Wikidata:Requests for comment/Personal names

From Wikidata
Jump to navigation Jump to search
An editor has requested the community to provide input on "Personal names" via the Requests for comment (RFC) process. This is the discussion page regarding the issue.

If you have an opinion regarding this issue, feel free to comment below. Thank you!


THIS RFC IS CLOSED. Please do NOT vote nor add comments.

(Some discussions relating to this topic: 1 2 3 4 5 6 7 8 9)

Personal names are a rather complicated issue for Wikidata, and setting up a standard data structure for them is probably going to be pretty difficult. I'm hoping we can come to a consensus on how to deal with some of the relevant issues. To start off, as a general question:

Should name data be included at all?[edit]

Assuming we do include name data...

Dealing with different name components[edit]

The following are among the numerous things that can make up a person's name that we can include data for:

And then we have things like pseudonym (Q61002), art name (Q39646) (could possibly be considered the same as pseudonyms) and nickname (Q49614).

One way to deal with these would be to create a separate property for each name component type. Alternatively, we could just create "Name" (or "name part") as a property, and set the type of name by using a qualifier.

Create separate properties for each name type[edit]

Use qualifiers to indicate name type[edit]

Even if used as a qualifier there is still a requirement for properties for each of these to be used as qualifiers and these properties need to be named. --Filceolaire (talk) 01:01, 19 June 2013 (UTC)[reply]

Discussion[edit]

Which name types to include?[edit]

Discussion[edit]

For some of these types, such as given names and surnames, a person can have multiple of one type at a time, in a set specific order. Two ways of dealing with this are to either have a qualifier to specify name position with a number, or to create separate properties/type target items ("first given name", "second given name", etc.). (If your preference on this depends on the resolution of the previous issue, please say so in your comment.)

I have added a third option below Filceolaire (talk) 21:31, 19 June 2013 (UTC)[reply]

Specify name position with numerical qualifiers[edit]

<Item A> 

   <GivenName> = AAAA
       <Name No> = 1
   <GivenName> = BBBB
       <Name No> = 2
   <FamilyName> = aaaa
       <Name No> = 1
   <FamilyName> = bbbb
       <Name No> = 2

or

<Item A> 

   <Person Name No> = 1
       <GivenName> = AAAA
       <FamilyName> = aaaa
   <Person Name No> = 1
       <GivenName> = BBBB
       <FamilyName> = bbbb

--Napoleon.tan (talk) 23:40, 18 June 2013 (UTC)[reply]

I think the first option would be much better. --Yair rand (talk) 00:12, 19 June 2013 (UTC)[reply]
The first option is appropriate for people with multiple names 'Charles Gordon Patrick Saxe Coburg Gotha' while the second option is better for people who change their name. Note that we need to specify how all the elements are ordered - some cultures put the family name first. --Filceolaire (talk) 01:12, 19 June 2013 (UTC)[reply]

Use separate "first given name", "second given name", etc. properties/items[edit]

Use both the full name and the components[edit]

A third option is to include the full name with the component names as qualifiers. e.g.

name<Sir James Alexander Gordon FitzAllen-MacHeath OBE> (string datatype)
given name<Alexander> (item datatype)
given name<Gordon>
given name<James>
family name<FitzAllen> (item datatype)
family name<MacHeath>
Honorific<Sir> (item datatype)
Decoration<OBE> (item datatype)
end date<1979> (date datatype)
Comment<name changed by deed poll in 1979> (multilingual string)
name<Dizzy MacAllen> (name 2 for the same person)
given name<Dizzy>
Family name<MacAllen>
Start date<1979>
comment<This name used as a stage name since 1965. Name adopted officially in 1979>

Does this work? Filceolaire (talk) 21:26, 19 June 2013 (UTC)[reply]

 Support --Filceolaire (talk) 21:35, 19 June 2013 (UTC)[reply]

How would you indicate the order of the given names or family names? --Yair rand (talk) 23:37, 25 June 2013 (UTC)[reply]
It's shown in the <name> string. I couldn't think of a way to describe it semantically. Filceolaire (talk) 20:19, 3 July 2013 (UTC)[reply]

