Identification of Legislative Errors

We present an approach designed to support the process of legislative drafting by helping to detect errors in a normative text. It is based on a framework allowing for representation and comparison of structure and semantic content of legal provisions. Such comparison serves as a starting point for detection of (potential) legislative errors. The approach provides in particular criteria to select provisions to be compared, related to the phenomenon of provisions overlapping. We show that specific cases of such an overlap may amount to legislative errors. The presented framework enables a precise and transparent account of these errors. We also acknowledge that textual provisions enable various interpretations, while the error methodology detection assumes that the semantic representation of provisions is a result of a specific interpretation. We introduce the notion of Constraining Interpretive Rules which are used to evaluate the acceptability of specific interpretations of legal provisions. We discuss the features of the model on a real example and we present an implementation of the approach by using semantic technologies.