Rule-Based Lexical Modelling for Translating Proverbs in Kashmiri and Kannada to English: Corpus Creation and Analysis

Kashmiri and Kannada are two linguistically rich but underrepresented Indian languages, each with deep cultural heritage expressed through proverbs. This paper presents a rule-based system for the identification and translation of proverbs from Kashmiri and Kannada into English. By leveraging a self-constructed bilingual proverb database, the system detects proverbs from user-provided text inputs and generates accurate English translations, supported by transliteration and lexical mapping. By focusing on multi-word expressions (MWEs), this system addresses the challenges in which literal translations often fail to preserve meaning. This approach demonstrates improved contextual fidelity in proverb translation and contributes valuable linguistic resources to low-resource languages.