A database of German definitory contexts from selected web sources

A database of German definitory contexts from selected web sources Adrien Barbaresi, Lothar Lemnitzer, Alexander Geyken Berlin-Brandenburg Academy of Sciences J¨agerstr. 22/23 – 10117 Berlin – Germany {barbaresi,lemnitzer,geyken}@bbaw.de Abstract We introduce work on the detection of definitory contexts designed to speed up two lexicographical tasks: searching for the exact meaning(s) of terms and providing usable input for paraphrasing. Our database is built from a specialized web corpus using a robust pattern-based extraction method. The corresponding interface displays information for a large range of lexical units. The contributions of this article are threefold: we describe both acquisition and extraction, provide a qualitative assessment of the method, and present an interface to access the data. Keywords: Definition extraction, definitory contexts, computational lexicography, resource acquisition 1. Introduction German is well-known for making extensive use of com- pounds (Schl¨ucker, 2012), as composition is a productive process of German word formation. In the context of lexical resources and dictionaries, the number of potential words to define is very large if not infinite. For that matter, cor- pus data is crucial in order to perform lexicographic work. Furthermore, automatic definition extraction can be used to complement existing entries or help lexicographers creat- ing new ones. This article introduces both a corpus and a database built to support lexicographers as well as in- terested users in sorting out the meaning(s) of lexemes. The acquisition draws on selected web sources containing a large number of lexical descriptions, which are expected to contain potential definitory contexts, from which hand- written definitions will be derived. We describe an exper- imental setting and front-end for pattern-based definition extraction from such resources. We first describe the context and motivations of our work. In section 2 we summarize related work and place our con- tribution into this context. In section 3 we present our ac- quisition and extraction method. Section 4 describes the database, an interface to it as well as the results of a quali- tative evaluation. In section 5 we conclude with the insights gained from our investigation. 1.1. Context and motivation The Digital Dictionary of the German Language (Digi- tales W¨orterbuch der deutschen Sprache, DWDS1) is a long term project of the Berlin-Brandenburg Academy of Sci- ences (BBAW). The goal of the DWDS project (Klein and Geyken, 2010) is to create a large-scale aggregated word information system based on legacy dictionaries, large cor- pora, word statistics and automated methods to provide ad- ditional types of linguistic information as well as to speed up the process of updating and amending the existing lexi- cal resources. The DWDS web platform provides access to this infor- mation and is among the most frequently used academic dictionary platforms. The dictionary component of the DWDS is based on high-coverage and detailed dictionar- 1https://www.dwds.de/ ies of contemporary German, including the W¨orterbuch der deutschen Gegenwartssprache and Duden Großw¨orterbuch der deutschen Sprache. Thus, the DWDS dictionary com- ponent with more than 170,000 entries provides a good cov- erage of the core vocabulary of German. Having covered this core …

Content truncated to 500 words