Credits & data sources

Multilinguo's typology data is sourced entirely from open academic databases and open-access reference works: WALS, Grambank, Autotyp, Universal Dependencies, WOLD, LangSci Press, SIL grammar sketches, and canonical author-cited reference grammars. Every feature value promoted to the app carries a citation back to a source on this page.

Each source below lists its license, canonical citation, and how it feeds the app. New sources appear here as they enter the pipeline.

Typology databases

These databases feed Multilinguo's 100-feature typological curriculum. Cells are cross-validated between multiple sources before being promoted to `live` in the app.

WALS — World Atlas of Language Structures

Typological feature values (192 structural features across ~2,600 languages) · CC-BY 4.0

Dryer, Matthew S. & Haspelmath, Martin (eds.) 2013. The World Atlas of Language Structures Online. Leipzig: Max Planck Institute for Evolutionary Anthropology.

How Multilinguo uses it: Track A primary source. Cross-validated against Grambank per feature × language cell. Values feed lg_features_reference with source_track='A' and wals_agreement_with_grambank flag.

https://wals.info/ ↗

Grambank

Typological feature values (195 binary/categorical features across ~2,400 languages) · CC-BY 4.0

Skirgård, Hedvig et al. 2023. Grambank reveals the importance of genealogical constraints on linguistic diversity and highlights the impact of language loss. Science Advances 9(16).

How Multilinguo uses it: Track A cross-validation source. Cells where WALS and Grambank agree promote to `live` automatically; disagreements hold for staff linguist review.

https://grambank.clld.org/ ↗

Autotyp

Fine-grained morphosyntactic feature values · CC-BY-SA 4.0

Bickel, Balthasar, Johanna Nichols, Taras Zakharko, Alena Witzlack-Makarevich, Kristine Hildebrandt, Michael Rießler, Lennart Bierkandt, Fernando Zúñiga & John B. Lowe. 2022. The AUTOTYP database (v1.0.1).

How Multilinguo uses it: Track A v1 addition (planned). Third cross-validation source for morphosyntactic features where WALS and Grambank leave residual disagreements.

https://github.com/autotyp/autotyp-data ↗

Universal Dependencies

Cross-linguistically consistent morphosyntactic annotation across 100+ treebanks · CC-BY-SA / CC-BY-NC-SA per treebank (see de Marneffe et al. 2021)

de Marneffe, Marie-Catherine, Christopher D. Manning, Joakim Nivre & Daniel Zeman. 2021. Universal Dependencies. Computational Linguistics 47(2): 255-308.

How Multilinguo uses it: Track A v1 addition (planned). Provides morphosyntactic feature values derived from treebank statistics. Attribution per contributing treebank preserved.

https://universaldependencies.org/ ↗

Reference grammar corpus

General typology references and per-language canonical reference grammars used as source material for Track B (LLM-assisted extraction) and Track C (staff-linguist authoring). Multilinguo does not distribute these works — they are consulted via open academic access routes (LangSci Press, SIL, university open-title archives, and public library access). Extracted values are always staff-verified before entering the app.

Comrie (ed.) — Language Universals and Linguistic Typology

General typology reference · Copyrighted — used under academic fair use

Comrie, Bernard. 1989. Language Universals and Linguistic Typology (2nd ed.). Chicago: University of Chicago Press.

How Multilinguo uses it: Track B LLM-extraction reference for cross-linguistic pattern claims.

https://www.wiley.com/ ↗

Croft — Typology and Universals

General typology reference · Copyrighted — used under academic fair use

Croft, William. 2003. Typology and Universals (2nd ed.). Cambridge: Cambridge University Press.

How Multilinguo uses it: Track B LLM-extraction reference for typological framework claims.

https://www.cambridge.org/ ↗

Haspelmath — Understanding Morphology

General morphology reference · Copyrighted — used under academic fair use

Haspelmath, Martin & Andrea D. Sims. 2010. Understanding Morphology (2nd ed.). London: Hodder Education.

How Multilinguo uses it: Track B LLM-extraction reference for morphological feature descriptions.

https://www.taylorfrancis.com/ ↗

Etymology authorities

Etymology stages (Proto-Indo-European → modern forms) are hand-authored by Multilinguo's staff linguist against these standard authorities. No LLM-generated etymology content appears on user-facing paths.

Buck — Dictionary of Selected Synonyms in the Principal IE Languages

Etymology stages authoring authority · Public domain (published 1949)

Buck, Carl Darling. 1949. A Dictionary of Selected Synonyms in the Principal Indo-European Languages. Chicago: University of Chicago Press.

How Multilinguo uses it: Primary authority for the hand-authored etymology trails feature (staff linguist).

https://press.uchicago.edu/ ↗

de Vaan — Etymological Dictionary of Latin

Latin etymology authority · Copyrighted — used under academic fair use

de Vaan, Michiel. 2008. Etymological Dictionary of Latin and the other Italic Languages. Leiden: Brill.

How Multilinguo uses it: Romance branch etymology stages (French, Italian, Spanish, Portuguese, Romanian).

https://brill.com/ ↗

Kroonen — Etymological Dictionary of Proto-Germanic

Germanic etymology authority · Copyrighted — used under academic fair use

Kroonen, Guus. 2013. Etymological Dictionary of Proto-Germanic. Leiden: Brill.

How Multilinguo uses it: Germanic branch etymology stages (English, German, Dutch, Swedish, Danish, Norwegian).

https://brill.com/ ↗

Derksen — Etymological Dictionary of the Slavic Inherited Lexicon

Slavic etymology authority · Copyrighted — used under academic fair use

Derksen, Rick. 2008. Etymological Dictionary of the Slavic Inherited Lexicon. Leiden: Brill.

How Multilinguo uses it: Slavic branch etymology stages (Russian, Polish, Czech, Slovak, Ukrainian, Croatian, Slovenian).

https://brill.com/ ↗

Oxford English Dictionary (OED)

English etymology authority · Subscription — used under academic fair use

OED Online. Oxford University Press.

How Multilinguo uses it: Modern English stage citations + English-side sound change dating.

https://www.oed.com/ ↗

Mallory & Adams — Encyclopedia of Indo-European Culture

PIE reconstruction authority · Copyrighted — used under academic fair use

Mallory, J.P. & Douglas Q. Adams (eds.). 1997. Encyclopedia of Indo-European Culture. London: Fitzroy Dearborn.

How Multilinguo uses it: PIE root reconstructions for etymology stages.

https://www.taylorfrancis.com/ ↗

Rix et al. — Lexikon der indogermanischen Verben (LIV)

PIE verb reconstruction authority · Copyrighted — used under academic fair use

Rix, Helmut et al. (eds.). 2001. Lexikon der indogermanischen Verben (2nd ed.). Wiesbaden: Dr. Ludwig Reichert Verlag.

How Multilinguo uses it: PIE verb root reconstructions where nominal sources leave gaps.

https://reichertverlag.de/ ↗

Sourcing method

Multilinguo's content pipeline is intentionally sourced from open academic databases (CC-BY / CC-BY-SA) as the primary layer, with open-access reference works (LangSci Press, SIL grammar sketches, Berkeley & MIT open titles) and canonical author-cited grammars as the secondary layer. Where nothing else is available, Wikipedia grammar sections serve as a low-confidence baseline, flagged as such in staging and always staff-verified before reaching users. No feature value is ever attributed to a source we cannot cite back to a public URL or a canonical print citation.

Methodology

For the full pedagogical framework these sources feed, see How it works.

Have a source we should add or correct? Email romain@multilinguo.app.