Morphology is a key component for many Language Technology applications. However, morphological relations, especially those relying on the derivation and compounding processes, are often addressed in a superficial manner. In this article, we focus on assessing the relevance of deep and motivated morphological knowledge in Natural Language Processing applications. We first describe an annotation experiment whose goal is to evaluate the role of morphology for one task, namely Question Answering (QA). We then highlight the kind of linguistic knowledge that is necessary for this particular task and propose a qualitative analysis of morphological phenomena in order to identify the morphological processes that are most relevant. Based on this study, we perform an intrinsic evaluation of existing tools and resources for French morphology, in order to quantify their coverage. Our conclusions provide helpful insights for using and building appropriate morphological resources and tools that could have a significant impact on the application performance.
Examining the role of statistical and linguistic knowledge sources in a general-knowledge question-answering system
