Property talk:P37

From Wikidata
Jump to: navigation, search


official language
language designated as official by this item
Description Official language of an organization (Q43229), including municipalities and states
Represents official language (Q23492)
Data type Item
Template parameter en:Template:Infobox country, official_languages
According to this template: organization (Q43229) : state (Q7275), local government (Q6501447), but also smaller organizations
According to statements in the property:
organization (Q43229), administrative territorial entity (Q56061) and fictional organization (Q14623646)
When possible, data should only be stored as statements
Allowed values language (Q34770), mainly natural language (Q33742) (note: this should be moved to the property statements)
According to statements in the property:
Czech Republic (Q213)Czech (Q9056)
Hawaii (Q782)Hawaiian (Q33569) and English (Q1860)
France (Q142)French (Q150)
Austria (Q40)German (Q188)
Kazakhstan (Q232)Russian (Q7737)
When possible, data should only be stored as statements
Tracking: usage Category:Pages using Wikidata property P37 (Q23908973)
Proposal discussion Originally created without a formal discussion
Current uses 4,378
[create] Create a translatable help page (preferably in English) for this property to be included here
Type “organization (Q43229), administrative territorial entity (Q56061), fictional organization (Q14623646): element must contain property “instance of (P31)” with classes “organization (Q43229), administrative territorial entity (Q56061), fictional organization (Q14623646)” or their subclasses (defined using subclass of (P279)).
Exceptions are possible as rare values may exist.
List of this constraint violations: Database reports/Constraint violations/P37#Type Q43229, Q56061, Q14623646, SPARQL
Value type “languoid (Q17376908): This property should use items as value that contain property “instance of (P31)”. On these, the value for instance of (P31) should be an item that uses subclass of (P279) with value languoid (Q17376908) (or a subclass thereof).
Exceptions are possible as rare values may exist.
List of this constraint violations: Database reports/Constraint violations/P37#Value type Q17376908, SPARQL
Qualifiers “end time (P582), start time (P580): this property should be used only with listed qualifiers.
Exceptions are possible as rare values may exist.
List of this constraint violations: Database reports/Constraint violations/P37#Allowed qualifiers, SPARQL
This property is being used by:

Please notify projects that use this property before big changes (renaming, deletion, merge with another property, etc.)

Label change[edit]

I've changed this label from "national language" to "official language", since "national language" is somewhat vague. If this puts anything in the wrong place, the solution would be to just create a property for unofficial widely-spoken languages or something. — PinkAmpers&(Je vous invite à me parler) 00:40, 5 February 2013 (UTC)

Why only country?[edit]

Why does this only cover countries? Regions, municipalities, organizations (EU, UN), etc, may also have official languages, without there really being any difference to the meaning of "official language". I suggest that this property should cover all uses. Jon Harald Søby (talk) 18:44, 5 February 2013 (UTC)

Good point. We've now pretty much gone in the opposite direction of the original "national language" label, but I, for one, would Symbol support vote.svg Support amending the description to allow for non-country entities as well. — PinkAmpers&(Je vous invite à me parler) 14:58, 6 February 2013 (UTC)
In Norway and Finland, each municipality has it's own "official language". I see no point in not allowing this property to be used for such purpose. -- Lavallen (block) 15:53, 28 April 2013 (UTC)
A language can be assigned to any region, not only nations and subregions of them, but an official language should be assigned only to official regions. Jeblad (talk) 14:03, 24 February 2016 (UTC)

I think the best solution is the one Snipre proposed in the Project Chat a while ago: just merge all the language properties (P:P37, P:P103, P:P424) in P:P407 and use qualifiers when needed. There is no way we can maintain decicated properties for all cases (language with regional status but not official, language widely spoken but not really official, language spoken by a person but not natively, etc). I cannot see any real use in having separate properties that force us to bend reality. --Zolo (talk) 16:17, 28 April 2013 (UTC)

I cannot see that P424 can be merged since it is a string-property, but I see your point for the rest. -- Lavallen (block) 16:28, 28 April 2013 (UTC)

"No value" as "None"?[edit]

If an entity has no official language, is it correct to add this property and set it to "No value", as in Q30? --Yair rand (talk) 15:46, 28 April 2013 (UTC)

Is it possbile to add "none" as prefered value and add others with a qualifier? -- Lavallen (block) 16:31, 28 April 2013 (UTC)


I've commented a couple of constraints because they have so many violations that they are arguable. Let's discuss them separately. Infovarius (talk) 09:42, 18 October 2013 (UTC)

