Elamur Research⌕
☰
home / papers / 2026 / A Mīmāṃsā-Based Framework for Disambiguating Compounds in Sanskrit

A Mīmāṃsā-Based Framework for Disambiguating Compounds in Sanskrit

Paper · 2026 · Devulapalli Koushik Suraj and Pavankumar Satuluri

Summary

Sanskrit compounds (samāsa) often allow several grammatically correct internal analyses, and each analysis gives a different sense. Poetry exploits this deliberately through śleṣa, but technical and śāstric writing usually intends only one reading. Existing Sanskrit tools for sandhi splitting, morphology and compound generation either list every possible analysis or expect the user to specify the samāsa type. The authors draw on Pūrva Mīmāṃsā hermeneutics, specifically the Niṣādasthapati-nyāya. This maxim favours readings obtained through primary signification (abhidhā) over those that need secondary signification (lakṣaṇā), provided grammar and context do not rule them out. The authors treat the maxim as a conditional preference and build a layered decision procedure around it. First, candidate analyses are filtered grammatically. Next, the procedure flags candidates that rely on lakṣaṇā. In the last step, contextual fit is assessed using the tātparya-liṅgas, yogyatā, and the relation between qualified and qualifier (viśeṣya–viśeṣaṇa). The study is purely conceptual, illustrated with Vedic and classroom examples and a methodology flowchart. It offers no implementation, dataset or quantitative evaluation. Its contribution is a principled basis for interpreting compounds without user input, which could later support Sanskrit WSD, parsing, anvaya construction and machine translation.

Key points

  • The core principle is the Niṣādasthapati-nyāya. It prefers lakṣaṇā-free (abhidhā-based) compound readings, but only when nothing grammatical or contextual blocks them, so it guides rather than dictates.
  • Of the six tātparya-liṅgas listed in Mīmāṃsā, the authors consider two most useful for compounds: abhyāsa (repetition) and upakrama–upasaṃhāra (opening and closing).
  • The paper distinguishes three ambiguity scenarios:
    • bahuvrīhi vs. tatpuruṣa, where both involve lakṣaṇā: context decides, through the verb, the qualified noun or another related word, as in the pītāmbara examples.
    • a lakṣaṇā-free reading with little context: the nyāya alone suffices, as with nīlotpalam read as karmadhāraya.
    • both nyāya and liṅgas applicable: they work together.
  • No system is built or evaluated, and no corpus, model or accuracy figures are reported.
  • Planned future work includes:
    • resolving conflicts among tatpuruṣa subtypes;
    • adding discourse-level and pragmatic signals;
    • implementing the framework and testing it empirically on parsing, semantic interpretation and WSD.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Fields, topics and concepts

📚 Shelves 2The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 works

Cite

Cite as Devulapalli Koushik Suraj and Pavankumar Satuluri. 2026. A Mīmāṃsā-Based Framework for Disambiguating Compounds in Sanskrit. Proceedings of the 8th International Sanskrit Computational Linguistics Symposium.

The paper itself, not this page. BibTeX from the library's record of the work.

@inproceedings{suraj2026mimamsa,
  title = {{A Mīmāṃsā-Based Framework for Disambiguating Compounds in Sanskrit}},
  author = {Devulapalli Koushik Suraj and Pavankumar Satuluri},
  booktitle = {Proceedings of the 8th International Sanskrit Computational Linguistics Symposium},
  year = {2026},
  url = {https://aclanthology.org/2026.iscls-1.7/}
}

