A Study on the Cross-cultural Conversion Mechanism and Compensation Strategies of Conceptual Metaphor in the English Translation of Chinese Classical Literature

Abstract

This paper proposes an end-to-end metaphorical vocabulary recognition model named CBAM, which combines character-based and vocabulary-based representation methods to automatically obtain lexical information. The BiLSTM network is used to capture the context semantics of the target word, and the multi-head attention mechanism is introduced to enhance the model's perception of metaphors. The similarity network and Softmax classifier are used to determine the metaphorical nature of the vocabulary. In the translation task, an English-to-English neural machine translation method for metaphorical vocabulary is proposed, integrating the recognized metaphorical vocabulary information into the model's encoder to help the model learn deep semantic information. Through the gating mechanism, the translation information of the vocabulary is controlled for input. Experimental results show that the metaphorical vocabulary recognition model outperforms the best comparison model in all parts of speech and in the case of only verbs, with an F-score increase of 5.0% and 5.5% respectively. The classical literature English-to-English neural machine translation model achieved a BLEU4 evaluation index value of 45.78%. This paper further summarizes the compensation strategies for the conceptual metaphors in the translation of Chinese classical literature, providing methodological references for the cross-language dissemination of classical literature.