Family Historian User Group

On entering a census using AS for 1891, the entry screen shows the wife's maiden name in the entry, but shows the married name in the text.

By 1901, with Jane now a Widow, her maiden name is shown in both sections.


I've never noticed this before, but is it intentional that the text now shows her with her maiden name, and if so why? There are no other marriages/divorces etc for either partner, and the same thing occurs with other couples I have tested.

This happens if you have no marriage date or details entered for the Widow. AS doesn't know that she has been married.

Ok I see that (partly).

But AS knew she was married in 1891, so with no other info regarding a 2nd marriage should it not default to the married name? After all putting in the marriage date, does not mean she is still known by that surname in the 2nd census. The logic for using the marriage date to call her by her married name is just as liable to error as using the fact that there is (or was) a spouse!

Both could be overwritten by circumstances, but the fact that she had a spouse is not really different from the fact that her marriage date is known. Both indicate nothing more than that she was known at one stage as (in my case) Jane Piper.

It is nothing as clever as using Marriage events or dates. You are reading too much into the process.
Working out who somebody is married to or widowed from at a particular date is quite tricky.
I know because my Lookup Missing Census Events plugin performs that algorithm.

In the main Census Grid, as elsewhere in FH & AS, the Primary Name of each Individual is always used.

If both partners in a Family record are included in the main Census Grid, then in the Auto-Text the female is given her male partner's surname, regardless of whether married, divorced, widowed, etc. (Not sure how same-sex partnerships are handled.)
They don't have to be entered as Head and Wife or given any relationship at all.
As the Help states, the Auto-Text is an educated guess at the transcript and must be reviewed before saving.

BTW: The default Relation droplist doe not include Widow, but it can be added to the template or simply typed in.

But I tested adding a marriage date, and the Married surname was then displayed

Sorry, you are absolutely correct. I was investigating the case when there is no Marriage date.

However, AS also uses the married surname if the census date is after her husband's date of death and there was a marriage date before that.

So is your original example explained by there being no Marriage date for George & Jane?

Then, as I said, her partner' surname is used when both are included in the Census Grid even without a marriage.
But after his death, Jane reverts to her maiden surname because there is no Mariage date.

tatewise wrote: ↑

03 Oct 2020 09:52

BTW: The default Relation droplist does not include Widow, but it can be added to the template or simply typed in.

Widow isn't a relationship, it's a condition and Widow is one of the options in the condition column.

Sorry for the delay in responding, yes, my original example had no marriage date for George & Jane.

I fully understand that the existence of what FH calls a spouse does not mean there is a marriage - although in the 1800s this is likely to have been truer in most cases, whereas today it is far less likely to be the normal assumption.

However, it does seem that FH and AS both do assume that the addition of a spouse does generate a marriage. In particular, the Default treatment for Marriage status, currently does generate a marriage.

Here I added two test records, making one a spouse, and entering absolutely no other data - in particular no Marriage date - simply a spouse. The Individual Narrative report for each of them includes a sentence saying they are married. I am aware that I could enter "unknown" as the marriage status, but in the case I am dealing with, it would be wrong, as I know from other documents that there was a marriage.


AS also regards them as married if the male is still alive, so it is not logical to treat the widow by reverting to her single name simply because the male is dead.

If it is right to call her by her maiden name after the death of her "spouse", then exactly the same logic should apply while the "spouse" is alive.

However it does seem a far more sensible treatment that if AS & FH both use the married name for a female when the male is alive, then the logical step is to continue to use the married name after the males' death. There was no evidence to confirm the marriage in the first place, but equally, there is no evidence to show that there is now no marriage.

I agree that there could be improvements, but some of your arguments are inaccurate.

FH never ever uses the married surname for women in any circumstances in any displays, Diagrams or Reports.
The only assumption FH makes is that Family partners are 'married' in Reports, unless the Marriage Status says otherwise.

It is only AS that attempts to adjust the married women's surname in its Auto-Text display to try and make the transcript as accurate as possible, but in general, that is never going to be 100% achievable.
Nick may be able to modify the Auto-Text for the Census Grid so that female partners retain their male partner's surname, even after their Death, even when there is no Marriage event.
The novel feature is that surnames are being adjusted for the specific date of a Census.

The complexity is that the software must build a timeline of multiple Marriage & Death events to determine which surname applies at any chosen date. That is further complicated by Divorce & Annulment events that may affect the women's surname.

ColinMc wrote: ↑

04 Oct 2020 09:09

AS also regards them as married if the male is still alive, so it is not logical to treat the widow by reverting to her single name simply because the male is dead.

It's been many years since I've looked at this aspect of AS but I've just done some testing and also looked at the program code and my findings are (assuming one marriage only):

• If there is no marriage date then AS doesn't change her name. If, however, you add her spouse to the census entry then it does.

• If there is a marriage date then AS will change her name if it is after the date they married. The death of her spouse isn't relevant to this.

So in your first post in 1891 Jane is shown as Piper because George is included in the same entry. But in 1901, because there is no marriage date and no husband present, AS doesn't change her name because it doesn't know when they married. I agree that AS could do better by taking into account the spouse death date. The system that tries to predict marriage status does take into account death of the spouse and hence correctly records her as Widow.

In the next version of AS, the autotext prediction of a woman's surname has been improved so the death of a spouse is taken into account if a marriage has no date. I also fixed a bug with the marriage status prediction if a widow subequently remarried. Thanks for reporting this Colin.

Best wishes

Nick

Just to clarify, I never suggested that Married surname was used in any circumstances. Further, the topic is in the AS forum, and I did say that it was AS that I was querying, not FH, so not sure my arguments are inaccurate.

Anyway, thanks Nick for agreeing to look at it.

Sorry, but just to clarify my comment, on Sun 4th Oct 2020 10:09...

ColinMc wrote: "... that if AS & FH both use the married name for a female when the male is alive ...

That suggestion that FH uses "the married name for a female when the male is alive" is the inaccuracy I was correcting.
FH never uses the married name for a female.