Behavioral vs. Verbal Methods in Translation Quality Evaluation: A Cognitive Experimental Study

This study explores the sensitivity differences between behavioral experiments and verbal reports in translation quality evaluation. Results indicate that behavioral metrics (e.g., response times) are significantly more sensitive to syntactic-pragmatic manipulations (phrase order) than verbal reports. Translations with congruent phrase order received higher ratings and faster response times compared to those with incongruent order. However, most participants explicitly denied phrase order's influence in verbal reports. Lexical equivalence showed no significant impact on explicit ratings but increased cognitive effort, as indicated by slower response times for approximate lexical matches. These findings reveal a critical dissociation between implicit cognitive processes and explicit awareness in translation evaluation. The study highlights that translation assessment involves both implicit System 1 processes and explicit System 2 reasoning, offering new cognitive insights for translation research and practical implications for translator education and machine translation assessment. By bridging cognitive science and translation studies, this research contributes to a paradigm shift: translation quality is not merely what evaluators say it is, but what their cognitive behavior reveals it to be.