Concept pipeline / candidate discovery
Which words deserve
a concept page?
Start with terms that recur across major Sanskrit works, then use dictionary coverage to distinguish substantial concepts from incidental forms.
— candidates20 primary documents5+ dictionaries
RankTerm
OccurrencesDocumentsDictionariesDecision
Counting surface forms across the corpus…
Method
This is a directional view, not lemmatized corpus truth. It counts normalized surface forms across the 20 largest primary Sanskrit documents, excludes common function words and apparatus labels, and retains terms found in at least five documents and five Toroid dictionaries.