Colloquium: Erik J. Olsson (Lund)
Department: PhilosophyDate and Time: October 6, 2017 | 3:00 PM-5:00 PM
Event Location: HIB 55
Title: "The Google Theorems: Applying Epistemology to the Web"
Search engines like Google play a huge role in society as sources of information. Determining what information we find online they influence and shape our view of the world. And yet, strangely enough, philosophers (including epistemologists) have had very little to say on the epistemological and societal effects of relying on search engines.
What webpages we look at is often determined by what comes on top of Google’s result list. The result list is determined in part by what webpages people link to from their own webpages, the so-called link structure of the web. Ceteris paribus, the webpages that receive more links are considered more important and ranked higher. (There are a lot of other things going on as well in Google, but we focus on linking here.) This idea, which is an appeal to the so-called “wisdom of crowds”, derives from the original technical report by the Google founders and associates that laid the foundation for one of the world’s most successful enterprises (Brin et al, 1998).
Hence, in order to understand societal and epistemological effects of search engines we need to understand the link structure and its significance. Again, epistemology is largely silent on this crucial topic, or was silent until just recently (Masterton, Olsson, and Angere, 2016; Masterton and Olsson, 2017).
A webmaster’s decision to link to a webpage can be interpreted as a “vote” for that webpage. But how far does the parallel between linking and voting extend? In the talk I will provide several theorems showing that link-based ranking (including ranking by Google’s PageRank algorithm) tracks “importance” on the web in the limit as the number of webpages grows, given linking independence and minimal linking competence. The theorems, which will be explained on a philosophical level, are similar in spirit to the voting, or jury, theorem famously attributed to the 18th century mathematician Nicolas de Condorcet.
Finally, I argue that these “Google theorems” provide a fundamental epistemological justification for link-based ranking on the web, analogous to the justification that Condorcet’s theorems bestow on majority voting as a basic democratic procedure. In both cases there are also important limitation. To deal with one such limitation, I will also quickly look at various ways of incorporating “group think” into the framework, i.e. a bias for linking to what other people link to and not only to what is important.
Brin S., Page L., Motwami R., and Winograd T. (1998), “The PageRank Citation Ranking: Bringing Order to the Web, Stanford University Technical Report.
Masterton, G., Olsson, E. J., and Angere, S. (2016), “Linking as Voting: How the Condorcet Jury Theorem in Political Science is Relevant to Webometrics”, Scientometrics, March 2016, Volume 106, Issue 3, pp. 945-966.
Masterton, G., and Olsson, E. J. (2017), “From Impact to Importance : the Current State of the Wisdom-of-Crowds Justification of Link-based Ranking Algorithms”, Philosophy and Technology, published online: https://link.springer.com/article/10.1007/s13347-017-0274-2