Translation Evaluation - a State of the Art Survey

Measuring translation quality is a subjective process which relies on human judgements. This paper discusses different frameworks used in the process of translation evaluation with special focus on error classification schemes used both in the translation industry and in translation teaching institutions. Such error-based models, like BlackJack or SAE J 2450, allow human judgements to benefit from a consistent and systematic error classification. Apart from offering an overview of the existing error-based translation evaluation systems, this paper describes two recent projects which further the research in this field. This paper does not claim to provide an exhaustive list of existing error-based translation evaluation tools nor does it aim to impose or offer a particular model for translation analysis.