Open main menu

Wikidata:Property proposal/Nationality

< Wikidata:Property proposal

nationalityEdit

Originally proposed at Wikidata:Property proposal/Person

   Not done
Descriptionwhat nationality (Q231002) is person identified with
Representsnationality (Q231002)
Data typeItem
Template parameter"nationality" in en:Template:Infobox person; "nationality" in c:Template:Creator
Domainpeople
Allowed valuessubclass of nationality (Q231002), like Ukrainians (Q44806), Indonesians (Q14535097) or Canadians (Q1196645)
Example
SourceRKDartists (Q17299517), Union List of Artist Names (Q2494649) and others
Robot and gadget jobsbot import e.g. from RKDartists (Q17299517) and Union List of Artist Names (Q2494649)
See alsoethnic group (P172), country of citizenship (P27), languages spoken, written or signed (P1412), native language (P103), allegiance (P945)
Motivation

Item descriptions of people often start with phrases like "German painter", "Welsh writer" or "Japanese Ukiyo-e artist". The first word describing a person is often nationality and most biographical infoboxes on Wikipedia projects list it. Also majority of biographical databases list it, see for example RKD or ULAN. Currently there is no good way to figure out nationality of many people on Wikidata. Quite often ethnic group (P172), country of citizenship (P27) or native language (P103) are synonymous with nationality or can be used to decipher it. country of citizenship (P27) is often too precise and lists multiple items for different periods of country's existence (see Adolf Hitler (Q352)). ethnic group (P172) is also problematic as it often does not define a person. We refer to Barack Obama (Q76) as "American president" not "Kenyan American President". native language (P103) often determines nationality: mostly Poles (Q1026) speak Polish (Q809), but not all speakers of Spanish (Q1321) are Spaniards (Q160894). See also this discussion. --Jarekt (talk) 14:53, 9 August 2017 (UTC)

