What this shows
Every wordform in the corpus can potentially be the surface form of multiple lemmas (dictionary headwords). For example, the wordform "laulu" could be a genitive of "laul" (song) or a partitive of "laulu". This page lets you explore that morphological ambiguity.
Data sources
- Corpus count: How many times this wordform-lemma pair appears in the raw runosong corpus annotations.
- DS count: How many times DeepSeek AI analysis attributed this wordform to this lemma across 165K poems.
- Candidates: The number of distinct lemmas a wordform can map to.
How to use
- Search for a specific wordform to see all its candidate lemmas with frequency breakdown.
- Browse the top ambiguous wordforms table — sorted by number of candidate lemmas.
- Click any wordform row to see its full lemma distribution.
- Click a lemma to look it up in the main dictionary.
Loading wordform distribution data (23 MB)...