Rule-Based Semantic Tagging. An Application Undergoing Dictionary Glosses

Abstract: The project presented in this article aims to formalize criteria and procedures in order to extract semantic information from parsed dictionary glosses. The actual purpose of the project is the generation of a semantic network (nearly an ontology) issued from a monolingual Italian dictionary, through unsupervised procedures. Since the project involves rule-based Parsing, Semantic Tagging and Word Sense Disambiguation techniques, its outcomes may find an interest also beyond this immediate intent. The cooperation of both syntactic and semantic features in meaning construction are investigated, and procedures which allows a translation of syntactic dependencies in semantic relations are discussed. The procedures that rise from this project can be applied also to other text types than dictionary glosses, as they convert the output of a parsing process into a semantic representation. In addition some mechanism are sketched that may lead to a kind of procedural semantics, through which multiple paraphrases of an given expression can be generated. Which means that these techniques may find an application also in 'query expansion' strategies, interesting Information Retrieval, Search Engines and Question Answering Systems.