Seeing Algorithms as Infrastructure: Making Sense of the Reddit Voting System in Narrative and Practice

• In Event: Making Sense of Political Calculations

Abstract

Recent crises in American politics have heightened visibility and concern about the influence of online social media on culture and politics. In order to understand these effects, technologists and scholars attempt to locate the ‘responsibility’ of algorithms in the technical platforms facilitating these social trends. How may we begin to understand the material effects of algorithms, beyond the logic that constitutes an algorithm, and the common narratives used to define and make sense of them?

This paper proposes seeing algorithms as infrastructure as a means to uncover the material effects of algorithms. We use Susan Leigh Stars’ definition of infrastructure to frame our Trace Ethnography of the voting system that is used to curate the content on Reddit. We examine the narratives historically used to make legible the algorithms that constitute this system by both its developers and its users, which characterize the voting system as a fair and transparent model that democratizes the curation of content. However, while the surrounding infrastructure changes, algorithms are themselves distorted as they wrestle with the structures that they’re embedded in—as Reddit scales and changes in structure both technically and socially, the voting algorithms themselves perform differently.

We observe the frictions and crises that arise from the performance of this voting system diverging from its narratives to surface the assumptions that underlie its development and the way it changes materially alongside the rest of the platform.

Authors

• Katherine Lo, University of California, Irvine

• Alex S Taylor, City, University of London