Adjectives: Constructions vs. Valence
Abstract
The paper approaches adjectives in French and Polish from two perspectives: a linguistic description and an automatic text analysis. In particular, we aim at specifying adjective valence and distinguish it from components with which they occasionally occur in various syntactic constructions. Then, we apply linguistic knowledge to annotated data and automatically extract valence lexicons for adjectives. For French, a richly annotated treebank is available whereas the Polish corpus we use currently contains only morphosyntactic information. The paper focuses on results obtained for French as valence extraction for Polish requires additional data processing.