What is needed in a Name?

Question

I wanted to find a way of raising this question within the limited strictures of SE so please bear with me on the explanation.

When we think of personal names, we immediately think of given-names and surnames, often using the more informal and less-correct terms forename and last-name. In almost all places where names are stored - in genealogy and in other areas such as government databases - there is an attempt to formalise names and categorise their elements. Hence, the parochial Western knowledge of worldwide names often gets enshrined in data storage and it can adversely affect people from other cultures.

So, for Western type names, we obviously need to account for possible multiple middle names. Then we find there may be academic titles (e.g. Dr. or Prof.), honorific prefixes (e.g. the honourable, or his holiness), honorific titles (e.g. Sir, Lord, Dame, Lady), or post-nominal letters (e.g. VC, OBE, PhD), generational titles (e.g. .Jr, Sr, I, II, III, etc).

As we move out of the English-speaking world, we find cultures with multiple surnames. The surname category itself has to include patronymic or matronymic names which are a different type of inherited name element. In Far Eastern cultures, there is a generational-name concept that we don't have in the West. There is also a general class of name element called a 'name particle', analogous to a grammatical particle. This includes all those small joining words such as: “von”, “van”, “der”, “de [la]”, “d′”, “the”, “[son] of”, “mc”, “mac", "Ó", "Ní", "Nic", "Mhic", "Bean", "Ui", "y", etc.

Even if the storage supports extended-Latin or non-Latin alphabets, we find different rules for capitalisation (sometimes it is not the first letter of an element, and sometimes it is more than one letter), and sorting (e.g. sorting on the first, last, or other name element).

The Native American cultures effectively have unstructured names that defy all attempts at formalisation, so this is a problem even within the US. A name like Running Deer has no surname concept. If our representation of names is to include pseudonyms, stage names, and other alternatives - any of which could be a simple mononym - then we have to abandon any formalisation.

In my opinion, we're chasing a non-existent formalisation, and all such approaches are doomed. So, my question boils down to: What are the essential properties of a name that we need to record and to distinguish, over and above a simple sequence of name elements? If a product (or data format) supports alternative names for a person, each of which is a simple sequence of elements, then do we really need to know about surnames any other element types?

You might be prepared to answer about inheritance of names and helping the end-user. Well, even ignoring the unstructured Native American example, above, we find multiple rules for inheritance. Some cultures have matrilineal inheritance as opposed to patrilineal. In the Irish culture (and others) the surname is not simply tacked on and there are different rules for men and women of different generations.

It seems for every rule, there will always be valid cases that don't fit.

Answer 1

If you are looking for a format to formalize the structure of a name I would think a key/value collection would get you pretty far. (Us computer geeks commonly call these things a "dictionary", so that's the term I'll use going forward.)

Although the solution I am steering towards is somewhat human-readable, the idea is that it is easily parsed by a computer program where it can then be reasonably constructed into a very human understandable "name". The intent of the solution is to break down the name into atoms, or parts of the name that cannot be further broken down. From there it is a matter of tagging each atom with a well-documented and unique key. The collection of atoms is then bundled into a dictionary that represents the complete state of the name.

So, e.g. an American name like "Dr. John Santa Claus Doe III" might break down to:

{
    title_professional: "Dr.",
    first: "John",
    middle: [ "Santa", "Claus" ],
    last: "Doe",
    title_generational: "III"
}

Note in our example there are two middle names: I chose to represent them as an array under a single middle key, though you could have middle, middle_1, etc.

Also note that the order of the k/v pairs must not matter. It is the responsibility of the program to know the correct order of information given the context of use (First Last or Last, First?) and do the right thing.

You could also nest dictionaries, allowing you to glom related atoms:

{
    first: "John",
    middle: [ "Santa", "Claus" ],
    last: "Doe",
    title:
    {
        professional: "Dr.",
        generational: "III"
    },
}

For this kind of representation the keys are the key: they provide the context for the atomic segment of a name and should be well documented in terms of what the value for that key means as well as what types of values should be expected.

You could even provide nonprinting contextual information in your name dictionary as well:

{
    origin:
    {
        culture: "native_american",
        tribe: "apache"
    },
    complete: "Running Deer"
}

For the above example a program would simply output "Running Deer" as the name, and would be able to provide the cultural background of the name in some meta field.

Another example. Say our name was "John Santa Claus Doe III, 7th Earl of Elgin". The atomic decomposition might be something like:

{
    first: "John",
    middle: [ "Santa", "Claus" ],
    last: "Doe",
    title:
    {
        generational: "III",
        noble:
        {
            rank: "Earl",
            ordinal: 7,
            region: "Elgin"
        }
    },
}

Answer 2

The GEDCOM standard (specifically GEDCOM 5.5.1) keeps names down to the basics, and I really don't think it's that bad.