Discussion[edit]

Hyphenated names[edit]

There needs to be some way to indicate that two (or more) names are attached together with a hyphen or otherwise, and I have no idea how this could be done. I suppose we could just consider a hyphenated name to be one unit, but I really don't think that would be optimal. Any ideas? --Yair rand (talk) 01:30, 18 June 2013 (UTC)[reply]

See my proposal above. Filceolaire (talk) 21:42, 19 June 2013 (UTC)[reply]
For Catherine Scott-Clark, Q88312923, I plan to create "Scott-Clark" as a single surname. I can do this with what I know how to do. If Wikidata has a better way, I hope someone will fix this and let me know. Thanks, DavidMCEddy (talk) 20:58, 22 March 2020 (UTC)[reply]

Which data type to use for names?[edit]

(A lot of the above was written with the assumption that the item datatype would be used. Feel free to add sections about how to deal with issues that only come up if a different data type is used.) One option for the datatype of each of these properties/this property would be the Item datatype; the value could be set to the item for the name, with each name having its own item. In this discussion, User:Šlomo proposed that MultilingualTextValue be used instead, saying "MultilingualTextValue seems more suitable since the use of personal names' translations/transliterations is different for particular names, languages and even persons and can't be simply linked to a Wikidata item dealing with name in general.", and User:Beta16 said that "String or MultilinguaText are better." User:Snipre opposed having data for specific parts of names at all, saying "Better to have one property "Full name" including last and surname".

Use Item datatype[edit]

Use MultilingualTextValue datatype[edit]

Use Monolingual Textvalue datatype[edit]

For the full name - see my proposal above. Filceolaire (talk) 21:46, 19 June 2013 (UTC)[reply]

Use String datatype[edit]

 Support Items are about concepts, not words, but names are words. That creates two sorts of issues:

  • practical: it is very sensitive to small imperfections like approximate interwikis. If someone is called en:John, should we use Q397210 that links to both en:John and ja:ヨハネ ? I dont think so, as ja:ヨハネ appears to include de:Johannes and fr:Jean that are different names for most practical purposes.
  • conceptual: if an American of Korean descent is called "Lee", should her surname be en:Lee (Korean name) or en:Lee (surname) ? I would say en:Lee (surname) , because really she is just called "Lee" no matter whether her grand-father was called 이, Lee, or 李. And if so, "Lee" is really just a meaningless string, and it does not make much sense to use items. --Zolo (talk) 05:15, 19 June 2013 (UTC)[reply]
    I would think that if the names an article describes don't match up to that of the rest of the articles linked to by the item, the sitelink could be removed and a separate item created for the article. --Yair rand (talk) 23:39, 25 June 2013 (UTC)[reply]
    My mistake, ja:ヨハネ actually does not include Johannes. Still even if my first point is solved, I do not think it adequately solves the problem, as point 2 remains (two names that are spelled the same way can have different items). We can probably create items to solve that as well, but I am not sure it really makes sense. Why should we use items ? The reasons I see are:
  1. attach a meaning, like "he is called John - a name given in memory of John the Baptist". But in many cases, that is raher tricky. What is common between the "Lee" of en:General Lee and the Lee of en:Lee Kuan Yew ? Sure, in this case, we could make a bit of genealogical research and find meaningful items, but doing that for every person seems next to impossible to me.
  2. make translations easier. But I am not convinced that it is better done through items than through templates. For instance, the same Chinese name will be transcribed differently in English depending on whether the person is Chinese or Taiwanese. --Zolo (talk) 06:40, 26 June 2013 (UTC)[reply]

Discussion[edit]

Start/end dates[edit]

Should start date and end date qualifiers be used for a person's original names and names that a person had until death, respectively? Over at WD:Requests for comment/Time DataType Properties, it was suggested that (within the general topic of start/end dates) start and end dates should be given even if they go all the way to the start or end of a person's life, as otherwise it would be impossible to tell if perhaps the data hadn't been added yet. In my personal opinion, even if that is to be the standard, names should be an exception, as it is extremely common that names don't change. Thoughts? --Yair rand (talk) 01:30, 18 June 2013 (UTC)[reply]

