Past and future goals of Computational Linguistics

For different reasons, Computational Linguistics is at the beginning of a new era. One of these reasons is that the technological progress makes it possible to process complex algorithms in real time and to handle large amounts of data in a such way that storage size and fast data access are no longer the bottleneck of language processing. The technological progress makes it possible to recognize the regularities of language production and language reception from the statistical probability of their occurences. A second reason is that Computational Linguistics is no longer restricted only to the channel of orthographical language. The restrictions of this channel have dominated research for more than 40 years. We now are able to use parameters of spoken language like word accent and intonation patterns for the description and disambiguation of sentence and discourse structures. In the near future the visual channel will be available for transcribing gestures and face expressions and to use them for structural analysis. From a third point of view Computational Linguistics becomes more and more interesting for commercially applications. It is not only the old idea to overcome the language barriers through mechanical translation, but also the growing needs of our society, to get information as fast and precise as possible, which require new products to decode text and speech (e.g., speech recognition, speech synthesis and information extraction). On this general background this paper discusses the position of Computational Linguistics between academic research and technological applications. On the basis of a detailed analysis of the history and current situation of Computational Linguistics future goals and trends of academic research are to be outlined. Especially the progress of computer oriented discourse analysis and of multimodal research methods will be mentioned.