Assessing semantic equivalence for shared understanding of international documents: An application of semantic network analysis to multilingual translations of the Universal Declaration of Human Rights

Abstract

The purposes of this thesis are: (1) to examine if the translation issue influences international disagreement surrounding the UDHR and (2) to suggest the objective methodology to assess the translation quality. Among two different factors determining translation quality, the technical factor and the language factor, the study focuses on how the language factor, which is inherent in the culture, influences translation equivalence. The study conducted semantic network analysis based on the matrices of the forty most frequently appeared words and of the eighteen commonly appeared words. The quadratic assignment procedure analysis and the Spearman's rank order correlation resulted in high correlations throughout the seven translations. The high correlations imply that the translations are roughly equivalent enough to reach the shared understanding of the documents. However, the centrality measures revealed that the usage of some concepts were different across languages. The concepts showing different centralities reflect the each language's cultural predisposition. By showing how centrality measure can present the subtle differences across the translations, the study demonstrates the usefulness of semantic network analysis as the methodology to examine the language issue influencing translation quality. The study discusses further the importance of exploration of translation and language issue as the extension of the intercultural communication studies.