In October, I presented at the Internet Research 13.0 Conference on algorithms and networks. My paper, “Creating Networks Through Search: PageRank, Algorithmic Truth, and Tracing the Web”, was included in the conference proceedings, Selected Papers in Internet Research, and is now available for reading. The abstract is below.
This paper analyzes PageRank, a key feature of Google’s search algorithm, showing how its primary function is not to identify quality Web pages but rather to identify hubs within a network defined by the Internet’s link structure. While PageRank’s method has been compared to the process of using the wisdom of crowds to determine quality, by relying on network effects to identify hubs, the algorithm does not allow users the independence and diversity necessary for crowdsourcing to be completely effective. For these reasons, Google and other search engines cannot be simply understood as information providers, for their role in defining the network structure of the Web makes these search companies the holders of a significant form of network power: programming. However, users can offset this power by becoming switchers who actively connect networks in order to diversify their information sources.