With the globalization o f the w orld markets, the need for m ultilingual inform ation processing increases, because the users are increasingly forced to deal w ith inform ation available in m ultiple languages. D ue to the fact that translation services are qu ite expensive, the m ultilingual society needs smart technologies to access foreign language docum ents at least to a degree which allow s the user to make use o f the inform ation provided by them. These technologies should take into

more »

... ount that a user has often only som e com prehension ability for a given foreign language. R ecent developm ents in Inform ation R etrieval, therefore, try to support users by providing so-called cross-language retrieval techniques. These system s allow the user to query inform ation about a particular subject in their ow n language and retrieving relevant docum ents written in foreign languages. In this paper, the focus is on a cross-language inform ation retrieval com ponent integrated into the m ultilingual inform ation system EMIS which is dedicated to the dom ain o f European M edia Law. This com ponent, MproIR, makes use o f linguistic know ledge provided by a m orpho-syntactic analysis to increase its effectiveness. Based on stem inform ation as well as com positional and derivational know ledge, autom atic query expansion and translation as well as a classification o f the retrieved docum ents is carried out. The languages currently covered by the system are Germ an, English and French. *As the German company ELEXIR reported during System ^, AltaVista.de uses their retrieval system. Unfortunately no further information how the search is improved is publicly available.