Question
What is lost when Indigenous-origin vocabulary is processed without cultural context, and how can culturally situated evaluation data be structured so that uncertainty is shown instead of hidden?
ResearchArchive / dataset · prototype
Research · Released prototype · 50 entries
Chilean Spanish Micro-Archive
A bilingual annotated lexicon of Chilean Spanish terms with Indigenous substrate: words from Mapudungún, Quechua and Aymara in which meaning, class coding and colonial history are compressed into everyday speech.
Chilean Spanish, Mapudungún, Quechua and Aymara. Lexical borrowings, semantic shifts, toponyms, polysemous terms and hybrid formations. 50 entries in this prototype; target 100+.
Question · methodChilean Spanish Micro-Archive
Question and method
Method
Each entry is bilingual and carries a verification-as-claims layer: etymologies are claims with a status, a method and sources. A claim is verified when at least one solid source supports it with an explicit method, disputed when sources conflict, unverified when it is analytically productive but not yet confirmed.
Sources / corpus
Lexicographic sources (Lenz 1910; Manquilef 1911; Valdivia 1684, all public domain through Memoria Chilena), corpus attestation and native-speaker testimony, labelled as such.
Current state
51 etymology claims: 38 verified, 9 unverified, 4 disputed.
Positionality
Curated by a native speaker of Chilean Spanish who is not Mapuche; that position limits access to the original Mapuche meanings and is documented as a methodological boundary. Background in Theory and History of Art, not formal linguistics.
Limitations
A proof-of-concept dataset, not a comprehensive dictionary. The web edition was last updated in July 2026 and the dataset in October 2026 (schema v0.4); field-level parity between them has not been verified yet.