Type of administrative division[edit]

1) There's another candidate for the root of divisions: no label (Q4057763). 2) At present the tree of the divisions is rather random and uncomplished so the constraint doesn't work good. 3) What about countries? Should they also have the same claim? Infovarius (talk) 09:42, 18 October 2013 (UTC)

I put back the domain constraint on organization (Q43229), which is the high-level parent of state (Q7275) and local government (Q6501447), but also smaller organizations. I believe that only organization (Q43229) have authority to designate some language as "official". Note that the concept of country (Q6256) is unclear and thus is not a good base for the domain of a currency. LaddΩ chat ;) 16:30, 5 April 2014 (UTC)

is a language[edit]

I (and obviously many others) can't solidly distinguish between 3 items for language: language (Q315), language (Q34770) and natural language (Q33742). Some languages have "is a q315" statements, some "is a q34770" and some "ïs a Q33742". Until the mess persists the constraint has no sense. Infovarius (talk) 09:42, 18 October 2013 (UTC)

The purpose of having a constraint and clear property documentation is to resolve such issues. I hope usage is clearer now. LaddΩ chat ;) 21:48, 1 April 2014 (UTC)

Allowed value[edit]

(Allowed value => range) An official language can be a language family. In some cases variants are listed as a dialect and in other cases a variant is a new language. In the later case an official language can be a language family. Jeblad (talk) 15:21, 7 July 2015 (UTC)

I changed it to languoid (Q17376908) used for the constraints of some of the other properties. --- Jura 07:32, 8 July 2015 (UTC)

Label, alias et use of this property[edit]


With Ludo29 with have a discussion (in French, but I'll sum it up here in english) on Talk:Q39 about how handle official language (P37) on Switzerland (Q39). Right now, there is little date, only  :

added and removed periodically...

I want to get out of this messy situation. For a start and to open the discussion, I propose this :

I suggest to use official language (Q23492), national language (Q319972) and minority language (Q61566) but in the long run it would probably be better to create specific Swiss sub-items like : « Swiss official language », « Swiss national language », « Swiss national language ​​without territory » (at least for the last one, as it's a very specific status). I suggest as (P794) but maybe instance of (P31) would be better (instance of (P31) already used on Sweden (Q34)). If everybody agree, the constraint need to be completed.

Ludo29 disagree on using official language (P37) for non-official language (the two last) and suggest to split official language (P37) and create a new property for national language. I must say I'm puzzled, I've got a wide interpretation of official language (P37) but the label is confusing. What do you think? Alternatively, we can maybe use language (P2439) instead. If we keep only one property for all « officially designated language » maybe we should adapt the label to be less confusing (see too the first thread on the talk page by PinkAmpersand about the label change from « national language » to « official language »).

I don't know if ranking could be use. I suggested to use : « preferred » for the three official and « normal » for the two others or « normal » for the three official and « deprecated » for the two others ; but as Infovarius pointed it out, this is not how ranks are usually intended to be used.

@Jon Harald Søby, Lavallen, Zolo, Yair rand, Laddo, Jeblad: who wrote on this talk page. @Pikolas: because I talked about

< France (Q142) View with Reasonator View with SQID > official language (P37) View with SQID < French (Q150) View with Reasonator View with SQID >
has dialect (P134) View with SQID < French of France (Q3083196) View with Reasonator View with SQID >

Does anyone have an example with an other country (I see that the situation is a bit exotic on the items Sweden (Q34) and Spain (Q29) but I don't know well the language status on these country).

Cdlt, VIGNERON (talk) 11:49, 24 February 2016 (UTC)

