Advanced search
Start date
Betweenand

Back to Cratylus

Abstract

This project aims at developing the computational linguistic resources (dictionary and grammar) needed for the automatic generation, from UNL (Universal Networking Language) onto Brazilian Portuguese, of an English version of Cratylus, by Plato. The project is divided into two different steps: 1) human encoding, from English into UNL, of the English version of Cratylus, by Plato, translated by Benjamin Jovett; and 2) the automatic decoding, from UNL onto Brazilian Portuguese, of the UNL version derived from (1). In the former step, the integral text of Cratylus will be manually analyzed and represented according to the 2005 version of the UNL Specifications, i.e., as a hyper graph, where nodes (the so-called Universal Words) will stand for concepts conveyed by English lexical items, which will be linked one another through labeled arcs defining UNL semantic relations. Additionally, nodes will be annotated by UNL attributes conveying pragmatic information. The decoding step, which will be carried out automatically, will require modeling and implementing both a UNL-Portuguese dictionary and a generation grammar, which will follow the UNDL Foundation recommendations, in order to be compiled by the UNL Center existing tools (the Dictionary Builder and the Deconverter). (AU)