The Development of an Advanced SLP-based System for the Individual Learning and Fast Training of Speaking Skills in a New Foreign Language

This paper describes the organization of a concerted action of German and Chinese phoneticians, IT speech technologists and specialists in the field of foreign language teaching in order to create a new research program which has been proposed by the first author as a so-called graduate college (entitled “Chinesisch fr deutsche Ingenieure und andere L1-Sprecher des Deutschen”) to the Federal Government of Germany. The paper has four parts: We first start by describing the prime goals and conditions of our application-oriented basic research plans for the development of a new SL/T-system (Speech Learning/Training-system). Then we argue why — in phonetic speech research — the classic paradigm of “analysis-by-synthesis” should be replaced by a the programmatic approach which follows the new “synthesis-by-analysis”-principle. In the third part we look at certain types of teaching speech acts and teaching dialogs. In the last part we will be trying to consider the fact that such complex approaches of application-oriented basic speech research programs as in our case needs a concerted effort in the collaboration of all relevant disciplines within the field of SLP (including also NLP1), and here we end with the question of how the so-called dilemma of text-tospeech developers can be systematically avoided by practically involving the addressed users of the developed System (engineers and language students) from the very beginning. As described by H. Fujisaki in his introduction to the First International Conference on Spoken Language Processing (ICSLP) in Kobe, 1990.