Standards Development

Scope

This project describes the Syntax and Semantics (SynSem) module of the Lexical Markup Framework (LMF), a metamodel for representing data in monolingual and multilingual lexical databases used with computer applications. The SynSem module allows for the description of specific syntactic and semantic properties of lexemes, as well as the complex interactions between them. More specifically, the syntax part of the module describes the properties of a lexeme when combined with other lexemes in a sentence. When recorded in a lexicon, these properties make up the syntactic description of a Lexical Entry instance. The semantics part of the module, on the other hand, describes the sense of a lexeme and its relationship with other senses belonging to the same language. The SynSem interface describes the predicates, and the mapping between syntactic and semantic arguments. This serialization covers the classes of ISO 24613-1 (the LMF core model), ISO 24613-2 (the machine readable dictionary, MRD, model) and ISO 24613-4 (TEI serialization). 

Purpose

The ISO 24613 (LMF) multi-part standard is based upon the definition of an implementationindependent meta-model combining a core model and additional modules accounting for the variety of forms that onomasiological lexical content may take.

Standardising of the representation syntactic and semantic properties of lexical entries, as well as their mapping will allow for important lexical information to be correctly encoded. This involves the relations between senses - such as the ones encoded in well known resources, for instance the SIMPLE lexicons (Lenci et al 2000), semantic predicates and sub categorization frames (for instance those in Framenet and Verbnet) as well as mapping between syntactic and semantic arguments.

The modelling of syntax and semantics had already been thoroughly described in ISO 24613:2008, which contained two normative annexes (Annex E and Annex G) for syntax and semantics respectively. In its core, the ISO 24613:2008 modelling of syntax and semantics can be considered as being still valid, has been used to model lexical resources, and was the inspiration for other initiatives, such as the W3C Ontolex-Lemon model.

The new ISO 24613-6 SynSem module will update the old modules where necessary, to ensure the compatibility with ISO 24613:1-5 as well as backward compatibility with the old standard and with more recent initiatives.

From an industrial point of view, the syntax and semantic module of LMF can be particularly useful to represent domain ontologies and terminologies which contain semantic predicates and or automatically extracted subcategorization frames (see Caselli et al 2012 for the environment domain).