On learning discontinuous dependencies from positive data

This paper is concerned with learning in the model of Gold the Categorial Dependency Grammars (CDG), which express discontin- uous (non-projective) dependencies. We show that rigid and k-valued CDG (without optional and iterative types) are learnable from strings. In fact, we prove that the languages of dependency nets coding rigid CDGs have nite elasticity, and we show a learning algorithm. As a standard corollary, this result leads to the learnability of rigid or k- valued CDGs (without optional and iterative types) from strings.