Scale-free topology of the interlanguage links in Wikipedia

Abstract

The interlanguage links in Wikipedia connect pages on the same subject written in different languages. In theory, each connected component should be a clique and cover one topic. However, incoherent edits and obvious mistakes result in topic coalescence, yielding a non-trivial topology that is studied in this paper. We show that the component size distribution obeys the power law, and we explain anomalies in the distribution as results of certain edit conventions. Next, we propose a method of filtering out the cliques and study basic properties of the resulting skeleton, which turns out to be scale-free.