Two-Dimensional Centrality of a Social Network

A procedure of deriving the centrality in a social network is presented. The procedure uses the characteristic values and the vectors of a matrix of friendship relationships among actors. While the centrality of an actor has been usually derived by the characteristic vector corresponding to the largest characteristic value, the present study uses not only the characteristic vector corresponding to the largest characteristic value but also that corresponding to the second largest characteristic value. Each actor has two centralities. The interpretation of two centralities, and the comparison with the additive clustering are presented.