Forums: Index > Watercooler > Person name properties

This thread is intended to pull together two separate threads that are highly inter-related. Formerly, this was in response to Forum:Dutch patronyms are not given names. It covers issues discussed in Forum:Middle names as well.


I have seen the term "patronomic" used as a genus term for both patronym and matronym, but we could name this m-patronym or matronym-patronym for such patronomizations. You don't have to go far from the Netherlands to find matronym use (common in Scandinavian countries). Although they are in disuse in much of Europe, a patronomic is commonly heard in my household. It is the proper way to address a senior adult with whom one is familiar. While it is customary for Russians to see a "middle name" on a form and use the patronym, for cultures like Iceland or Burma, there is no family name and the patronomic is the surname.

There are so many other weirdnesses with names that we should probably produce a help article to give guidance on these issues. A patronomic might be functionally equivalent to a middle name in Russia, but not in Iceland or Burma where it is a Surname. There are other cases where surname is not equivalent to the family name. In slavic languages and others that are highly inflected languages (have many word forms), the surname can be treated as an adjective that is alterred for agreement with gender of the person. For instance, the first female in space was Valentina Tereshkova- her family name is Tereshkov. Other cases: Polish ~ Phlox 18:36, 26 July 2009 (UTC)

Greek similarly: a Facebook friend of mine gives her married surname as "Bulukaki" but her husband's is "Bulukakis". — Robin Patterson (Talk) 12:48, 27 July 2009 (UTC)

Conventions on patronomics, family, middle and last names

The following needs to be noted in our documentation:

  • "A surname is a name added to a given name and is part of a personal name." (wikipedia definition). In more formal terms, in the context of familypeidia, property:surname has the highest taxonomic rank for a person name. The second order taxonomic rank is held by the property:given_name. The third order (tertiary) rank is held by property:middle_name. The general guidance is that a name should be placed in the property that corresponds to the taxonomic rank that the culture would assign to the name. Properties family_name and patronomic have no taxonomic significance. If the values for them are to have such significance, they must also appear in the surname, middle_name or given_name properties.
  • Surname is not the same as family name. (EG: Patronomics are not family names, and inflected surnames can be orthographically different than family names, and systems that allow compound surnames using as many as six family names, for instance the Spanish system requires a "double-barrelled name" but allows family names from mother father and different grandparents). This is correctly explained in wikipedia's current article on family names, but our current copy equates family name with surname in the topic sentence. Practical impact: Valentina Tereshkova should be included in category:Tereshkov (surname).
  • Surname is not the same as last name. For Chinese names, it comes first, even in western script, for instance wikipedia:Lee Kuan Yew. Such chinese ordering is common practice in non chinese cultures of eastern asia.
  • Surname is not the same as sort name. Tereshkova should be sorted under Tereshkov. For Portuguese surnames (double system), the last surname is the one to sort on, not the first.
    • familypedia convention: "family_name", if present should be used in place of surname for sorting purposes. (This is how we get brothers and sisters from slavic countries into the same surname category).
    • familypedia convention: Surname is a required field. If the patronomic or family_name is filled out, the user must still specify surname even though it may be the same. (This is because possible inflections of the family name or whether the patronym used in place of middle or last name is variable between cultures). If there are multiple names in the surname, they should appear in the display order expected for reports.
    • No assumption should be made that a surname category is the same as a family name category or vice versa. A surname could be a patronomic.
  • "middle name" refers to taxonomic rank only, not the meaning it commonly has in western europe of a secondary given name, nor the practice in slavic countries that it is a patronym.
    • familypedia convention: If patronomic is specified, no assumptions are made about how it is used. If it is intended to affect display (eg. in showfacts biography) it should be included in property:full_name in the correct position. If it is expected to affect sort sequence, then it must also be coded with either the surname or middle_name properties, depending on its taxonomic rank for that culture. In cases where there are multiple patronomics (for example [[wikipedia:arabic name#Arab family naming convention|arabic names), all should appear in property:full name, but only the most significant should be used in property:patronomic. Others may appear in alternate names if desired.
  • Father's name can be very different from the son's name that is based on it. eg: Magnús Eiríksson's father was Eiríkur Grímsson.
    • familypedia guidance: Template and bot writers need to understand that assumptions about person names are error prone. Eg- that fathers and children will appear in the same surname category.

Note that encoding patronomic does not affect the display of an article, and this has search relevance only. Users may search for Matronyms or Patronyms by querying the patronomic field.

Encoding family_name does affect the display of the article (sorting eg in the case of Tereshkova).

Disambiguation: 100 Chinese surnames comprise 85% of the population of China. Even using middle names, the numbers of collisions for 1.3 billion living individuals and just 3 generations of ancestors will be exceptionally high.