The latter is a misuse of has dialect (P134), meant to be used strictly on languages. Why not simply use language (P2439) directly on the country's item for non-official ones? LaddΩ chat ;) 12:03, 24 February 2016 (UTC)
@Laddo: Yes has dialect (P134) is odd (plus there is no reference). Yes again, language (P2439) is a solution but is it a good one and the best one? And if we use language (P2439) we should remove the aliases like « fr: langue nationale » or « de: Landessprache » on official language (P37) to be consistent. Cdlt, VIGNERON (talk) 12:53, 24 February 2016 (UTC)
I think the main issue seems to be that there are different opinions about the status of Romansh. Usually, we would keep different statements in such cases. A qualifier that may help could be "when communicating with persons who speak Romansh". This seems to be in line with how qualifiers at Wikidata work. Not sure how qualifiers that state the contrary could help.
--- Jura 12:31, 24 February 2016 (UTC)
@Jura1: I'm not sure. Nearly everybody is agree on « Romansh is offically a 'national language' and not an 'official language' of Switzerland ». The question is more : is official language (P37) strictly for « official language » or more broadly for « officially designated language » ? Cdlt, VIGNERON (talk) 12:53, 24 February 2016 (UTC)
Given the edits on Q39, it's hard to determine. In any case, the status changed twice since 1848. Your sample only includes one.
--- Jura 12:57, 24 February 2016 (UTC)
@Jura1: from what I've seen in the history, the edit are going back and forth from 'nothing' to official language (P37) = Romansh (Q13199) without any qualifiers. I think qualifiers can help solving the problem and improve this item. My sample is just a proposition (and I hope an improvement from the current statements), please refine and improve it ! Cdlt, VIGNERON (talk) 17:39, 24 February 2016 (UTC)
Much to long! Sorry!
A nation has one or more majority languages, and zero or more minority languages. All of them may have an official status. The minority languages can have a status as endangered languages, and then the state should have some kind of support for them them. Some of the languages can be listed as official in some sub-part of the nation, without being official in the larger state.
Note that a written form of a language is not necessarily a dialect, although it can be a dialect. The difference between a dialect and a language can be somewhat blurred, and changing over time.
In Norway we have several languages, with different degree of official status. There are some information about this at Facts about the Nordic Region. Note that the page describes the situation in several other Nordic countries in addition to Norway.
The differences between a official national language and an national official language is interesting, as Norwegian is an official national language while Northern Sami language is an national official language. Northern Sami language is spoken in Norway, Sweden, Finland and Russia. A specific language does not imply nationality, but different states may try to normalize the language according to different rules.
Norwegian language (no) is the majority language, with two official written forms, Bokmål (nb) and Nynorsk (nn), and two unofficial forms Riksmål and Høgnorsk. Written names in the later forms follows spellings from Bokmål or Nynorsk.
Sami languages are minority languages that has status as official languages in Norway, and are used as an administrative language in the municipalities Kautokeino (Northern Sami: Guovdageaidnu), Karasjok (Northern Sami: Kárášjoga), Kåfjord (Northern Sami: Gáivuotna), Nesseby (Norther Sami: Unjárgga), Porsanger (Northern Sami: Porsáŋgu), Tana (Northern Sami: Deatnu), Tysfjord (Lule Sami: Divtasvuodna, Northern Sami: Divttasvuotna) og Snåsa (Southern Sami: Snåase). Different municipalities use different Sami languages, and the municipality Tysfjord even have two different Sami languages. The official Sami languages in Norway are Northern Sami, Southern Sami and Lule Sami language. Southern Sami and Lule Sami are listed as endangered languages. All three Sami languages are covered by the w:en:European Charter for Regional or Minority Languages
Kven language is a minority language that has status as official language in Norway, and are used as an administrative language in the municipality Porsanger (Kven: Porsanki). Kven language is listed as an endangered language, and is covered by the w:en:European Charter for Regional or Minority Languages
Scandoromani language (Romani) and Romani language (Romanes), yes the Norwegian names are confusing, are minority languages that has status as official languages in Norway, but are not used as an administrative language in any local region. I'm not sure, but I think Scandoromani language is covered by w:en:European Charter for Regional or Minority Languages.
Norwegian Sign language is a minority language that has status as official language in Norway, but are not used as an administrative language in any local region.
I don't think it is important to note whether a language is a majority or minority language in the property itself, but it is probably important to note if it has an official status. It is also important to distinguish between official languages, for example in Brazil and Russia there can be a lot of them, and administrative languages, which can be quite few. When used as a statement on a region it should be possible to give them qualifiers for at least minority language and majority language, but I'm not sure what such a property should be called.
Looking at articles about official languages in Wikipedia it seems like they describe administrative languages. Languages used by the the administration in some region is obviously official languages, but they are usually just a subset of the recognized languages in the state.
There is also the problem how we describe an language that isn't recognized by the state itself. This situation exist in Sweden with Elfdalian (no it's not from Lord of the rings) which has more in common with West Norse dialects or really languages than Swedish. According to the Swedish Language Act, Swedish is the main official language and there are six official minority languages; Finnish, Meänkieli (Torne Valley Finnish), Saami, Yiddish, Romani Chib and Swedish sign language. Elfdalian is not recognized as an official language.
It could be interesting to add qualifiers describing when a language become official or became an administrative language, and why. Perhaps "why" would go in the sources section. Jeblad (talk) 00:37, 25 February 2016 (UTC)
Very long but very interresting. Thank you Jeblad.
More pragmatically, how would you put these informations on Norway (Q20) ?
Cdlt, VIGNERON (talk) 08:58, 25 February 2016 (UTC)
Given the present available properties, I would do something like this for Norway (Q20)
I've somewhat misused a few items. national language (Q319972) double as majority language, and I'm not sure that is correct. official language (Q23492) is also linked as administrative language on enwiki, and I use it as that for Norwegian.
Note that I would link to Norwegian (Q9043), and not the individual written form of Bokmål (Q25167) and Nynorsk (Q25164). Perhaps we need a way to specify written forms. This is not a dialect! Source for the statement would be Lov om målbruk i offentleg teneste.
Note also that I would link to Sami languages (Q56463), and not the individual languages Northern Sami (Q33947), Southern Sami (Q13293) and Lule Sami (Q56322). This is because the law actually states that it is Sami languages (Q56463) which is the officially designated language, that is the whole language group.
I would do something like this for no label (Q5262464)
Again, note the interpretation of official language (Q23492) as administrative language. In this area both Norwegian (Q9043) and Sami languages (Q56463) has similar status. Souce for the statements would be Forskrift om forvaltningsområdet for samisk språk
Finally I would do something like this for Porsanger (Q483885), which is part of no label (Q5262464)
Something along these lines I guess. Jeblad (talk) 02:53, 26 February 2016 (UTC)
Good ! Thank you Jeblad
I've been thinking on your examples during the week-end and they seem fine but I've got a big and naive question. Why do we need a specific property? what is the benefit? and, as Laddo, said, why not simply use language (P2439)? The more I think about it, the more I tend to believe it would be a good thing to replace all official language (P37) by language (P2439) with qualifiers.
Cdlt, VIGNERON (talk) 19:23, 29 February 2016 (UTC)
In general you can remove a lot of props and just keep a few core props, and then add qualifiers to describe the finer points. This is what reified statements is all about. That would although lead to a lot of statements with repeating qualifiers. The opposite would be to make a very fine grained set of props, but then the set of props would be huge. So in between we try to make a balance. Don't make the props to fine grained, don't repeat the qualifiers to often. (And as I said Q23492 should be administrative language and not official language. That is confusing.) Jeblad (talk) 00:23, 1 March 2016 (UTC)

