My article [Non-autonomous accusative case in Estonian][1] references two corpus studies that I conducted to investigate whether...
1. the relative pronoun alternation between `mis~mille` was restricted to object position,
2. or if it was available in all genitive positions (objects, adposition complements, possessor/genitive position)
In the article, I give information about how I queried the corpora, but here I actually upload the (partially annotated) data that I pulled for these studies. Unfortunately, it isn't completely annotated. Since I can read some Estonian, I didn't need to fill out every column.
`estacc-relpn-corpussearch-adposition.csv`
* Pulled from [EtTenTen corpus][2]
* Query: `, +mis@word +(k)`
* Examples: 200/4969 total hits
* Columns:
* `token`: example + source
* `AdP`: the adposition next to `mis`
* `meaning`: adposition translation
* `here`: roughly, the adposition's syntactic function in the token
* `ekss`: roughly, what the adposition lemma's range of syntactic functions is
* `mis`: roughly, what `mis`'s role is in the relative clause
`estacc-relpn-corpussearch-possessor.csv`
* Pulled from [EtTenTen corpus][2]
* Query: `, +mis@word &nn>@syn +*mine(s)`
* Examples: 82/82 total hits
* Columns:
* `id_no`: an id number to be able to return to original sort order
* `example`: source + example text
* `coded_subj`: was the relative pronoun tagged as subject in the corpus (some were tagged multiple roles)
* `syn`: roughly, what I think the relative pronoun's syntactic role actually is
* `nominalization_translation`: What the `-mine` nominalization's translation is
[1]: https://ling.auf.net/lingbuzz/004500
[2]: https://www.keeleveeb.ee/dict/corpus/ettenten/