ABSTRACT
In this paper we propose a system allowing natural language queries for data in a structured archive. The system not only reacts appropriately to spoken remarks, but also is minimally restrictive on the users language customs in order to meet real world requirements. Thus, it accepts specifics of human communication on several levels, such as indirect speech acts, non-grammatical morphological and syntactical phenomena. Two very important features of our system are its flexibility to be modified to work on another natural language and to be ported to different data archives with a minimum of effort. The system currently runs on a lap top in two languages, English and German.
