Credits
Built on the work of others.
Lexis exists because people shared their knowledge freely. Here's what we use, and under which terms.
Words and languages
| Source | What | Licence |
|---|---|---|
| PanLex, a project of the Long Now Foundation | All words and translations (25,598,007 words, 7,027 sources) | CC0 1.0 |
| Glottolog | Language families, location, endangerment status, English language names | CC BY 4.0 |
| Wikidata | Speaker numbers, countries, Dutch language names | CC0 |
| Wikipedia | The text about a language (credited in the app) | CC BY-SA 4.0 |
| OpenStreetMap | The small maps on language pages | © OpenStreetMap contributors, ODbL |
| LDNOOBW | Starting point for the list of sensitive words | CC BY 4.0 |
We cite PanLex as requested: Kamholz, David, Jonathan Pool and Susan M. Colowick. 2014. PanLex: Building a Resource for Panlingual Lexical Translation. Proceedings of LREC 2014.
Images
| Image | Maker | Licence |
|---|---|---|
| The Earth (Blue Marble) | NASA, Apollo 17 crew, 1972 | Public domain |
| Example images on this site | Made for the Lexis prototype | © the maker |
Pronunciation in other scripts
For most scripts, the pronunciation shown under words in another script is computed with fixed rules, using pykakasi (Japanese, GPL-3.0), pypinyin (Chinese, MIT) and Unidecode (others, GPL-2.0). In nine scripts rules aren't enough, because vowels aren't written or the spelling is too complex: Arabic, Persian, Urdu, Hebrew, Thai, Khmer, Lao, Burmese and Tibetan. For the core words in those languages, the pronunciation was made once, behind the scenes, with AI (Claude). Where it wasn't sure, nothing is shown.
Numbers on this site
Words, languages and sources: counted in PanLex Lite (June 2024). "Languages" means languages with at least one word. Endangered languages: Glottolog status (3,878 of 6,607 living languages with a known status).