Assessing the Annotation Consistency of the Universal Dependencies Corpora

Abstract

A fundamental issue in annotation efforts is to ensure that the same phenomena within and across corpora are annotated consistently. To date, there has not been a clear and obvious way to ensure annotation consistency of dependency corpora. Here, we revisit the method of Boyd et al. (2008) to flag inconsistencies in dependency corpora, and evaluate it on three languages with varying degrees of morphology (English, French, and Finnish UD v2). We show that the method is very efficient in finding errors in the annotations. We also build an annotation tool, which we will make available, that helps to streamline the manual annotation required by the method.