NAME_PERSONAL:= {Size=1:120}
   [
   <NAME_TEXT> |
   /<NAME_TEXT>/ |
   <NAME_TEXT> /<NAME_TEXT>/ |
   /<NAME_TEXT>/ <NAME_TEXT> |
   <NAME_TEXT> /<NAME_TEXT>/ <NAME_TEXT>
   ]
   The surname of an individual, if known, is enclosed between two slash
   (/) characters. The order of the name parts should be the order that
   the person would, by custom of their culture, have used when giving it
   to a recorder. Early versions of Personal Ancestral File® and other
   products did not use the trailing slash when the surname was the last
   element of the name. If part of name is illegible, that part is
   indicated by an ellipsis (...). Capitalize the name of a person or
   place in the conventional manner — capitalize the first letter of each
   part and lowercase the other letters, unless conventional usage is
   otherwise. For example: McMurray.

They include the very sensible rule:

The name value is formed in the manner the name is normally spoken, with the given
name and family name (surname) separated by slashes (/).

Multiple names can be assigned to an individual, and you can assign a NAME_TYPE to any name as follows:

NAME_TYPE:= {Size=5:30}
   [ aka | birth | immigrant | maiden | married | <user defined>]
   Indicates the name type, for example the name issued or assumed as an immigrant.
   aka = also known as, alias, etc.
   birth = name given on birth certificate.
   immigrant = name assumed at the time of immigration.
   maiden = maiden name, name before first marriage.
   married = name was persons previous married name.
   user_defined= other text name that defines the name type.

(Note the error in the definition above. "aka" has size 3 and is therefore less than the minimum size of 5 that is allowed.)

GEDCOM allows PERSONAL_NAME_PIECES that include the prefix, given name, nickname, surname prefix, surname and suffix. However, GEDCOM 5.5.1 does not recommend using them:

Based on the dynamic nature or unknown compositions of naming
conventions, it is difficult to provide more detailed name piece
structure to handle every case. The NPFX, GIVN, NICK, SPFX, SURN, and
NSFX tags are provided optionally for systems that cannot operate
effectively with less structured information. For current future
compatibility, all systems must construct their names based on the
<NAME_PERSONAL> structure. Those using the optional name pieces should
assume that few systems will process them, and most will not provide
the name pieces.

This basically says that you should not try to get too complicated in defining names in a standard. If you do, you will make it harder for programs to be compatible with each other.

Answer 3

My sense is that what you're after is a name object that consists of 1) a unique identifier 2) a human-readable form governed by some number of culture-specific rules 3) a number of rules that specify how a child's name might be derived from those of his parents

Current implementations may conflate 1) and 2), and hard-code 3).

Answer 4

It may be useful for me to sketch in some elements of names and / or titles for British aristocracy in order to broaden the view. I have no direct need for these things so I haven't really sorted them out in my head. The fuller version of what follows can be found in the archives for the Rootsweb Mailing List FAMILY-HISTORIAN-USERS three years ago. Specifically in the thread "Hi, I'm new here & 1 question about entering titles" where I blundered in with a little knowledge and Patrick Cracroft-Brennan was good enough to correct me.

Trying not to commit the same errors again, and summarising, for British titles, there can be two elements: "style" and "title". To take the example of Prince Philip, Duke of Edinburgh, husband of Queen Elizabeth II, his titles (in the plural) are Duke of Edinburgh, Earl of Merioneth and Baron Greenwich. He bears the style and precedence of a Prince of the United Kingdom of Great Britain and Northern Ireland. Part of that is the right to the style of "His/Her Royal Highness" ("HRH"). So one particular "name" for him is "HRH The Duke of Edinburgh". So in this case the "name" is a sort of concatenation of the style and the title.

Switching to my Clan Chief, the current Earl of Elgin bears the titles Baron Elgin, Earl of Elgin and Earl of Kincardine. He is the 11th Earl of Elgin and 15th Earl of Kincardine. According to Wikipedia, a Baron would usually have the style "The Right Honourable The Lord", so if Baron Elgin were his only title and the usual rules applied, then he would be (in full) The Right Honourable the Lord Elgin, a.k.a. Lord Elgin for short. Except that I don't know whether the Scots or UK rules over style should apply.

Clearly the concepts of style and title are linked to name, don't always form sub-elements of any sort of name, but do influence the "name".

Also, adding other elements into the mix, we have the concept of courtesy titles. Here, one of the titles belonging to the father (usually) is used by the son (usually) but this is purely as a courtesy - the peerage actually remains with the father. So the Earl of Elgin's heir apparent is the present holder's son Charles Edward Bruce, Lord Bruce.

Then there's Prince Charles, heir to the throne. In England and Wales, he is referred to as HRH The Prince of Wales. In Scotland, he is known as HRH, the Duke of Rothesay. Two different contexts.

Now, I've probably made all sorts of errors in the above. My objective is not to document styles and titles of the UK, but rather to illustrate some of the concepts, the complex way in which they react together and form "names", the way in which contexts can apply (e.g. the courtesy titles of the heir) and generally to pose the question whether imposing a structure makes any sense. Because, of course, this has been a totally Anglo-centric view.