Similar Stanzas in the Vedic Saṃhitās

About this Resource

What is this?

This tool helps you discover semantically related stanzas within and across Vedic texts. Using a machine learning model trained on multiple English translations of the same stanza, we have identified stanzas that express similar concepts or themes, even when the exact wording differs.

The system was built by fine-tuning a sentence embedding model on aligned stanza pairs from the Rigveda (Jamison & Brereton vs. Griffith translations) and Atharvaveda (Whitney & Lanman vs. Griffith). The model learns that different translations of the same stanza are semantically equivalent, then uses this understanding to find conceptually related stanzas across the corpus. Each stanza is compared to all others using cosine similarity, and links with scores above 0.55 are displayed here.

How to use this tool

1. Select a Samhita from the first dropdown
2. Select a Book number from the second dropdown
3. Select a Sukta number from the third dropdown
4. Select a Stanza number from the fourth dropdown
5. The selected stanza and its semantically related passages will appear below.

Understanding the links: Each linked stanza shows its location (Book.Sukta.Stanza), translation, and similarity score. The colored bar on the left indicates the strength of the semantic relationship: red/orange for weaker matches (0.55), transitioning through yellow to green for very strong matches (approaching 1.0). A score of 1.0 indicates identical or nearly identical passages.

Note: This is an experimental tool. The semantic matches are based on English translations and computational similarity — they must be verified against the original Sanskrit texts for scholarly work.