Non-official languages[edit]

Just been reading through the previous discussions, very interesting stuff. I want to pick up on @VIGNERON:'s comment:

"The more I think about it, the more I tend to believe it would be a good thing to replace all official language (P37) by language (P2439) with qualifiers."

I think that would make sense, since Wikidata is currently lacking language related information for anything that's not official. Spanish is spoken by more than 40 million people in the US but since there is no "official" language in the constitution there's also no link on Wikidata. I found language used (P2936) but that seems rather vague (events? places?) and unused currently. Why don't we use a property which is not based on any legal/official status but just on empirical facts (= number of speakers)?

The number of speakers for a given language in a geographical location would be interesting information on its own. Right now I can get the number of Dutch speakers globally (via number of speakers (P1098)) but not the number of Dutch speakers just in Belgium.

Or perhaps this is already possible? – Jberkel (talk) 17:03, 14 August 2016 (UTC)

Applies to part[edit]

Several items use applies to part (P518) as a qualifier for this property, to describe an official language of a subdivision. Is there any reason why this can't be stored only on the item for the subdivision? Seems redundant. If it is necessary, P518 should be added as an accepted qualifier in the constraints. (Pinging @Thayts, Infovarius, Thryduulf:, who have added P518 as a qualifier to P37.) --Yair rand (talk) 20:19, 12 December 2016 (UTC)

