Capturing the comparable : a system for querying comparable text corpora

We discuss the importance of bilingual and multilingual text corpora in many types of cross language investigations and illustrate the differences between parallel and comparable text archives. The advantages of comparable over parallel data for certain kinds of contrastive linguistic studies arc outlined. A prototype version of a system for querying comparable text archives is then described and examples nf the first results arc given. The system will form part of an integrated works talion for mono- and bilingual lexical and text database management and interrogation under development at the Istituto di Linguistica Computazionole, Pisa.