms are immanent to ecologies and that they are enacted across all the modalities of algorithmic ecologies.