Wikidata:Property proposal/Ethnologue language status

Ethnologue.com language status

Originally proposed at Wikidata:Property proposal/Authority control

   Done: Ethnologue language status (P3823) (Talk and documentation)

Description	language status identifier by Ethnologue.com using EGIDS scale
Data type	Item
Domain	language in the world
Allowed values	listed here
Example	Indonesian (Q9240) → 1 national
Source	https://www.ethnologue.com/about/language-status

Motivation

This property is important, we have UNESCO language status (P1999), we should have Ethnologue.com language status too. Beeyan (talk) 10:00, 2 November 2016 (UTC)

Discussion

• Keen support, but would prefer an item data type, like UNESCO language status (P1999). John Vandenberg (talk) 13:25, 2 November 2016 (UTC)
•  Support with datatype of "item". Only 13 items should be required. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:01, 2 November 2016 (UTC)
• External ID is absolutely wrong for this, this is a classification, not a unique id that would be applied to each item where the property is used. I support if datatype is "item", or "string". ArthurPSmith (talk) 18:25, 2 November 2016 (UTC)
•  Comment last property proposal: Wikidata:Property proposal/Archive/33#EGIDS language status --Pasleim (talk) 21:48, 2 November 2016 (UTC)
  • Thanks @Pasleim: for referring previous discussion with the same property. --Beeyan (talk) 08:39, 3 November 2016 (UTC)
• Note I can see the merit of having the EGIDS status, but ethnologue isn't always a reliable source for that (see the previous discussion). Uanfala (talk) 10:26, 5 November 2016 (UTC)
  • For most languages there simply haven't been any language vitality studies published, so if the property values get filled in from ethnolgue's best guesses, they'll be largely unreliable. Uanfala (talk) 11:59, 5 November 2016 (UTC)

•  Comment If we talk about reliable source, UN Atlas of the World's Languages in Danger also has some flaws. They're using old resources and not showing the recent status. Reliable sources are hard to find for defining language status, as we encounter so many contradictory resources. If we can get this property on Wikidata, we could see the differences between sources :) --Beeyan (talk) 05:15, 6 November 2016 (UTC)
•  Comment One of the previous objections was that Wikipedia would use this data, e.g. in an infobox, and all of the caveats are lost. Perhaps we can avoid that problem by always using 'deprecated' for any value which isnt from a reliable vitality study. We should be able to build a constraint to enforce that - "if no source, or source is ethnolgue, warn if the value status is not deprecated". John Vandenberg (talk) 06:24, 6 November 2016 (UTC)
  • It would not be appropriate to use the deprecated rank with the property proposed here (it would be fine for instance of (P31) statements though). Deprecated means that a statement is wrong and always was wrong, but if Ethnologue says that the status is X, then "the status in Ethnologue is X" is a true statement, even if people think Ethnologue is wrong to say that. For example, if Ethnologue says a language is extinct but we know it isn't because there are native speakers, we can mark a instance of (P31) dead language (Q45762) statement as deprecated because we know that "it is an extinct language" is wrong, but "Ethnologue language status" "extinct" shouldn't be marked as deprecated because "Ethnologue says that it's extinct" is still completely true. - Nikki (talk) 10:35, 22 December 2016 (UTC)
•  Comment Please provide an example that uses an item. ChristianKl (talk) 19:27, 15 November 2016 (UTC)
• Support Per nom. Iazyges (talk) 21:51, 15 November 2016 (UTC)
•  Oppose per previous discussion. Visite fortuitement prolongée (talk) 22:25, 15 November 2016 (UTC)
•  Comment Is there any outside discussion by linguistics about the status of Ethnologue that we can use to guide our decision? ChristianKl (talk) 10:45, 18 November 2016 (UTC)

• • Example: Indonesian (Q9240) → 1 national (it should be filled by an item data, just like UNESCO language status (P1999), cc: @ChristianKl:. Beeyan (talk) 08:14, 29 November 2016 (UTC)
    • @Beeyan: Saying "it should be different" isn't the point. Please actually create the example that you are looking for. This example just refers to a string. ChristianKl (talk) 15:54, 29 November 2016 (UTC)
    • @ChristianKl: there's no point creating the items if the property is not approved, and it's clear the intention is to use items as for similar properties, so I don't see why you are requiring Beeyan to do this now? ArthurPSmith (talk) 19:16, 29 November 2016 (UTC)
      • I don't think that this property needs to exist for the categories to exist. In general I think the point of a property proposal is to decide on the structure of something is supposed to be modeled. It doesn't make sense that the actual acct of modeling happens outside of the visibility that the property proposal process provides. ChristianKl (talk) 11:00, 30 November 2016 (UTC)
•  Support with item data type. Ethnologue is one of recommended sources at en:w:Template:Infobox language, so I have no concerns about general reliability. Also we already have tens of properties linking to obscure unreliable sources.--Jklamo (talk) 09:54, 15 March 2017 (UTC)
• @Jklamo, ArthurPSmith, Visite fortuitement prolongée, Iazyges, Beeyan, Nikki:  Done ChristianKl (talk) 08:15, 28 March 2017 (UTC)