Pseudonym[edit]

Pseudonyms are a bit more complicated than the other name elements, as they frequently take the form of full names themselves, complete with pseudonymous given names and surnames. Even more problematic is that they occasionally get inserted into a person's real name, as in the case of Joanne Kathleen Rowling. Any ideas as to the best way to store this data? --Yair rand (talk) 01:30, 18 June 2013 (UTC)[reply]

People with different names in different languages[edit]

Some people have had different names in different languages, while others have been referred to by different names in places where other languages are spoken, without it ever having actually been their name. How should such information be included, if at all? --Yair rand (talk) 01:30, 18 June 2013 (UTC)[reply]

Use some string-related type, with language qualifiers :). I do not think there is any other way to say that Aristotle is called Aristote in French. --Zolo (talk) 19:48, 19 June 2013 (UTC)[reply]

Fictional character names[edit]

Should fictional characters have name data provided, and if so, should they use the same type of naming system as people? If they do, and the item type is chosen for name properties, should separate items be created for fictional names even if only held by a single character? Also, how should characters that have different names in different language versions be dealt with? --Yair rand (talk) 01:30, 18 June 2013 (UTC)[reply]

Name signs[edit]

We might be best off just ignoring these until real support for SignWriting is available... --Yair rand (talk) 01:30, 18 June 2013 (UTC)[reply]

In principal this depends on whether the name sign is considered to be a transliteration of a name (sign name is a qualifier to the name) or a separate name (sign name is a separate name entry) Filceolaire (talk) 21:51, 19 June 2013 (UTC)[reply]

Exceptional names[edit]

Occasionally, there are names that don't fit into any normal system. One example is w:Prince (musician), who changed his name to a symbol that can not be pronounced or represented in unicode. Should such cases be simply ignored, or should we try to find some way to include it? --Yair rand (talk) 01:30, 18 June 2013 (UTC)[reply]

Titles, honorifics[edit]

The properties noble title (P97) and honorific prefix (P511) have already been created. Are these a full solution to titles and honorifics? Also, do they overlap? --Yair rand (talk) 01:30, 18 June 2013 (UTC)[reply]

They don't overlap. You could be styled "the right honourable" but not "third Marquess Calgary", the former being an honourific prefix and the latter being a noble title. The honourific prefixes get shoved in front of the name, the noble title replaces the name itself (or is used in conjunction with, e.g. John Snow, fifth Duke of Windsor). Ajraddatz (Talk) 14:12, 18 June 2013 (UTC)[reply]

Anything else?[edit]

How about the pronunciation of names? actually this is a general issue, even for name places? Should this be just a qualifier? For names that are not latin characters, maybe romanticized pronunciation qualifier could be used. --Napoleon.tan (talk) 15:58, 18 June 2013 (UTC)[reply]

I think that pronunciations are more within the scope of Wiktionary than of Wikipedia-linked Wikidata items. Names that are not in Latin characters would typically have a transliterated label in languages with other scripts, so that doesn't seem to be an issue. --Yair rand (talk) 23:43, 25 June 2013 (UTC)[reply]

I've requested a new bot task related to this RFC at Wikidata:Requests for permissions/Bot/SLiuBot 3. Feel free to leave your comments. --Stevenliuyi (talk) 18:40, 22 June 2013 (UTC)[reply]

Single word names[edit]

Some culture doesn't have surnames -- their names consist only one word. (w:Mononymous person). Bennylin (talk) 16:12, 5 September 2013 (UTC)[reply]

Closing this RfC[edit]

I created two of the discussed properties today. You can see the reason here. So the decision was made more based on the properties discussion itself, than this RfC, which didn't proceed. Hope everyone can understand my decision. --Nightwish62 (talk) 20:29, 31 July 2013 (UTC)[reply]

Retrieved from "https://www.wikidata.org/w/index.php?title=Wikidata:Requests_for_comment/Personal_names&oldid=1140884717"