References

30 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Sri G. Mahabaleshwara Bhatt. 2006. Samāsaḥ Saṃskṛtabhāratī. find
  2. Johannes Bronkhorst. 2001. Language and Reality: On an Episode in Indian Thought. Brill, Leiden. find
  3. George Cardona. 1997. Pāṇini: His Work and Its Traditions, volume 1. Motilal Banarsidass, Delhi. find
  4. Andrea Ciabattoni, Björn Lellmann, and Elisa Freschi. 2019. A deontic logic for mīmāṃsā. Journal of Philosophical Logic, 48(6):1127–1160. find
  5. Andrea Ciabattoni, Janneke Dik, and Elisa Freschi. 2025. Mīmāṃsā on “better-not” permissions. Journal of Philosophical Logic, 54(3):651–689. find
  6. Francis X. Clooney. 1990. Thinking Ritually: Rediscovering the Pūrva Mīmāṃsā of Jaimini. De Nobili Research Library, Vienna. find
  7. Madhav M. Deshpande. 1991. Sanskrit and Prakrit: Sociolinguistic Issues. Motilal Banarsidass, Delhi. find
  8. Pawan Goyal, Gérard Huet, Amba Kulkarni, and Shripad Sarma. 2012. A natural language processing framework for sanskrit. In Proceedings of the 24th International Conference on Computational Linguistics (COLING 2012), pages 25–34, Mumbai. find
  9. Oliver Hellwig. 2010. Design and development of a sanskrit treebank. In Proceedings of the Fifth International Sanskrit Computational Linguistics Symposium, pages 161–170, Kyoto. find
  10. Oliver Hellwig. 2015. Sanskrittagger: A pos tagger for sanskrit. In Proceedings of the 12th International Conference on Natural Language Processing (ICON), pages 252–259. find
  11. M. Hiriyanna. 1952. Outlines of Indian Philosophy. Allen & Unwin, London. find
  12. Gérard Huet. 2009. Sanskrit computational linguistics: A survey. Journal of Indian Philosophy, 37(4):403–428. find
  13. Daniel H. H. Ingalls. 1965. An Anthology of Sanskrit Court Poetry (Vidyākara’s Subhāṣitaratnakoṣa). Harvard University Press, Cambridge, MA. find
  14. Ganganatha Jha. 1916. The Pūrva Mīmāṃsā Sūtras of Jaimini. Panini Office, Allahabad. Translation. find
  15. Shivram D. Joshi and Johannes A. F. Roodbergen. 1991. Patañjali’s Vyākaraṇa-Mahābhāṣya. Sahitya Akademi, New Delhi. find
  16. Pandurang Vaman Kane. 1941. History of Dharmaśāstra, volume 1. Bhandarkar Oriental Research Institute, Poona. find
  17. Amba Kulkarni, Soma Paul, Malhar Kulkarni, Anil Kumar, and Nitesh Surtani. 2012. Semantic processing of compounds in indian languages. In Proceedings of COLING 2012, pages 1489–1502, Mumbai, India. find
  18. The COLING 2012 Organizing Committee. find
  19. Anil Kumar, Vipul Mittal, and Amba Kulkarni. 2010. Sanskrit compound processor. In Proceedings of the International Symposium on Sanskrit Computational Linguistics, Lecture Notes in Computer Science, pages 57–69. Springer. Available: https://sanskrit.uohyd.ac.in/faculty/amba/PUBLICATIONS/ papers/scl… ↗
  20. S. Nehrdich, O. Hellwig, and K. Keutzer. 2024. By t5-sanskrit: A unified model for sanskrit nlp tasks. arXiv preprint arXiv:2409.13920. — [arXiv:2409.13920](https://arxiv.org/abs/2409.13920) find
  21. S. B. Pancoli. 2021. Vaiyākaraṇabhūṣaṇasāraḥ. Chaukhambha Sanskrit Sansthan, Varanasi. find
  22. Amba Kulkarni pavankumar Satuluri. 2013. Generation of sanskrit compounds. In Proceedings of the International Conference on Sanskrit Computational Linguistics, pages 19–28. AISC. Available at: https://www.researchgate.net/profile/Amba-Kulkarni/publication/ 259655936_Generation_of_Sanskrit_Compounds… ↗
  23. Kunjunni Raja. 1969. Indian Theories of Meaning. The Adyar Library and Research Centre, Madras. find
  24. K. K. Raja. 1977. Indian Theories of Meaning. The Adyar Library and Research Centre, Madras. find
  25. P. Ramanujan. 1993. Principles of Mīmāṃsā for Natural Language Processing. Centre for Development of Advanced Computing (C-DAC), Pune. find
  26. Peter Scharf and Malcolm Hyman. 2009. Linguistic Issues in Encoding Sanskrit. Motilal Banarsidass, Delhi. find
  27. Jean-Marie Verpoorten. 1987. Mīmāṃsā hermeneutics and the interpretation of vedic sentences. Indo-Iranian Journal, 30:1–22. find
  28. Sadānanda Yogīndra. 1931. Vedāntasāra of Sadānanda. Advaita Āśrama, Almora. find
  29. Āpadeva. 1972. Mīmāṃsā Nyāya Prakāśa: With Sarva Vivecinī Explanation. Chowkhamba Sanskrit Series, Varanasi. find
  30. Dharmendranātha Śāstrī. 1953. Nyāya Siddhānta Muktāvalī. Rashtriya Sanskrit Sansthan. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/suraj-2026-mima-sa-based-framework-disambiguating/