In 1998, ELSNET started a small-scale resource-building exercise, to produce a mini-corpus of sentences annotated at both syntactic and semantic level. The resource produced is a pair of small sample corpora of parallel structure for German and Italian, about 1000 sentences of each language, illustrating 20 verbs, and their syntactic and semantic subcategorization. The annotation concentrates on the verbal predicates and their subcategorized complements, as well as on a few relevant modifiers.
This resource is to be seen, among others, as an example of a corpus annotated at several levels of linguistic description, including word classes, syntax and lexical semantics. As it is the case with all experimental resources produced by the ELSNET Resources Task Group, the corpora are intended to be an example of the kind of resources that may, in the medium term, be produced in a wider framework. Thus, the production methodology is seen as equally important as the data outcome.