Discussion
  •   Comment so having grown up in Canada and hearing of many people born there (or who lived there) who moved to or worked in the US and then became celebrities, I always found it amusing that in Canadian media the person (in a report about them having won an award or something) was referred to as "Canadian", and yet there would be no mention of nationality or there would be an assumption of "American" on the US side. US media seemed more careful to call them Canadian in the case of negative news (arrest etc.) The quintessential example I can remember is Alexander Graham Bell (Q34286) - born in Scotland, later lived and died in Canada, but he worked and is credited with inventing the telephone in the US. To Canadian's, Bell is Canadian. Few Americans think of him that way. And the enwiki page lists him as "Scottish-born". None of those three "nationalities" is listed on the current wikidata page. I think this is a rather slippery concept, likely to be historically and culturally contingent and readily challenged. On the other hand, if you have a reliable source stating the nationality in some clear way, I guess it's fine to have it here in wikidata. I just have some doubts about whether it will be a useful generally single-valued property. ArthurPSmith (talk) 19:42, 9 August 2017 (UTC)
  • How is this different to country of citizenship (P27)? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:28, 10 August 2017 (UTC)
  •   Comment While I understand the rationale of this request, I am dubious about the adoption of such property. Yes, we have clear cases (i.e. Nazario Sauro, Austrian Empire-born but Italian, condemned for treason by the Austrians but honoured as hero by the Italians), but beyond these obvious cases I am afraid that this property is prone to host a series of POVs that the disputations on Commons over the borders of Kashmir or the former republics of Yugoslavia are going to look something like a kindergarten fight on comparison. -- Blackcat (talk) 15:22, 13 August 2017 (UTC)
    • Blackcat, Many sources, like RKD or ULAN, list nationality, so we should not have namy POVs. Also most biografical infoboxes use it, and although country of citizenship (P27) works for many contemporary people and some countries, in many cases there is no way to easily and predictably look up people's nationality based on current data model. Most nationalities are not controversial and Commons Creator templates and wikipedia infoboxes hosted them for years without much fighting. Most Wikipedia biographical articlses state nationality and occupation in the first sentence, also without much controversy in most cases. --Jarekt (talk) 12:06, 29 August 2017 (UTC)
  •   Weak support Not a very clean property, but there are reliable sources, it's useful to have this information (even if it's badly formalized) and I don't really see a way to formalize it better, it's mainly about historical persons and here history is hard to put into well-defined terms of a database. Actually, this property could be a means to clean ethnic group (P172) from bad uses. --Marsupium (talk) 17:19, 29 August 2017 (UTC), 19:11, 29 August 2017 (UTC)
  • Note: Project chat discussion has been archived. See Wikidata:Project_chat/Archive/2017/08#How_to_look_up_persons_nationality.3F.
  •   Oppose. Nationality is incredibly ambiguous. If we have a source saying "this person is an Arab", "...is Chinese", "...is Scottish", or "...is a Muslim", should we put that as nationality? Some people may self-identify their nationality as such, but not all people who consider themselves part of such groups would consider their membership in the group to be a person-nation relationship. --Yair rand (talk) 02:52, 30 August 2017 (UTC)
  •   Oppose I don't think there is sufficient clarity about how this would be used. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:55, 11 September 2017 (UTC)
Oh I can clarify that. We can populate it from sources like RKDartists ID (P650), Benezit ID (P2843) or ULAN ID (P245), which all state nationality, and I will call it from c:template:Creator and it can be used by en:Template:Infobox person "nationality" field. --Jarekt (talk) 12:28, 21 September 2017 (UTC)
  •   Oppose Nationality is too complex and at the same time too vague of a construct to be captured like this. We may limit it to its legal use, but that would hardly differ from our current citizenship property. For the other common use of the word we already have our ethnic group property. – Máté (talk) 21:00, 21 September 2017 (UTC)
  •   Support I struggle to correctly identify artists using "country of citizenship" especially for prior centuries, when the concept "Germany" or "Italy" (for example) meant something different from the modern country. The property is stated in reliable sources such as ULAN and RKDartists (mentioned above), and using it in Commons is standard; therefore, we should have a property for it and we should reference a source for the value. (And they do sometimes disagree, but as Wikidata is an aggregator that's perfectly okay.) - PKM (talk) 23:56, 21 September 2017 (UTC)
I don't quite see your point. The concept of nationality is way more modern than that of citizenship (and ethnicity). Citizenship has been around since ancient times, while the first nation state came about in the 16th century. – Máté (talk) 04:39, 22 September 2017 (UTC)
@Máté: and others: To turn the table how else shall we state that Albrecht Dürer (Q5580) is German or that Leonardo da Vinci (Q762) is Italian? Shouldn't we have it at all? In fact, we already have that information on Wikidata in unstructured form in the descriptions in dozens of languages (not to mention all the country of citizenship (P27) statements where the person died (centuries) before the creation of the object state which might express some need). I'd prefer to get it in an at least slightly more structured form that also can be queried. I admit it isn't a very clear concept. That's why I'd support strong conditions to use the property, e.g. disallowing not sourced (and imported from Wikimedia project (P143) sourced) statements. But still it's a historic concept, Albrecht Dürer (Q5580) and Leonardo da Vinci (Q762) were also considered in their times to be German and Italian respectively (although not calling that nationality). --Marsupium (talk) 16:46, 22 September 2017 (UTC)
I think the alternative is to work off of the fact that Albrecht Dürer (Q5580) was a citizen of Duchy of Bavaria (Q47261), one of the historical German states, and that Leonardo da Vinci (Q762) was a citizen of Republic of Florence (Q148540), one of the historical Italian states. Better than creating a vague new biographical property, perhaps we should work on using an existing or possibly a new property that will affiliate Duchy of Bavaria (Q47261) with a Q item for "historical German states" and Republic of Florence (Q148540) with a Q item for "historical Italian states". That way it is structured, queryable, and you could even display it in infoboxes if you like.--Pharos (talk) 17:45, 22 September 2017 (UTC)
Often their isn't any relationship between a person and state-like thing in medieval and early modern period beyond the residence of the person on the territory of the state-like thing. Albrecht Dürer (Q5580) actually is an exception, it might be said that he was a citizen of Free Imperial City of Nuremberg (Q117020) (not Duchy of Bavaria (Q47261), Nuremberg didn't belong to Duchy of Bavaria (Q47261)/Kingdom of Bavaria (Q154195) until 1815) since Nuremberg had something like a citizenship, but that is a special case among the "historical German states". --Marsupium (talk) 14:34, 23 September 2017 (UTC)
@Marsupium: If there is no relationship at all to a state-like thing, we're not talking about nationality, we're talking about ethnicity - for example, I don't think we should say an Italian-speaking person born in the Duchy of Bavaria was of "Italian nationality". I think the vernacular use of country of citizenship (P27) to include countries that didn't have citizenship as such is correct, and perhaps it should be relabeled "country of citizenship or nationality". Leaving aside the question of whether proper citizenship existed, wouldn't both Nuremberg and Bavaria be considered "historical German states"?--Pharos (talk) 04:43, 14 October 2017 (UTC)
