Phase 3: language packs that prefer or restrict their language #753

Activity

1. derek73

   self-assigned this

   on Oct 19, 2026
2. derek73

   added

   design

   vocabularyPer-word vocabulary model and language packsPer-word vocabulary model and language packs

   on Oct 19, 2026
3. derek73

   added a parent issue on Oct 19, 2026

   • Vocabulary model and language packs (tracking) #747
4. derek73

   added sub-issues on Oct 19, 2026

   • Locale packs for Alexandre Dumas fils and Jiří Dienstbier ml.: generational words that collide outside their language #684
   • Ship hi and bn locale packs for Latin-transliterated Indic honorifics #345
   • Thai honorifics are glued to the given name (นายสมชาย → given "นายสมชาย") #317
   • Should Juan Sta. Maria read family Sta. Maria? Today it reads middle Sta., family Maria #725
5. derek73 commented on Oct 24, 2026

   

   derek73

   on Oct 24, 2026

   OwnerAuthor

   More actions

   A first measured case for "a pack prefers its language": Nepali (ne) against the Hindi and Marathi vocabulary.

   The nepal-candidates local corpus (#866) writes each of 125,331 local-election candidates in Devanagari beside the Election Commission's own romanization, so it shows two languages' vocabularies disagreeing about one word, with the writer's intent recorded beside it.

   श्री (#870). hi.toml and mr.toml list it as a salutation (Shri, Mr.; #269), and in a Hindi or Marathi name that is right. In Nepali names it is also the first element of a given name:

   	names	today	the Commission's romanization
   spaced श्री opening the name	215	title in all 215	a name word in 201 (Shree 193, Shri 6, Sri 2), Mr in 2
   श्री glued into the given name	386	given	—

   So श्री कृष्ण श्रेष्ठ reads title श्री while its own romanization Shree Krishna Shrestha reads given Shree.

   Buddha, Siddha, Gyani, Guru (#869). Title words in the default vocabulary (buddha/siddha religious in en.toml, gyani in pa.toml, guru the renunciate title of #346), and the opening element of common Nepali compound given names: Buddha Bahadur Tamang reads title Buddha, given Bahadur. Of the names opening with one of them, 172 read it as a title.

   Why this is a pack question and not a default change. Both readings are right for some language: the default loads every language, so it can only pick one, and either pick breaks the other language's names. Decision 6 (decisions/vocabulary-model) is the mechanism for exactly this: a ne pack that prefers its language settles the clash where Nepali has a vocabulary reading, reading श्री and Buddha as given-name elements, while the default and a hi pack keep the salutation. The Phase 2 spec (#752) decides how the default gates title words that are also names (its D12); the Nepali cases may still belong to the pack rather than to the default gate, and that is recorded there.

   What a ne pack would carry, from these two issues: given-name readings for श्री (spaced, opening a name) and for buddha, siddha, gyani, guru opening a compound given name, with श्रीमती (92 names, romanized Shreemati in 62 and Mrs/Ms in 11) left a salutation. The corpus's two scripts give the pack its own test: a Devanagari name and its romanization should read the same roles.