Lexicalized Proof Nets in Pomset Logic and TAGs
First introduced by Retore (1993), pomset linear logic can deal with linguistic aspects by inducing a partial
order on words. Lecomte & Retore (1995) use this property: they define modules (or partial proof-nets) which
consist in entries for words, describing both the category of the word and its behavior when interacting with
other words. Then the natural question of comparing the generative power of such grammars with Tree
Adjoining Grammars (Josji, Levy, Takahashi 1975),as Joshi & Kulick (1996) pointed some links out, arises.
To answer this question, we propose a logical formalization of TAGs in the framework of linear logic proof-nets.
We aim to model trees and operations on these trees with a restricted part of proof-nets, and we show how this
kind of proof-nets expresses equivalently TAG-trees.
Proc. LACL 98, Grenoble, France. LNCS/LNAI vol. 2014.
pogodalla-lacl98.pdf (182.56 kB)