KURD - A Formalism for Shallow Post Morphological Processing

In most NLP applications an input text undergoes a number of transformations until the desired information can be extracted from it. Typically, such transformations involve part of speech tagging, morphological analysis such as lemmatization or full derivational and compositional analysis, context dependent disambiguation of tagging results, multi-word recognition, shallow, partial or full syntactic parsing, semantic analysis and so on. It is not always evident what level of analysis should be involved. For instance, whether a certain task requires a full parse or whether some `shallow' operations may be su cient is often difcult to determine. The choice of the involved tools can be guided by the data or the requirements or premises of the goal to be reached. These considerations may depend on the availability of a grammatical model, the required standard of the results and processing time constraints. However, the optimization of this task remains an unresolved area until now. The interest of the NLP community for 'shallow' processing has grown recently (cf. (Skut et al., 1997),(Abney, 1996)). In this paper, we describe a simple formalism (KURD) that is designed to perform some `shallow' operations on morphologically analyzed texts. The output can be used directly, or be redirected to further (e.g. linguistic or statistic) processing.