Abstract
This paper introduces a fairly recent corpus-based approach to lexical semantics, the Behavioral Profile (BP) approach. After a short review of traditional corpus-based work on lexical semantics and its shortcomings, I explain the logic and methodology of the BP approach and exemplify its application to different lexical relations (polysemy, synonymy, antonymy) in English and Russian with an eye to illustrating how the BP approach allows for the incorporation of different statistical techniques. Finally, I briefly discuss how first experimental approaches that validate the BP method and outline its theoretical commitments and motivations.