• Summary
• Related Documents
  • Active Bibliography
  • Co-citation
• Version History

Abstract:

Synchronous TAGs are a variant of TAGs introduced by [Sheiber and Schabes, 1990] to characterize correspondences between tree adjoining languages. They can be used for relating TAGs for two different languages for the purpose of machine translation [Abeille et al., 1990], or for relating a syntactic TAG and a semantic one for the same language [Sheiber and Schabes, 1990, Abeille, 1992] for the purpose of generation [Sheiber and Schabes, 1991] or semantic analysis. For example, consider the synchronous transfer between grammars of English and French. Although the approach is not directional, call English the source and French the target language. The transfer lexicon puts into correspondence a tree from the source grammar instantiated by lexical insertion (all its nodes and their attributes) with a tree from a target grammar. First the source sentence is parsed according to the source grammar. Each elementary tree in the source derivation tree is then mapped to a tree in the target derivation tree by looking in the transfer lexicon. The target sentence is then read off the target derivation tree. Consider the fragment of the transfer lexicon given in Figure 1. The transfer lexicon consists of pairs of trees one from the source language and one from the target language. Within the pairs of trees, nodes are linked. Adjunction or substitution is generalized from occuring on one node in a tree to pairs of nodes in tree pairs which are linked together. Suppose we start with the pair / and we operate the pair ff on the link from the English node NP 0 to the French node NP 1. This operation yields the derived pair ff 1. ff 1

Citations