  • Proposal: Extend the genealogical convention of specifying place with a name. EG: John Smith (1830) born Denver, Colorado becomes article John Smith (1830) (Denver).

Western name: It is common in Eastern asia for individuals to have both a given name and a western given name (for instance, Lee Kuan Yew, Harry. Until a substantial case is made for storing such western names as a separate property, the western name should be stored as one of the alternative_names for the individual.

Template writers are discouraged from the practice of assembling names for display. Rather than attempt to assemble names from constituent properties, contributors are encouraged to use the property:short_name or property:full_name and for the vernacular name to use the article name for that person minus the parenthetical items for language, birth date and place of birth. This is not a blanket admonition, since templates that assume a particular ordering are perfectly acceptable when restricted to sets of names for which that ordering applies. -~ Phlox 18:36, 26 July 2009 (UTC)

I have no problems with any of the above. Two additions, though.
First, we need to accommodate the Spanish custom of double surnames (mother's and father's). I do not know which of the two counts as a family name, and how this is passed on to grandchildren.
Second, it would be good to distinguish formal and colloquial first names. In German, Dutch and Russian, there can be a huge difference. We could use "first name" for the latter, and rename "middle names" to "given names". This also solves the problem of people being called by their second or third given name. rtol 06:34, 27 July 2009 (UTC)

What are the practical user scenarios driving this "need"?

Actually, Spanish can have up to 6 names as part of the surname. The contributor is free to cram them all in the surname field if they want to do that. We explain to them what will happen. As I recall from my rudimentary knowledge of Spanish, a name like Vicente Fernández Gómez would mean the father's family name would be Fernandez, and his mother's is probably Gomez. In Portugal & Brazil, I believe from the wikipedia article on family names that it is the reverse. If folks want to encode Fernandez Gomez as surname, then his children will not appear in the same category as him, since his cat would be Category:Fernandez Gomez (surname), and his children's would be [[:Category:Fernandez <wife family name> (surname)]]. Relatives that use the surname field to encode their relatives in a florid manner citing selected grandparent family names with desirable names would be even further pigeonholed into oblivion. They need to know that this is not the best place to indicate this information. They should instead put all that stuff in the property:full_name including the appropriate honorifics, titles etc etc in the proper order.

Why do I say oblivion: For instance, a query on fernandez would not yield a hit on his article, since the query would have to be for the real value which is fernandez gomez. The user is going to have to realize that if they do this, then they are making the classification extremely unique and thereby pretty useless for referencing. Ok- we hand them a gun and they blow their foot off. We can't prevent them from doing that. Personally, I would tell them to put Fernandez in the surname, with the full_name=Vicente Fernández Gómez, and the short_name=Vicente Fernández. Say the name was actually portuguese and the contributor really wanted to put both family names in the surname. They could do that and still get the sort right by specifying family_name=Gomez. (The pt convention is that you are supposed to sort on the last name in the list, not the first as is the case with the spanish). Our systems don't need to know any of that junk. We tell them what the effect of specifying the family name is, and let them make up their mind. Technically, it might be better to call internally name this surname-sort-key since that is functionally what it is (not really used as a family name).

There are similar implications for multiple given names (French: Henri Philippe Petain: his given names are Henri and Philippe). The french do not consider Philippe a middle name, and that they don't have middle names. Fine. Let them eat cake and cram them both into the given name property with similar consequences. Encode short name as Philippe Petain since that is what he was known as. Ditto with multiple middle names, such as the newly obscure George H.W. Bush.

Re your proposal on formal vs colloquial first names. Are you talking about William vs. Bill in English, and familiar and formal given names in German? Why aren't these alternate names to search on? What is the user scenario that is not provided for with the system I specified?

Say you take the second taxon level and assign it the name "first name" as you propose. Well, that's odd for eastern asia because first name is the surname, so calling it "first name" is going to confuse well over a billion people. Ok so let's take the second part of the suggestion. We take what we used to call given names and merge it with any data we used to put in middle name property which becomes the newly branded given name property that now occupies the third taxon level. We could do that, but again, this would mean that those doing a query for Georges would get no hit on George H. W. Is that worth it? Again- why is property:alternative names not acceptable for this stuff?~ Phlox 07:59, 27 July 2009 (UTC)

Rename "middle name"? The Russians would enter a patronomic and the relatives of "Lee Kuan Yew, Harry" and George Bernard Shaw wouldn't know what to do with it. How about "other name" or "other names"? Or scrap it and advise users to put everything other than surname and given name in "alternative names"? — Robin Patterson (Talk) 12:48, 27 July 2009 (UTC)

You are mixing up the ideas of what a particular name parameter actually does with how names actually appear. If you as an author of an article on a russian person elects to leave what is labeled "middle name" on an english language form for the person blank, then Ivan Nicolavich Romanov may or may not be listed before Ivan Alexdrevich Romanov. Your choice. Looks messy to me.

