Formalizing MultiWords as Catenae in a Treebank and in a Lexicon

The paper presents formalization of multiwords as catenae in a treebank and in a lexicon. We view catenae as a dependency subtree, which reflects non-constituents and non-standard dependencies. Since the multiword classifications vary to great extent, starting from very narrow ones and proliferating to extended ones which include also valences, the focus in the paper is not on the multiword typology per se, but on the general formalization of multi-words.