Runosong texts often use dialectal or archaic word forms that differ from the standard dictionary lemma. For example, "taha" (dialectal) vs "tahtma" (standard Estonian).
These pairs were extracted from DeepSeek AI annotations that identified the standard lemma for each dialectal form in the corpus.
Studying these pairs reveals systematic sound correspondences and regional linguistic patterns preserved in folk poetry.
Sound correspondence patterns
The bar chart shows the most common character-level transformations detected between standard and dialectal forms.
Patterns like "aa→ea" or "h→∅" (h-dropping) indicate systematic phonological processes in different dialect areas.
Bars are split by language (blue = Estonian, red = Finnish) to show which patterns are language-specific.
Click any pattern bar to filter the table below to only show pairs with that pattern.
Using the table
Wordform: The form as it appears in the corpus text.
Standard: The standard dictionary lemma this form maps to. Click to open in the main lexicon.
Dialectal: The dialectal lemma identified by the annotation system.
Pattern: The detected sound change pattern (character-level transformation).
Sort by any column by clicking the header. Filter by language (Estonian/Finnish) or search by text.