Similarly for George Bernard Shaw. In what order do you want the following names to appear?

"George Bernard" could very well be assigned to given name field, just as french might do with their double given names. Fine with me. But if you expect a query with George as given name and Shaw as last name, you are not entitled to be surprised if the article on George Bernard Shaw does not show up. If a contributor wants it to show up, they better only put George in the given name field. Would the report only say George Shaw? Absolutely not. The creator of the report would use property full name, short name or article name in the report. This just affects how the articles are selected and sorted. If they don't put Bernard in the middle name field that is their choice, but the sort of the above list might not be what they want. As I said, this is the contributor's choice.

Regarding the idea of eliminating one of these sort names- of reducing the number of criteria to sort on from three to two: Well, seems like an odd idea to me. We expect to have large numbers of articles and yet we want to reduce the sorting resolution of our classification system, so everything is less clearly ordered in lists. Very odd idea indeed.

That means is that when you have a million records, you can't find anything because all the Harry Lee's are next to each other, ignoring any third order there may be for the person. By the way, the suggestion that chinese don't know how to use middle name fields is ill informed. Note the conventions on how middle names are used for western names versus chinese given names (see wikipedia:chinese name#diaspora). We don't make pronouncements about where they stick western name- Singapore Chinese might prefer the given_name field. Hong Kong Chinese might prefer the "middle name" property on the english language form. Note that the distinction is as I specified in the formal definition above. Understand it. Middle name means the lowest taxonomic sort order significance. If they want chinese name first, then they would put western name in the "Middle name" field. Similarly, the Chinese language version of the form might use the Chinese term for "Western name" to label the middle_name property.

Re:'Why not rename middle name to other name? If you are discussing the internal parameter name, a technical person is directed to read the docs on the function rather than make intuitive assumptions about it (including cultural baggage). The doc would describe it as a third order person name with a formal definition given above regarding taxonomic sort order. So actually other name is even more inaccurate. We could rename given name, surname, and middle name properties with more technically accurate abstract names, for instance person_name.first order key, person name.secondary order key, person_name.third order key. I would do this with engineers maintaining these templates, but not with weekend programmers. It would make already unintelligible templates even more difficult to understand. If you are discussing the name on the form, it is an internationalized string that is edited on Genealogy:Multilingual messages. On an English language form, the name of the field where a person would write Nicolaivich would most often read "Middle name". On a Russian form, the field name where this value would be written would be "Отчество". It is an utterly irrelevant point that Russian does not literally have "Middle names". I can't say this often enough. We are guided by pragmatics not theoretical idealistic classification schemes. Russians do use a tertiary key in sorting, and they do ask for it on forms. We simply use whatever they use. Literally, this term that always appears on Russian forms is the term for patronym. (Not "other", not "middle" and not Среднее имя ("secondary name") which is the translation of the Russian term that corresponds to what westerners call middle names). Should it be "other name" for English? That's up to the community. Certainly, any administrator could go to the Multilingual messages page and change the string to "Other name". I'd suggest that is exceptionally ambiguous, inviting confusion with the similar "property:alternative_names" that has no taxonomic significance. Chinese doesn't have middle name either, but they certainly do have tertiary sort keys. Different systems use different name types for the tertiary sort name. Maybe the Chinese contributors will decide that the best choice is to use the term "Western name" for fm-middle_name. That's up to them. ~ Phlox 18:23, 27 July 2009 (UTC)

The current classification of names on the info pages works. Ditto for the forms. It ain't broke, so we do not need to fix it.
However, we duplicate data input. Short name = given name + surname. Full name = given name + middle names + patronym + surname. If we want to cut out the duplication, we need to think about how names are stringed together from their components. If we are happy to keep duplication, then adding patronyms suffices.
The issue with Spanish names arise if and when we would use the parents' data to auto-fill the forms for their children. rtol 20:47, 27 July 2009 (UTC)

Of course it works. It classifies less than 10K individuals. It has two sort keys available- surname and given name. In China, 100 surnames cover 85% of the chinese population, so surname is not much of a discriminator. This means for one third of the population of the world we want to get by with 1 sort key? Sure when we have a grand total of 82 "Smith" articles, it isn't to surprising that it "works". But do a google search on John Smith. 5.9 million hits. Still think we don't need a third key for sorting folks with identical given and surnames?

The duplication claim is nonsense because it is not possible to reliably form names from the constituent parts as you describe. None of your formulas above work across European cultures let alone global practices for names.

I don't understand what the big issue is with Spanish names. Autosuggested names based on parents won't work for Iceland, Burma and countries with double barreled names such as Spain. So what. The user ignores the suggestion. ~ Phlox 04:16, 28 July 2009 (UTC)