Semantic and Grammatical Issues in Translating Idioms with Automatic Translation Systems

Idiomatic expressions are not only a part of phraseology; they are also the linguistic and cultural wealth of every nation. They reflect the culture and perspectives of nations. Translating idioms from one language (L1) to another (L2) and finding their equivalents are important not only in Natural Language Processing (NLP) but also in translation studies, including literary translation, synchronous and diachronic translation, and even text expertise in cognitive linguistics. Studying idioms and finding their equivalents in L2 has a great importance in machine translation of NLP and creating parallel corpora. It is impossible to translate idiomatic expressions literally using bilingual dictionaries. A number of translation programs and systems are currently being used for automatic translation. But the translation of the phrases in the text is still wrong or poorly translated. To overcome this problem, it is important to create a database of equivalent translations of phrases from L1 to L2. This article analyzes the results of automatic translation of idioms using translation systems, provides solutions to the problems identified, and underscores the need for modeling idioms for automatic morphological translation. Additionally, the issue of translating idioms using automatic translation services including Google translate and Yandex translate as well as artificial intelligence ChatGPT is studied in detail.