Relationship names

Started by Private User on Sunday, November 7, 2010


Profiles Mentioned:

Related Projects:

Showing all 23 posts
Private User
11/7/2010 at 6:24 PM

Hello all

It's annoying having no way yet to translate relationship names such as "father", "mother", "second great grandmother" etc. in phrases such as "X added his second great grandmother, Y" or "Y is your mother". Even in the Immediate Family section on anyone's profile, it lists each of their relations with an English word. This must be especially annoying for people with non-Latin scripts (let alone RTL scripts).

I'm sure the Geni team are working on this one; it seems like there's no point in having Tr8n for a genealogy site if you can't label your relatives properly. I know it must be a difficult task, especially when it comes to relationship paths such as "X is your second great aunt's grandmother's brother's wife". Having to include relation name, word order, genitive (if applicable), etc. would require a great number of context rules, I imagine...

Especially, I suppose, because different languages and cultures have extremely complex ways of describing relations (far more varied than English) which are based on much more than simply gender. Wrap your mind around this, for example:

Just putting in a vote for this feature to be made available soon!


Private User
11/8/2010 at 10:25 AM

This is definitely something that has plagued the translation project since it's inception. As you point out, the problem is that each language has its own core definition of relationships. No one can deny though how essential it is to the functionality of Geni in other languages. Lets just hope they get to this sooner than later.

Private User
11/8/2010 at 5:48 PM

Being that the relationship names component is so fundamental to Geni but also so complex, I would advocate for an interface other than the simple (and awkward) list of phrases in use for the rest of tr8n.

A link from the Chinese page I referred to above ( shows all of the Chinese relationship names in a tree (just like any family tree, but with names of relationships rather than people).

I would like to see a visual interface for relationship names (as a separate tab in Tr8n) laid out like the above link. Because there's no one-to-one correspondence between English kinship terms and terms in other languages, and things like "paternal elder male cousin's wife" are hard to unpack, having a list would be very inefficient and liable to error.

In summary, a new tab in Tr8n dedicated to relationship names, laid out in a family tree format (coloured for gender, numbered for age) would be a good way to go about translating relationship names, in my opinion.

You should be able to specify gender, age difference (birth order?), generation, etc. and as many or as few options as are applicable to your language.


Private User
11/8/2010 at 5:54 PM

The most important part is that the relationship is based on the path taken to get there (as in the Chinese example) rather than the English preconception of the relationship name, so a limited list of terms seems tricky.

Note also that English, by way of example, has some words that go straight from person X to person Y, i.e. "uncle" (shortcut for "parent's brother"), "grandfather" (for "parent's father"), "niece" (for "sibling's daughter"), etc., whereas there are also relationships for which there is no one word in English: terms which have to refer to an intermediary, such as "cousin's husband".

I suppose all this would have to be taken into account. I almost envy the Geni team the fun of nutting this one out programmatically. :)

Private User
11/8/2010 at 11:09 PM

I spoke with Geni about this today and they told me we won't see this "anytime soon". He went on to say "I would be very surprised if we had [relationship name translations] in less than six months".

Conceding, I suggested maybe an interim solution of having Google automatically translate relationship pathways. Granted, Google is often wrong but given the unsightly wait-time, I don't see any other options.

What are other's thoughts on this?

11/9/2010 at 8:10 AM

I like Kit's idea of the visual interface for translators. Given such a tool, Geni could have a phased approach giving access to the simple stuff first (I make the assumption they are 1:1 translatable across all languages/cultures) such as mother, father, brother, sister, daughter, son, wife, husband, sibling.

A 2nd phase could be to release structures that could differ by language. Again, making an assumption that might not hold true, some languages have a simpler, more "mechanical" approach than the very complex Chinese example. These languages are not too far from the English "cousin removed" concept.

Some related languages use the same "core mechanics" and could then benefit from being released without needing to wait until every complexity is sorted out for every language.

Intuitively I think it is almost impossible to make one set of core code that can handle all these complexities. Having worked with software development I also know the pitfalls of splitting up code giving an exponential increase in complexity of maintenance. A tough one to sort out but they will need to get stuck in sooner or later and as the rest of the user interface nears completion (language by language), this stuff sticks out like a sore thumb !

Private User
11/10/2010 at 4:05 PM

Thanks, Victar, for the news from the horse's mouth.

Oddbjørn's suggestion is probably more unwieldy than helpful, I think. You're right about the pitfalls of splitting up code. I think we've just got to grin and bear it for the six months that this takes to develop.

If they're out there, reading it, thanks to the Geni team for the work they're putting into it; we're not impatient, we're just overenthusiastic! :)

11/21/2010 at 7:38 AM

There are also some languages with more than one relative naming conventions. For example, in Croatian there is a name for the wife of the husbands brother (jetrva) but a small percentage of croatians know these words exactly. It would be easier to say wife of husbands brother. So which to use? Is it possible to use both?

