Marathi WordNet (MWN)
Marathi WordNet is a digital lexical database that organizes Marathi words by meaning, grouping them into synonym sets (synsets) and connecting them through semantic relationships. Developed by the CFILT team at IIT Bombay, it uses the Hindi WordNet as a starting point: for every Hindi synset, Marathi experts select the best matching words and translate or compose natural Marathi example sentences. The resource covers all parts of speech and accounts for spelling and dialectal variations, following official Marathi standards. It also introduces language-specific synsets for concepts unique to Marathi life—like traditional foods, jewelry, dances, clothing, games, and rituals—ensuring the WordNet reflects the language’s true cultural flavor. All semantic relations and sense IDs are inherited from Hindi WordNet, supporting consistency and cross-language linkage.
Without a structured way to map out word meanings and relationships, it’s hard for both humans and computers to truly understand or process Marathi. This limits the quality of translation tools, search engines, and educational resources, and makes it harder to preserve the cultural uniqueness of the language.
- Comprehensive Synset Creation: Lexicographers review each Hindi synset, identify the matching Marathi words, and ensure conceptual consistency with natural Marathi examples.
- Rich Semantic Structure: Includes synonymy, antonymy, hypernymy, hyponymy, meronymy, and more, all linked through shared sense IDs.
- Cultural Representation: Adds language-specific synsets for Marathi foods (e.g., पाटवडी, कडबोळे), ornaments (कोल्हापुरी साज), dances (लावणी), clothing (नऊवारी), and games (झिम्मा, आट्यापाट्या).
- Accessible Online: The interface allows users to search for words, view all senses, and explore related words, with typing support for Devanagari and phonetic input.
The Marathi WordNet is available online at http://www.cfilt.iitb.ac.in/wordnet/webmwn/wn.php. Users can search for any word, view all its meanings (synsets), see example sentences, and explore semantic connections. The tool supports Devanagari keyboard and phonetic transliteration, making it easy to use for all. The database is regularly updated and is designed to link with other Indian language WordNets as part of IndoWordNet.
By making the structure and richness of Marathi accessible in a digital, systematic form, Marathi WordNet supports better language technology, preserves cultural heritage, and helps both learners and experts explore the language in depth.
- Multilingual machine translation engines (e.g., English–Hindi–Marathi systems)
- Lexical and semantic search engines in regional languages
- NLP education and linguistic research tools
- Chatbots and conversational AI in Marathi
- Cross-lingual information retrieval systems
- Cultural lexicons and local knowledge bases for Marathi digital content creators
- Language policy and curriculum design at educational and governmental levels