Knowing Algorithms

Abstract

Algorithms have become objects of broad concern, expanding beyond the purview of computer scientists to capture the attention of the general public and critical scholars in the humanities and social sciences. This expansion has made manifest a basic knowledge problem: How do we know things about algorithms, especially as this “we” expands beyond the people who design and work with them? This chapter pursues some of the epistemological challenges that face “outsiders” to algorithmic processes. It argues that many of these challenges stem from an overly narrow understanding of what an “algorithm” is, which leads critics to imagine them as simple processes that only need to be revealed to be known. Against this revelatory model, the chapter advances an understanding of algorithms as “algorithmic systems,” heterogeneously composed of computational processes, data flows, people, and a host of infrastructural supports. Thus defined, some of the problems associated with algorithms become easier to identify, and the epistemological challenges that face outsiders are reconfigured. Algorithms are not the source of hidden regularities to be revealed, but rather sociotechnical situations that can be entered into and studied ethnographically.