  •   Comment It seems like the proposer is really looking for an answer to, "What country would this person be a citizen of today?". That's an interesting question, and can be queried in various ways, but bringing in the "member of nation" concept as orthogonal to "member of state" is not I think the way to do it. The examples given seem to all be based on either ethnicity, or citizenship, or simply projecting a current state backward in time.--Pharos (talk) 16:57, 22 September 2017 (UTC)
Pharos, I am not an expert on "Nationality" it is just basic information about a person that most sources list and most infoboxes display, that we currently do not store in accessible format. I personally do not care how people assign those labels I just want to store them. The theory that they are assigned based on "what country would this person be a citizen of today?" is interesting, but I do not think it is correct. Nam Yeongun (Q6780884) is "Korean" ruler but we would not assign him to neither North Korea (Q423) or South Korea (Q884). --Jarekt (talk) 17:17, 22 September 2017 (UTC)
@Jarekt: In this instance it's perhaps, "what country would this person be a citizen of today, if not for the historical accident of division?". This is a time-reversed case of Marsupium's examples I just replied to above, about "historical German states". Here, I think we can rather approach this by the concept of "historical Korean states", of which there were many before Japanese rule, and currently two after. So I think it is better to query if someone is a citizen/national of a state like Goguryeo (Q28370) or South Korea (Q884) as "historical Korean states".--Pharos (talk) 04:43, 14 October 2017 (UTC)
  •   Comment. I have never understood why nationalities or citizenships should be dealt with through a property linking to countries. The property should link to items dedicated to nationalities or citizenships, such as what French nationality law (Q1164879) is intended for on the French Wikipedia (but not the English, which is about the legislation, and should technically be moved to a separate item). Let's create a nationality item for each one that might be necessary and then move all declarations from country of citizenship (P27). Thierry Caro (talk) 06:43, 23 September 2017 (UTC)
@Thierry Caro: That isn't a support, is it? --Marsupium (talk) 06:25, 27 September 2017 (UTC)
Not exactly. You suggest we use items such as 'Germans'. I suggest we use items such as 'German nationality', most of which have yet to be created. Thierry Caro (talk) 06:35, 27 September 2017 (UTC)
Thierry Caro, I would support creation of separate "nationality" items (like Hui nationality (Q5934909)) to use for this property. --Jarekt (talk) 13:30, 27 September 2017 (UTC)
  •   Comment The Union List of Artist Names (ULAN) uses the term very broadly: “In ULAN "nationality" is a shorthand for nationality, culture, race, ethnicity, religion, even sexual orientation: a significant social grouping or designation of the agent.” (http://vocab.getty.edu/doc/#ULAN_Nationalities (co-)authored by Vladimir Alexiev; 2267 possible values compared to currently 279 allowed by country of citizenship (P27)’s one-of constraint (Q21510859)). That’s very pragmatic and hence viable here, too, in my eyes; all the more so since we (can) directly use the data of the ULAN. I guess RKDartists follows a similar approach although I didn't find any documentation on that. (RKDdata, you probably know more?) Regards, --Marsupium (talk) 06:15, 27 September 2017 (UTC), 08:36, 27 September 2017 (UTC)
We might even make this an umbrella subproperty of (P1647) ethnic group (P172) and country of citizenship (P27) to be used wherever possible, so (in Europe) for 19+th century people. --Marsupium (talk) 06:25, 27 September 2017 (UTC)
  •   Comment There is an article on Mozart's nationality with interesting discussion about the concept of "Nationality". In short Mozart lived in today's Austria, but was referring to himself as "German". --Jarekt (talk) 13:40, 27 September 2017 (UTC)
  • This property is different from Citizenship as clearly demonstrated above. Country borders shift often, and populations don't always move with them. So eg you got Bosnians, Albanians and Serbs in Kosovo, etc.
The POV concerns (there may not be reliable sources and that there will always be political arguments) are IMHO not too serious:
- If someone wants to call Obama "Kenyan" and "American", that's fine (especially if that's Obama himself).
- It's wholly natural that Scotts would want to call Bell "Scottish", Canadians want to call him "Canadian" and the Americans call him "American". IMHO there's no problem to capture all 3 (backed by sources), then he'd appear on indexes of "famous Scottish people" etc
- By the same token, we call the St Cyrill & Methodius brothers "Bulgarian" while the Macedonians call them "Macedonian" (but we Bulgarians are right ;-)
ULAN is particular to capture mostly the self-determination of the person. It won't record someone as Gay if that person hasn't said it publicly.
@Marsupium: "umbrella subproperty of Ethnic group and Country of citizenship" Guess you mean "super-property"? But Citizenship is very different: Nazario Sauro would never call himself Austrian: maybe even while he held an Austrian passport, he despised everything about that culture.
  Oppose unfortunately I have to oppose this in favor of Ethnic Group, because no clear case about the distinction between them has been made. Just add "Nationality" (and "Culture" and "Peoples") as aliases of that prop "Ethnic Group" --Vladimir Alexiev (talk) 17:14, 29 September 2017 (UTC)
Sigmund Freud was born in the Czech Republic, is considered Austrian but was technically an Austro-Hungarian citizen, at least until 1918. That's the problem with the former Empires... -- Blackcat (talk) 18:37, 1 October 2017 (UTC)
  • @Jarekt: would you agree to lump Ethnic Group/Nationality/Culture/Peoples in one prop? ULAN basically takes this approach --Vladimir Alexiev (talk) 17:17, 29 September 2017 (UTC)
Vladimir Alexiev I am fine with that. I just would like to store the adverb stored in ULAN under "Nationalities" label somewhere. Looking at Wolfgang Amadeus Mozart (Q254) most databases store information that he was "Austrian" is different fields:
So others store that information in fields called either "Nationality" or "Country". Here I opted for "Nationality" as less confusing than country of citizenship (P27) since we already have that property but use it differently. --Jarekt (talk) 19:22, 29 September 2017 (UTC)
  • As per @Jarekt:, I propose to give "Ethnic Group" aliases "Nationality" and "Culture" and "Peoples", and reject this extra prop proposal. ("School" is not an appropriate alias because it's specialized to arts and implies a more self-recognized and intellectual grouping) --Vladimir Alexiev (talk) 17:42, 2 October 2017 (UTC)
I would be OK with adding "Nationality" alias and using ethnic group (P172) to store content on Creator template "Nationality" field. I will post information about it at Property_talk:P172 to make sure users working with that property see those as equivalent. --Jarekt (talk) 19:00, 2 October 2017 (UTC)
Added aliases to ethnic group (P172). Those were already present, confirming we did the right thing: peuple (fr), национальность (ru). @Jarekt:, please withdraw the proposal --Vladimir Alexiev (talk) 10:03, 3 October 2017 (UTC)
Vladimir Alexiev, "Ethnic Group" works well in Europe but makes little sense in North American context. For example Barack Obama (Q76) based on his "Ethnic Group" he would be described as "African / multiracial / Kenyam - American president", which might be true, but most people describe him as "American president". Similarly Donald Trump (Q22686) would be described as "white American president" or "German-American president". Similarly in Canada: look at ethic makeup of Canadians according to Wikidata: most of them are "African Americans". There is almost as many white people (Q235155) as Canadians (Q1196645). According to Wikidata, in Australia there are as many African Americans (Q49085) as Aboriginal Australians (Q12060728) and there is much more Serbs (Q127885) than Australians (Q1318423). Should I add Americans (Q846570) as ethnic group (P172) of Obama as you suggested in the examples of ethnic group (P172)? --Jarekt (talk) 12:52, 3 October 2017 (UTC)
SELECT ?ethnicityLabel ?ethnicity (COUNT(?person) AS ?count)
WHERE
{
  ?person wdt:P27 wd:Q16 .
  ?person wdt:P172 ?ethnicity .  
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
}
GROUP BY ?ethnicityLabel ?ethnicity
ORDER BY DESC(?count)

