Proceedings of the 1987 workshop on Theoretical issues in natural language processing

Natural Language Processing is now an industrial process and product as well as a form of theoretical enquiry. Indeed, with the industrial growth of Artificial Intelligence and Expert Systems there is now a danger of difficult and resistant theoretical problems being ignored or pushed aside while products are generated and sold. It is therefore very timely for those in theoretical research in the area of natural language processing and understanding to stand back and survey the partial solutions available to problems in this field, the blind alleys that have been closed off, those that had been closed for many years but have now been declared reopened, and the new theories and representations that have been developed and published since the last such inventory was taken.The role of Tinlap workshops, then, is not at all to set out and discuss particular applications and implementations of natural language processing systems, but to concentrate on the underlying issues, and to compare solutions and constraints across disciplinary borderlines by drawing into a workshop theoreticians in artificial intelligence, logic, psychology, philosophy and linguistics. The previous Tinlap Workshops were able to crystallize, in the published statements of position from panellists, particular discussions and disputes within the field; and a measure of this is the degree to which those published volumes have continued to sell and to be referenced in the general literature of the field.The main aim of Tinlap3 is again to crystallize the state of the field at the present moment, after a period which has seen a revival of interest in syntactic approaches to analysis; a new and pressing concern with aspects of parallel processing and their role in language understanding; a shift in the aspects of semantics and pragmatics that are of most concern to researchers, as well as an underlying worry that the field has not progressed in the linear fashion that was hoped for at the time of Tinlap1.