BOBRA: An Expert System for Automated Language Description as a Basis for Statistical Studies

Abstract

One of the fundamental shortcomings of computers with regard to language materials is the poor basis of information contained in the strings which they recognize The computer merely recognizes strings of letters In computational linguistics much creativity and effort has been spent on the design of computer programs in order to enrich this poor information delivered by computers. No decisive breakthrough, however, could be made, the enriching of the strings cannot be performed automatically as a processing step before other procedures, such as statistical operations are applied In my opinion there are two main reasons for the failure of program packages for linguistic encoding emerge, firstly the theoretical basis for such packages is lacking. Secondly the programming effort is too high to expect results in a short time span. The second reason being a serious drawback for all progress in computational linguistics, we developed the expert system BOBRA for automated language description In this paper we deal with two subjects in particular firstly, the design and implementation of the BOBRA expert system, which we introduce in connection with our views concerning methods for linguistic computing, especially the automated linguistic description of text. After that, we describe the use of the expert system in several applications