In this paper we describe the general architecture and the main structuring relations of the Portuguese WordNet, under construction within the WordNet.PT <a name=TX></a>project<a href="#NOTA">*</a>. We also present empirical and technical motivations for information encoding, focusing on generic criteria concerning the expression of polysemy. For an overview of the first results of this project we present some statistical data, which show the representativeness of the major categories and relations.