Situating Word Senses in their Historical Context with Linked Data

In this article we present a Semantic Web-based model for creating lexical resources in which the diachronic and, more broadly, contextual dimensions of word meaning can be explicitly represented as part of a graph-based data structure. We start by discussing why Linked Data is the right publishing approach for such diachronic datasets. We then describe our model, lemonEty, which utilizes the ontology engineering technique of perdurants in order to model lexical entries as dynamic processes. Next we go onto explain how to represent etymologies using our model, and in particular how to associate temporal information with word senses, taking examples from two different lexicographic resources. In addition, we will show how our model deals with cognates and attestations.