Machine vs Human: translating metaphor in The Picture of Dorian Gray

Abstract

This project examines how metaphors in The Picture of Doran Gray are translated by machine translation (MT) and human translators (HTs), by comparing solutions provided by Google Translate with existing published translations into Spanish, Portuguese and Polish. The aim of this study is to evaluate the performance of MT by comparing the choices made by GT with those made by HTs in order to see how and to what extent they differ. Five types of metaphor are considered in the experiment: lexical, multi-word, extended, idiomatic and dead. By applying these categories, the experiment investigates whether the quality of MT output is affected by the length of metaphor and the frequency with which it appears in everyday language. Moreover, this study also analyses the translation procedures used by both categories of translator, in order to observe patterns of behaviour, including commonalities and differences. As two Romance and one Slavic language are included in the experiment, it also asks whether similar patterns emerge across languages belonging to the same family. The hypothesis is that metaphors using relatively common collocations will be rendered literally by both MT and HTs, but the procedures will diverge in cases of metaphors that use more unusual and/or culture-specific imagery and concepts. The prediction is that HTs will replace these with alternative metaphors more often than MT, which will most likely only be able to translate them literally.