Try it!

We discussed this above: if Obama wants to call himself Kenyan and American, that's fine. If others want to call him other things, that's less ok, but both "ethnicity" and "nationality" will always be subject to controversies. "President" is neither. I'm against precoordinated groups like "Kenyan-American" because how many combinations we'll need? Those Serbs in Australia may well feel themselves to be both: Serbian and Australian. The point is, I don't see any fundamental difference between ethnicity and nationality, and "ethnic group" already has a label "nationality" in Russian. And I think you agreed.
About the examples: I've already added all examples from here to there. --Vladimir Alexiev (talk) 13:32, 3 October 2017 (UTC)
  •   Support Given that this gets covered in a lot of existing databases. We want to be a database that tells people what other sources say and plenty of sources are available. ChristianKl (talk) 13:00, 18 October 2017 (UTC)
  •   Oppose To the extent that "nationality" is a valid concept and is covered in existing databases, it is a synonym for country of citizenship (P27), which could perhaps be renamed to "country of citizenship or nationality" to cover instances where the modern concept of citizenship isn't involved. To create it as a separate concept that is vaguely ethnic-based is a troubling mess likely to lead to nationalistic edit-warring. As explained above, I believe the situation of people who were subjects of say, historical Italian or German states, can be dealt with by applying properties to the states themselves, not to the individual historical people.--Pharos (talk) 17:11, 19 October 2017 (UTC)
    • nationality is not synonym for country of citizenship, see en:Multinational state. --Pasleim (talk) 19:50, 19 October 2017 (UTC)
      • It is not a synonym, it is a vague term that often means either country of citizenship (P27) or ethnic group (P172), but for biographical database purposes it usually means the former. I'm not saying that the concept of a nation is invalid, certainly people write about the national nature of various states, but the idea that for an individual person, "member of nation" should be a new and separate thing to "member of state" and "member of ethnic group" is just not meaningful.--Pharos (talk) 04:25, 20 October 2017 (UTC)
I agree. Pharos, just look at Adolf Hitler (Q352): most sources describe him as "German", however not a single one of the 6 country of citizenship (P27) entries is "Germany". We need a single field where we can save information that he is German and I do not think anybody will be happy with us if we replace all those country of citizenship (P27) with :Germany". user:Vladimir Alexiev made quite convincing argument that we can use ethnic group (P172) for that: it works well for most European countries and we can rename it to "ethnic group or nationality" so we can use it for countries like Canada or Australia. However you nuked all those efforts. --Jarekt (talk) 03:08, 20 October 2017 (UTC)
Both country of citizenship (P27) and ethnic group (P172) are well-defined concepts, and using the latter one for Adolf Hitler (Q352) is already totally appropriate under current policy. "Nationality" is a vague term that is sometimes used for both, and there is overlap between the two, especially in Europe. However, "Canadian" is not an ethnic group, and neither is "Kenyan", the whole world is indeed not built like European nation-states.--Pharos (talk) 04:25, 20 October 2017 (UTC)
@Pharos: Yeah, why did you nuke my examples? You say "inappropriate examples trying to further an active debate": how exactly saying that Durer and Hitler are Germans is trying to further one argument or another?? --Vladimir Alexiev (talk) 12:57, 20 October 2017 (UTC)
@Vladimir Alexiev: I don't at all object to Durer and Hitler as ethnic group (P172) "German", although one should generally first include the property on their items, before putting them in as examples. I do object strongly to "American" and "Australian" and "Indian" as values for ethnic group (P172), as these are cases of country of citizenship (P27). I think we are also encountering a linguistic difference in what the cognate terms of "nationality" mean in different languages, so perhaps that is causing part of the confusion between us.--Pharos (talk) 15:25, 20 October 2017 (UTC)