12/18/2010 at 3:23 AM

Dear Kit,
Thank you for your EM which popped up on my Iphone but not on the computer hence my replying to you via geni. My great uncle and latterly myself have done extensive research into the family tree and have got back as far as one Henry Sheargold d1709 in Rock Staffs...therafter the family remained in Worcs and the chances are we are related!. It would appear that the name Sheargold was changed from Sheingold at some stage (Jewish in origin..which probably wont surprise you! Eastern European most probably Polish) Henry originally lived in the Southampton area so this was probably the port of entry. I have 2 David Sheargolds in my tree... 5th generation from Henry.
1. David Sheargold b 15/7/1842 in Bombay India appearing in the Droitwich Census (Witton) in 1881 aged 8 then in Newport St Worcester aged 29 and so on till he died in 1843. He was a carpenter. He had a son named David William Winter Sheargold b 1868 in Worcester he died 1874. If you think either of these David's belong to you I have a little further information plus background as to how they came to England and why. My mother was also a Sheargold hence the interest.We too have Sheargolds in America and Australia in our FT. Hope this is useful to you Regards JMB

12/18/2010 at 9:04 AM

Dear all

Small team of Estonians had now already prepared "Estonian" relationship model which is required for translation of all possible relationship names (and situations).
The spreadsheet of all related English and Estonian words is here:
As you can see we have already created full Family Tree
which starts here:
Focus Estonian
with relationship names where first name of it is in Nominative Case and the second name is in Genitive Case (Whos?)
(Ref: Estonian grammar - and Estonian language -
There is also related project:
Short English overview of it:

12/18/2010 at 11:56 AM

To make the project more comprehencive I just added the English summary there.

Private User
12/18/2010 at 2:22 PM

I am still wish that they implement the two-level relationship model I have suggested several times, which should be easy to implement on top of current model.

The model would get rid of the technical "removed" descriptions, and as far as I understand it will fit most languages.

The model is simple: When you have a path including "removed", which means a generation difference: Follow the path up or down until the removed level is zero, show the relationship to that profile and then the relationship further on.

To make a simple example: "First cousin once removed" would be described as your father/mother/son or daughters first cousin

The model will both show the direction the generation-shift goes and the gender of the common ancestor, and it will also be the way we describe such a relationship orally.

Translation of a direct relationship is however more difficult to implement.

In some languages like Sami you have as far as I understand a different name on each aunt depending on the birth order, - the oldest aunt have one "aunt" name and so on.

In Scandinavia we skip one level in cousins (second cousin is described as a "tremenning" - third cousin and so on), and we have an extra level on grandparents, "tipp-oldeforeldre", which means that a third great grandfather is "just" the second because you start the numbering on the extra level, and sone don't even number the first levels but repeats the description until a certain level: (tipp-tipp oldefar - second "tipp" level, - in the same way as you in English would use great great grandfather instead of second great grandfather)

12/18/2010 at 6:33 PM

@ Bjørn
"To make a simple example: "First cousin once removed" would be described as your father/mother/son or daughters first cousin
The model will both show the direction the generation-shift goes and the gender of the common ancestor, and it will also be the way we describe such a relationship orally.
The model will both show the direction the generation-shift goes and the gender of the common ancestor, and it will also be the way we describe such a relationship orally."

But hiding information is part of Geni's business idea !

"First cousin once removed" is your parent's first cousin or your first cousin's child. First cousin of your child is your nephew or niece :)

3/12/2011 at 9:27 PM

A slight bump due to a newly created and topic related project.

If you're familiar with, or even have some command over your English + Swedish skills—in combination with an itch to translate and knowledge on how to best navigate those curvy relationship paths, kinships etc., jump into this and have a go:

Translators: Swedish Kinship Proband

I'm far from 100% about the final usefulness of the initiative—or other similar endeavors—since we still do not know when, if, or in what shape or form Geni eventually intends to roll-out any solution in conjunction to our beloved relationship paths. But, if it's to be done somewhere down the line, it can't possibly hurt anyone to have a workable, translated kinship structure standing by in the form of a tree.

Just shoot.

Private User
4/25/2011 at 7:35 PM

I'm really happy to see this discussion, and would be happier if there were any update on this (after roughly 6 months).

Even the words brother and sister in Chinese make clear distinction of being elder or younger. Direct translation can never work. What Geni needs to do is to express the relationship name in a mathematical way, as a string that encodes the exact path unambiguously. Then a designated page in tr8n can ask users to translate them into actual languages.

Without getting the names right, Geni can never create a world family tree. Almost a quarter of the world's current population is Chinese, and perhaps even more percentage-wise for all people ever lived.

Private User
4/25/2011 at 7:39 PM

and ideally, the Chinese user should be able to further customize the (informal) names they are most comfortable with. The regional difference is as great as the different languages in Europe. The Kwan family site is mostly Cantonese version, if I'm not mistaken.

Private User
4/25/2011 at 7:56 PM

Private User, thanks for contributing. We sorely need more Chinese representation.

Private User
5/3/2011 at 6:57 PM
5/3/2011 at 8:05 PM

Perhaps it is, though a short description might still be in order here since it's seemingly one of those crème de la crème exclusive threads to which us regular users have no access.

Private User
5/3/2011 at 8:11 PM

Ah, sorry. Geni says they're planing to "implement translations of the relationship paths in the semi-near future."

5/3/2011 at 9:15 PM

Olav - you meant Creme de la Curator...

Private User
11/9/2016 at 11:22 AM

Sorry to resurrect this thread. Have there been any news on this issue?

Private User
11/10/2016 at 3:14 AM

Nope, but we have a hope since rumors tells that the mother company, MyHeritage, which also have taken over most of the development for Geni, have implemented the two-level relationship model I have suggested -with translations.

Showing all 23 posts

Create a free account or login to participate in this discussion