Quality in human and machine translation

Zusammenfassung

There have always been long-winded discussions on the role played by both human and MT in quality translation processes. Which one is better? Or, should they be used in combination to achieve a quality translation? The present paper provides an answer to these matters by means of the calculation of several evaluation metrics to study the quality offered by MT compared to human translation. Moreover, there is a implementation of a new tool based upon a reference model text with some indexes including Narrativity, Readability, Referential Cohesion, Deep Cohesion, and Concreteness, which is compared to the translated texts produced by humans. To calculate the evaluation metrics and indexes, chosen samples of scientific and literary texts were included. Mentioned texts were used in two final dissertations in the university course of Translation and Interpreting at the University of Murcia.