Limitations of Reverso and Google Translate in Rendering Mosuli Implicatures into English

Abstract

The research adopts a qualitative analytical method to examine the generated translations of Google Translate and Reverso. It also aims to identify and classify the implicatures in Mosuli linguistic heritage based on Levinson’s (2000) model of classifying implicatures. In addition, it uses House’s (2015) model to analyze the generated translations and to highlight the limitations of artificial intelligence-generated translations.The results indicate that both applications face difficulties in transferring implicit meanings related to specific cultures. The generated translations tend in most cases to focus on the surface, literal, and explicit meaning or produce new interpretations in the target text that do not fit with the source text, which leads to loss of the communicative function and distortion of the implicit meaning. The research concludes that AI-generated translation, with the significant advances it witnesses, still faces clear challenges in translating Mosuli implicatures, especially when its meaning depends on social context and cultural knowledge. This highlights the need for more advanced and context-aware AI translation applications that are capable of conveying literal and cultural meanings of the source text.