The search of Web resources is a very important topic due to the huge amount of valuable information available in the WWW. Standard search engines can be a great help but they are often based only on the presence or absence of keywords. Thus problems regarding semantic ambiguity appear. In order to solve one of them, we propose a new method for discovering lexicalizations and synonyms of search queries based on a previously obtained taxonomy of terms for the specified domain. Keywords. Web mining, web search, knowledge acquisition, information extraction