How to represent Stillborn in GEDCOM 7

[dthaler]

As far as I can tell, FamilySearch GEDCOM removed the ability to record in any interoperable way the fact that an individual was stillborn. One cannot rely on TYPE Stillborn, since that is freeform text and could be entered in any language. One cannot rely on DEAT AGE 0y since the spec explicitly says 0y can be used for anyone who died less than an year old. So if one needs to, say, determine whether a SLGC should be performed, one can tell with GEDCOM 5.5.1 but cannot tell with GEDCOM 7, which is a disincentive to switch to GEDCOM 7.

See https://github.com/FamilySearch/GEDCOM.io/issues/26 for more discussion.

[tychonievich]

What definition of "stillborn" are you assuming? The modern US English version "dead at birth"? The 19th century UK English version "born alive but not breathing"? The Roman Catholic traditional use "died before the parents could baptize it"? The Jewish traditional use "died before circumcision"?

There may be others too; the OED has this in its etymology quotes in the entry "still-born, adj. (also stillborn)":

1955 W. P. D. Logan in Holland & Bourne Brit. Obstetr. & Gynæcol. Practice: Obstetrics xxxix. 1140
In certain countries children born alive but dying within a stipulated number of days are registered as stillborn.

5.5.1 certainly doesn't make it clearer in it's definition "died just prior, at, or near birth, 0 years" nor it's example of DEAT.AGE STILLBORN "meaning this person died at age approximately 0 days old". The definition is broad, inclusive of all of the various "died very young" meanings of stillborn without giving any indication which was meant, while the example is narrow and in contradiction of the definition.

I believe that removing STILLBORN from 7.0 improved clarity by removing an under-defined tag with culturally-specific interpretation that caused more misunderstanding than it did accuracy. DEAT.AGE 0d is more precise than DEAT.AGE STILLBORN and appropriate in most cases; but when it was a before-a-rite version I'd prefer to see DEAT.AGE <8d than DEAT.AGE STILLBORN and have to guess that STILLBORN meant <8d not 0d because of the parent's INDI.RELI Jew.

I oppose bringing back AGE STILLBORN.

So if one needs to, say, determine whether a SLGC should be performed, one can tell with GEDCOM 5.5.1 but cannot tell with GEDCOM 7

We intentionally kept SLGC.STAT STILLBORN for this reason. Note that STAT STILLBORN and AGE STILLBORN were never equivalent; the former is defined by a specific church's current policy while the latter by an author's culturally-specific interpretation of a word.

[tychonievich]

If we do wish to make a normative path to representing some type of stillbirth, we need to

1. clearly define which type we mean; I think the most common that are not simply numbers of days are:
   • late-term miscarriage
   • died during labor or delivery
   • died after birth but before breathing
   • born without breathing, declared dead, but then subsequently revived
2. add a normative structure for the one we pick

If we decide that there's one most-important definition to include, I'd favor adding it as a new kind of individual event in 7.1. If we decide that there are several all worth defining, I'd favor an enumerated value substructure of BRIT or DEAT (whichever more naturally encompasses all of the definitions we pick) instead.

It is not clear to me that we have enough demand to warrant this kind of addition.

[dthaler]

What definition of "stillborn" are you assuming? The modern US English version "dead at birth"?

Yes

I'd favor an enumerated value substructure of BRIT or DEAT (whichever more naturally encompasses all of the definitions we pick)

Yes. Note that GEDCOM 5.5.1 allowed "AGE STILLBORN" under any event, not just DEAT. I don't think that's necessary though, either BIRT or DEAT is sufficient in my view.

BTW, I think this same question was asked in

• https://genealogy.stackexchange.com/questions/3795/how-to-record-a-stillbirth-as-a-genealogical-event

[tychonievich]

Discussed 2022-01-12

Agreed there is a problem, namely: (a) neither 5.5.1 nor 7.0 can reliably record "born dead" in source data, only sealing status data, but (b) the Church does check for stillbirth in some way to decide ordinance permissions.

• Proposed recording BIRT.DATE XYZ and DEAT.DATE BEF XYZ (this would not work with ABT/CAL/etc in BIRT.DATE)
• Could add a hint like "Programmatic checks for stillbirth should check for BIRT.TYPE Stillborn".
• Could add a new in-grammar rule, like defining AGE < 0d to mean "strictly before being born" (currently AGE < 0d would mean under 24 hours
• Could add a new structure in 7.1, such as
  • a new Event type "deadborn"
  • a new enumerated-payload substructure of BIRT and/or DEAT

[tychonievich]

Proposed:

1. Add a STILLBIRTH individual attribute (to match gedcom-x's stillbirth fact)
   • could alternatively be an event
2. Add an optional enumerated substructure to define what type of stillbirth it was:
   • Miscarriage
   • Dead before birth
   • Born alive but never breathed
   • Died shortly after birth
   • Died before a near-birth rite like Christening could be performed