Community Wishlist Survey 2022/Wikidata/Rework Lexemes to be more intuitive

Rework Lexemes to be more intuitive

  • Problem: As it stands now, we have L3257-S1 (apple - english) with glosses in 7 languages and translations in 5. One of these translations, L53267-S1 (manzana - espanol) has the same meaning but only 3 glosses and 3 translation statements. Unifying all glosses and their statements across all language variants of a single sense manually is very time consuming and tedious.
  • Proposed solution: I'd suggest improving this process by not binding senses to a specific language. Have all glosses and statements linked to a particular sense (as it is now), and then link that single sense to all respective languages, to avoid missing or duplicated data.
  • Who would benefit: All lexeme editors
  • More comments:
  • Phabricator tickets: phab:T290858
  • Proposer: —Ivi104 01:26, 12 January 2022 (UTC)

Discussion

Have you given any thought to where information on senses that is tied to particular lexemes with respect to those senses should go after such a transition? How should the language styles of senses (to separate e.g. L361-S1 vs. {L7669-S1}), the citable definitions (of a word for "fly" in a German dictionary vs. that of "letjeti" in a Russian one), different conceptions of antonymy across languages (the word for "dark" having a different antonym in French vs. Persian), regional indications of senses ("bubbler" being a generic English noun but with the specific meaning "drinking fountain" in Wisconsin, while other languages might have words for this without region constraints), indications of specialization of senses (military-specific use of the Bokmål "perm" meaning "leave of absence" vs. "permisjon" with the same meaning), different behaviors of arguments of lexemes with respect to senses (English "to like" having a nominative liker and accusative likee, but German "gefallen" with the same meaning having a dative liker and nominative likee), and so on be indicated if the senses themselves are forced to be centralized? Mahir256 (talk) 22:00, 12 January 2022 (UTC)

@Mahir256: I'm not a linguist so I haven't thought that far ahead. I think this could be handled by making a new "Sense:" namespace that could be tied to a particular meaning of a particular lexeme.
As it stands now, Lexemes are tied to languages - what if lexemes and senses were multilingual, and the properties of a sense could differ according to a specific language? All languages have a concept (a sense) of an apple, a concept of love, a concept of flying, and they are all the same. That same sense has a different name and different connections (properties) in a particular language - e.g. French and Persian "dark" have a different antonym. I am very much looking forward to feedback - this is just an idea off the top of my head. —Ivi104 00:39, 15 January 2022 (UTC)

Voting

  •   Support Love the lexeme project but yes it needs design to be more user friendly to input data. Bluerasberry (talk) 19:51, 28 January 2022 (UTC)
  •   Support Aca (talk) 15:46, 29 January 2022 (UTC)
  •   Support Warmglow (talk) 17:06, 29 January 2022 (UTC)
  •   Support Libcub (talk) 00:04, 31 January 2022 (UTC)
  •   Support Novak Watchmen (talk) 17:20, 31 January 2022 (UTC)
  •   Support Fix some lexeme problem now! Thingofme (talk) 14:30, 4 February 2022 (UTC)
  •   Support Waldyrious (talk) 00:44, 6 February 2022 (UTC)
  •   Support --Ciao • Bestoernesto 16:45, 6 February 2022 (UTC)
  •   Support Ayumu Ozaki (talk) 04:03, 7 February 2022 (UTC)