I was thinking about how one would get this information in the infobox of en:Netherlands, as it is now, when using Wikidata for that. I really don't know what the best way to do this is. Should the infobox otherwise iterate over all the defined regions to see if each has P37 defined? (That is expensive too...) Thayts (talk) 21:37, 12 December 2016 (UTC)
@Thayts: Do you think that the items for the subdivisions themselves should or should not also have official language listed? If they should, we have redundant data. If not, articles on the subdivisions would either need to iterate upwards to find languages particularly applying to it, or not rely on Wikidata for it. There are a lot of subdivisions with official languages, and duplicating them all, in data or in infoboxes, could get overwhelming. Perhaps infoboxes like the Netherlands' could be handled manually on Wikipedia's end? --Yair rand (talk) 21:26, 15 December 2016 (UTC)
I think it should at least be in the items for the subdivisions themselves. And I agree you have redundant data then. Would it be an idea to create some kind of a hardlink concept within Wikidata to link from the main item to subitems, leaving the data stored only at the subitems but with the data cached at the main item? Thayts (talk) 17:50, 16 December 2016 (UTC)
I would tend to storing this information in the main item. I know about arbitrary access, but for usual readers of the item a single language could be misleading. --Infovarius (talk) 15:59, 13 December 2016 (UTC)
@Infovarius: I don't think people reading the item directly are a target audience here... --Yair rand (talk) 21:26, 15 December 2016 (UTC)

'De facto' official[edit]

For the record,

There is 4 items with 'de facto' official languages right now : United States of America (Q30), Bosnia and Herzegovina (Q225), Nauru (Q697), Eritrea (Q986). See the corresponding query:

SELECT DISTINCT ?item ?itemLabel WHERE {
	?item wdt:P37 [] .
	?item p:P37 [ ps:P37 [] ; pq:P794 wd:Q712144 ] .
	SERVICE wikibase:label { bd:serviceParam wikibase:language "fr" } .

Try it!

For the moment, I can't see any reason for these claims which seems illogical to me. I've contacted directly Tcp-ip, the person who added these four claims.

Again, it poses the question of how we want to structure all the spectrum of status of languages can have (ping @Jeblad: from previous discussions).

Cdlt, VIGNERON (talk) 10:15, 13 September 2017 (UTC)

Not sure what I'm supposed to do here? Jeblad (talk) 10:57, 13 September 2017 (UTC)
Jeblad nothing Face-wink.svg, as I said it's « For the record ». Cdlt, VIGNERON (talk) 11:27, 13 September 2017 (UTC)
As you can see in my contributions page, I am checking the informations about official languages in all articles about countries. In the cases where an official language is set by the constitution, I am adding the exact article in the references. In some cases, e.g. the USA, there is no law, but a language is the language in which the constitution and all laws are written, the parliament holds discussions, courts hold trials and all communications with the authorities must be held. I think that, though these countries can't be described as having an official language proper, having no statements about the language used by their governments would mean omitting important informations from Wikidata knowledge base. The property "official language" with "de facto" seems to me a good way to explain the situation, in line, for example, with what is meant by "Official language" according to the English Wikipedia:
"An official language is a language that is given a special legal status in a particular country, state, or other jurisdiction. Typically a country's official language refers to the language used within government (e.g., courts, parliament, administration). Since "the means of expression of a people cannot be changed by any law", the term "official language" does not typically refer to the language used by a people or country, but by its government. [...] Many of the world's constitutions mention one or more official or national languages.[...] In countries that do not formally designate an official language, a de facto national language usually evolves."
BTW There are situations that I find rather hard to model. I see that Norway has already been discussed above. Bolivia has the constitution designing many languages as official, but then saying that the central government, as well as the local governments, are required to use only some of them.
Tcp-ip (talk) 22:08, 13 September 2017 (UTC)
@Tcp-ip: first, thank you for these explanations and for your work (indeed there is a lot of work to do).
Then, from the general point of view, it is sadly very true that language are hard to model. Still, I think there is something startling with the current structure (or more exactly the lack thereof) that should be corrected and improved. We put under the label official things that are not really official… Maybe we should change the label (going back to the original label 'national language'? qv. discussion #Label change, is was indeed vague but it's vagueness and flexibility was convenient), use another property (language (P2439) and let qualifiers dealing with subtleties?), or anything else (I'm open to suggestions Face-wink.svg). I don't know, but what I know is that we need more clarity if we want to solve it in a proper and stable way.
From a closer point of view, I fell that the claim doesn't fit the references on Q31#P37 (you changed the value and kept the old references but it doesn't really match). As the situation is complicated, it is even more important than usual to put references. And as there is different point of view for this value, I think we should add them instead of replacing them (which was already suggested by Lavallen on #"No value" as "None"?).
For Bolivia (Q750), maybe you can solve it with applies to part (P518).
Cdlt, VIGNERON (talk) 08:48, 14 September 2017 (UTC)
PS: as (P794) is not an accepted qualifier right now, any objection to add it? (likewise for applies to part (P518), both are already more commonly used than the current accepted qualifiers!)