Suppose Mary A marries John B and then (after his death or divorce) she marries James C.

Rightly or wrongly my impression is that conventional practice in GEDCOM based programs is to record only Mary A as her name because her married names are "obvious". It is clear that this practice is creaking where Mary uses married names that are not "traditional" (where tradition is that of the English speaking world). Problems here include

  • Mary A becoming Mary A B and then Mary B C (where the family-name is still one word and the previous family-name migrates to become a given name);
  • Mary becomes Mary A-B (a new compound family-name);

However, this conventional practice can give rise to issues even when Mary's names are traditional, as any diagram (or report) centred on her 2nd marriage only will show, unless adjusted, that Mary A marries James C, when the marriage certificate probably records Mary B marrying James C.

To help overcome this, whenever a woman marries twice, I give her a secondary / alias name of her first married name (Mary B here) and adjust my diagrams to show multiple personal names. Should I be recording all married names to remove any risk of incorrect assumptions?

So - what is the best practice (and / or society standards) for recording married names? And what should it be?

Note husbands changing their name and issues with children's names are out of scope of my question. Note also - I work within GEDCOM practice where (stupidly) no dates can be assigned to Names.

  • Wish I could add another plus 1 the point about assigning dates to names.
    – GeneJ
    Commented Nov 4, 2012 at 1:26

4 Answers 4


Should I be recording all married names to remove any risk of incorrect assumptions?

It seems to me that you should be recording all names -- married or otherwise. Use notes to add detail where necessary.

If a person is born Mary Alice Aston and marries John Bower and changes her name to Mary Aston Bower (i.e. dropping her given middle name and using her original surname as a middle name), you should record this new name and add a note that she took this name upon her marriage. If she is referred to in records (i.e. incorrectly -- it does happen) as Mary Alice Bower, you should record this name and add a note. If she divorces and changes her name back, you should add a note recording this fact. If she then marries James Carter and changes her name to Mary Alice Carter, you should record this name and add a note.

There are just too many possible variations to "assume" that the married name will follow the "conventional" (English) rules.


Genealogy is all about building personal identities using many pieces of evidence including names. The first identifier an individual get is usually a name, so it makes sense to use the same identifier throughout life. Unfortunately names change in many ways, not just on marriage, some of which are official and leave a record. Other types of name changes include official deed poll (e.g. men changing thier name to inherit a title or under the terms of a will of a wealthy relative), double barred surnames, nicknames ... So the name alone makes a very poor primary database key. It is important to track the name changes, as vital parts of the person's identity. I would want the option of recording the period a name was used as it is relevant to building a personal identity.

Naming conventions vary considerably as do the way they are recorded in official records. For example English baptisms mostly do not give the mother's maiden, but Scottish baptisms do. Spanish people keep the name they were born with for life so a woman does not change her name on marrying and uses two surnames the first being the father's surname and the second being the mother's maiden name. The practice of changing surname on marriage is now rejected by many women for professional reasons and personal beliefs.

I agree with other comments concerning the inadequacies of genealogical software.

  • 2
    In the 19th century and earlier, married Scots women were known just as much by their maiden name as married - arguably, more often for official purposes. Which caught me out on one occasion when I claimed the wrong 4G granny, as mine had "written down" her maiden name on the census form despite being in a full household of men with her married name. And naturally, there was another with her married name from the same town, of the same age, living on her own, who I thought was G-gran.
    – AdrianB38
    Commented Nov 8, 2012 at 10:13

The way I learned genealogy, prior to genealogy software, was that women always be referred to only by their maiden names.

But as you note, in many Western cultures, for many years it has been the practice that women change their surname at marriage to that of their husband's. As a result, almost all the women in your family tree would have been know most of their life, not by their maiden name, but by their married name.

As a result, if you list women by just their maiden names, most will be unrecognizable to the people viewing the information.

GEDCOM does have a NAME_TYPE that part of the PERSONAL_NAME_STRUCTURE that does allow you record the birth name separately from the maiden name separately from married names and even also known as names. If your software program supports this, then this is the proper way to enter married names that the person has adopted.

Personally, I would hope software developers implement this by giving you the option to let the program automate the adding of married names to women. It would then present your example woman as:

  • Mary (A B) C

and would index her in the name indexes under all three names, as: "A, Mary", "B, Mary" and "C, Mary"

In cases (which are more common now-a-days) where non-traditional surnames are adopted, the program should allow you to select which married names not to include, or which ones to override and what with (e.g. with Mary A-B) so that the person can be listed properly and indexed properly.

I don't agree with your suggestion that dates should be able to be assigned to Names in GEDCOM. Dates are assigned to events, not to facts. A name is a fact. The changing of a name is an event. So the date of the name change should be included as a custom name-change event in GEDCOM. This is often used when someone legally changes their name to something else, and many programs allow you to enter a name-change event.

For names changed at marriage, this event is usually assumed to coincide with the marriage date and an additional name-change event would not be required.

For best practice, I would record the birth surname as every person's surname. I would record non-marriage surname changes as name-change events. I would note and indicate non-traditional surname changes at marriage as notes with the name. I would try to find a program that would index people by all their surnames (birth, maiden, every marriage, other changes) and show all their surnames in reports and save all this surname information into GEDCOM so that it is potentially retrievable again.

  • 1
    Interesting you say "prior to genealogy software ... women always be referred to only by their maiden names". As such reports are composed by humans, it would allow the writer to add an extra (sur)name in where necessary - e.g. "Mary A (the widow of John B) married...". But software isn't usually intelligent enough to see the need and add such clarifications. Hence a workable standard for pre-software era becomes a disadvantage for the software era. I wonder how many other "standards" there are like that.
    – AdrianB38
    Commented Nov 6, 2012 at 16:44
  • 1
    Re "GEDCOM does have a NAME_TYPE that part of the PERSONAL_NAME_STRUCTURE". Another thing I hadn't realised - then, when I check it out, it turns out to be another thing that's in GEDCOM 5.5.1 and so not in my strictly GEDCOM 5.5 compliant program!
    – AdrianB38
    Commented Nov 6, 2012 at 16:47
  • 1
    @AdrianB38 - Programs today should all move to GEDCOM 5.5.1. Even though it was never declared the standard, it is the de facto standard and is the only version of GEDCOM that allows a UTF-8 translation format for the Unicode character set which is an absolute must-have, among other enhancements.
    – lkessler
    Commented Nov 6, 2012 at 17:51

My approach is similar to yours Adrian. STEMMA records explicit married names - rather than assuming them to be obvious - but indicates the dates (or events) at which they changed. The recording of time-dependent names, and other types of alias or diminutive name, has a direct analogy in place names.

However, this is just part of the total data associated with that person. It impacts no specific diagrams since the label/title used for display purposes can be picked independently of the personal names.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.