During my research I have found that in my village of focus, marriages were often registered in more than one parish, in multiple books. This means different locations, sources, dates (often more than 1 day apart) and the witnesses are not necessarily the same either, indicative that these were indeed separate events, even if only administrative for the family. I can confirm that these are not errors by the registrar but that perhaps someone has reported different dates/witnesses to the different parishes. These differences force me not to log these events under a single Marriage with two dates (or notes explaining).
I am aware of this question about multiple dates for one birth event, but my case is different. I cannot register these marriages as separate Marriage events either, as GEDCOM only allows a single Marriage per family, so I risk losing data whenever I export.
Reading the Gramps wiki on Alternative Marriage (“A relationship such as Cohabitation, Common Law Marriages, Civil Unions, Domestic Partnerships.”) clearly tells me that it does not apply: both parishes register the very same religious, bounding marriage, so neither is an alternate.
Any idea how to formalize this very human thing for the computer?
Oh, I didn’t know that, thanks for the correction!
I use a GEDCOM parser wrote for a third party application by someone else, and there the standard procedure is to emit a warning and remove redundant marriage events on import. So I assumed that is because of the GEDCOM format. Perhaps it was for an earlier GEDCOM version?
Anyway, then I guess the reasonable way would be to store separate Marriage events.
I do it that way. For example when they married in the office and the next day in a church.
One interesting question is how to handle it if a couple married, was divorced and married again. I prefer to use two families. The first marriage and divorce belongs to the first family. The second marriage belongs to the second family. That allows to put the children from the first and the second relationship to the correct family. It has the disadvantage that some genealogical programs label them as half-siblings.
Since this is “logically” the same single family, why don’t you assign first-marriage, divorce, second-marriage events to this family? This solves the half-sibling problem.
But, it may induce other problems for the “interim period” between divorce and second marriage. Your opinion?
Isn’t that covered by the time frame described by the marriage and divorce events?
More generally, this emphasises the need to define what a Family is in Gramps (or perhaps the significance a researcher gives to this record).
It can be seen as a relationship between two persons and events record the (legal status, time, …) history of this relationship. Then the type is no longer necessarily “constant” and the type field may not be meaningful over the whole period.
It can be seen as covering only a preiod of time where “properties” remain the same such as legal status. Then even with the same partner, there are several “families”.
There are pros and cons in both concepts. I think the choice depends on the proportion of “problematic” cases and the complication of each situation.