Until now, the contribution of world knowledge to the process of pronominal anaphora resolution has not received a thorough computational investigation, mostly due to the lack of a large scale implemented model of world knowledge. This paper proposes Latent Semantic Analysis (LSA) as such a model: word meaning representation it constructs can be used to rank potential antecedents according to how well they fit in the pronoun’s context. The initial results of incorporating LSA into a pronominal anaphora resolution algorithm are encouraging.