  Comment Recently User:Epìdosis was removing [1][2] referenced statements about people nationality from country of citizenship (P27) property, based on the fact that Italy (Q38)inception (P571)  2/6/1946 so "Italy" did not exist beforehand. I do not agree with it, but if we narrow the definitions of current country items to only include current political system than we are going to loose a lot of the nationality information currently stored on Wikidata, and used on other projects like Commons. Vladimir Alexiev proposal of broaden the definition of ethnic group (P172) to include nationality was opposed by Pharos (see discussion above). So at the moment the concept of nationality, which is used by infoboxes on many projects relying on Wikidata, and used by many other institutions we can use to reference the statements, does not have a designated property on Wikidata. It is waste of my and other volunteer's time to be adding references to claims of nationality, like here, and for those claims and referenced to the thrown away because someone decided to narrow down the concept of what Italy (Q38) is. --Jarekt (talk) 12:33, 2 November 2017 (UTC)

I'm glad ethnic group (P172) has alias "nationality", and I hope nobody deletes it from there.

  • @Pharos: said "Canadian is not an ethnic group", but let me tell you something: I'm proud to be Canadian (and Bulgarian), because Canadians are nice friendly people (just go see Canada Day celebrations to agree with me) and Canada is a great country.
  • It doesn't matter if I have a Canadian passport (9 of the 12 years I lived there I did not: but for at least 8 of these years I was proud to be Canadian).
  • How can I record this about myself? What prop would you use for "proud to be X"? I care not about your "linguistic differences"
  • If you want to get beaten up, go to a hockey or soccer game in Canada and try to explain to the fans that "Canadian" is only a citizenship ;-) --Vladimir Alexiev (talk) 18:30, 15 November 2017 (UTC)