Driven by the popularity of social networks, there has been an increasing interest in employing such networks in the context of named entity linking. In this paper, we present a novel approach to person name disambiguation and linking that uses a large-scale social network extracted from the English Wikipedia. First, possible candidate matches for an ambiguous person name are determined. With each candidate match, a network substructure is associated. Based on the similarity between these network substructures and the latent network of an ambiguous person name in a document, we propose an efficient ranking method to resolve the ambiguity. We demonstrate the effectiveness of our approach, resulting in an overall precision of over 96% for disambiguating person names and linking them to real world entities.