LVF-lemon ― Towards a Linked Data Representation of “Les Verbes français”

LVF-lemon – Towards a Linked Data Representation of "Les Verbes français" Ingrid Falk, Achim Stein Institut für Linguistik/Romanistik, Universität Stuttgart first.second@ling.uni-stuttgart.de Abstract In this study we elaborate a road map for the conversion of a traditional lexical syntactico-semantic resource for French into a linguistic linked open data (LLOD) model. Our approach uses current best-practices and the analyses of earlier similar undertakings (lemonUBY and PDEV-lemon) to tease out the most appropriate representation for our resource. Keywords: linguistic linked open data, LLOD, French valence lexicon, lemon, pdev, uby 1. Introduction In this paper we investigate how a traditional lexical syntactico-semantic resource, namely "Les Verbes Français" (The French Verbs, henceforth abbreviated LVF) (Dubois and Dubois-Charlier, 1997; François et al., 2007), can best be converted into a standardised and normalised linked open data model. Our motivation is twofold. On the one hand we aim to explain and make more accessible the encoded linguistic knowledge. On the other we plan to make LVF more interoperable with and comparable to other linguistic resources, as for example corpus data and subcategorisation lexicons. To convert LVF to LLOD format the following steps have to be undertaken: 1. RDF conversion 2. Data modelling: the content needs to be modelled in terms of well established vocabularies. These may be general vocabularies like RDFS, OWL, SKOS, lin- guistic vocabularies (lemon, LexInfo, OLiA, LMF), or finally vocabularies specific to LVF. 3. Linking the data. While the first step is rather straightforward in our case, the second and third step require substantially greater re- search effort due to the traditional way this resource has been designed and developed. A major strength of LVF lies in its syntactic and semantic description. We will therefore primarily investigate how this information and the intricate syntax-semantics interface can best be represented within a linked data framework. 2. The LVF resource “Les Verbes Français” (Dubois and Dubois-Charlier, 1997; François et al., 2007) is a detailed and extensive lexical resource providing a systematic description of the morpho- syntactic and syntactico-semantic properties of French verbs. The basic lexical unit are readings of the verbs, determined by their acceptable syntactic and semantic context. The LVF covers roughly 12 300 verbs with a total of 25 610 usages (readings). Each reading is associated with an elabo- rate morphologic, syntactic and semantic description. In the following we use sample entries for the verb élargir (enlarge) to illustrate LVF’s basic layout and give an idea of the underlying lexicographic and representation choices, in particular with respect to the syntactic and semantic descrip- tion. LVF lists four readings for élargir, illustrated by the sample usages shown in Table 1. Semantic description. Each reading is characterised by a semi-formal semantic description, called opérateur, which is meant to represent its meaning, e.g. r/d+qt large for élargir 01 (cf. Table 1a). The meaning of élargir 01 is thus associated with the semantic primitive r/d (‘render/become’ indicating transformation). This way the readings were as- signed one of a finite set of prototypical semantic predicates which, in a subsequent step, …

Content truncated to 500 words