| Sumario: | Electronic dictionaries are now the primary resource for second-language (L2) learners, providing effective support for vocabulary acquisition and lexical development. It is commonly presumed by users that entries in e-dictionaries are accurate, often without critical evaluation of their authenticity or reliability. The content of electronic dictionaries is seldom scrutinized for accuracy, and the underlying corpus design used in their development has received little scholarly attention. Moreover, research on dictionary usage remains limited, particularly regarding challenges unique to e-dictionaries. This research adopts a mixed-methods approach that integrates experimental testing, user surveys, and critical analysis of dictionary content to examine Youdao, the most widely used electronic dictionary in China. The experimental components integrate retrospective reporting with a controlled translation task to evaluate the influence of an electronic dictionary on student learning outcomes. More than 100 students from various academic disciplines participated in a translation task featuring sentences containing terms with inadequate definitions in the Youdao Dictionary. The consultation process was recorded to assess how such inaccuracies affected participants' comprehension of the target vocabulary. The findings suggest that electronic dictionaries containing incomplete or inaccurate entries may lead to sentence misinterpretation during translation tasks. Moreover, ineffective consultation behaviors, such as excessive reliance on the concise dictionary section, were found to further aggravate these challenges. The study attributes inaccurate definitions to the limitations in current artificial intelligence (AI) and machine learning technologies used in dictionaries. To mitigate these challenges, it is recommended that L2 learners receive targeted training to develop more critical dictionary consultation skills, while online dictionaries should incorporate more advanced AI technologies for greater accuracy and usability.
|
|---|