Modeling information structure in a cross-linguistic perspective

This study makes substantial contributions to both the theoretical and computational treatment of information structure, with a specific focus on creating natural language processing applications such as multilingual machine translation systems. The present study first provides crosslinguistic findings in regards to information structure meanings and markings. Building upon such findings, the current model represents information structure within the HPSG/MRS framework using Individual Constraints. The primary goal of the present study is to create a multilingual grammar model of information structure for the LinGO Grammar Matrix system. The present study explores the construction of a grammar library for creating customized grammar incorporating information structure and illustrates how the information structure-based model improves performance of transfer-based machine translation.