Use of Non-Roman Alphabets

Question

Hi, is there a single place on Wikitree that describes the protocol for using letters (other than from the Roman alphabet)?

I don't want to come across as culturally insensitive, but I am not familiar with the Cyrillic alphabet, the Greek alphabet etc, so I find this aspect challenging especially as the vast majority of profiles are in the Latin alphabet and the language of the forums (for eg) is predominately in English.  So, when looking for duplicates, how do I ensure that I am not missing a duplicate (for eg) because one surname is written in the Latinised form and the other in the original language?    Does the Wikitree search engine automatically resolve for this?  And are there guidance notes (in one place) that describes Wikitree's rules for using non-Latin script?

Thanks a lot

Comments

Hi I have been thinking about Lianne's answer, which is technically correct, but the more I think about this, this is not a great situation.  

I believe the aim of Wikitree is to have one single profile for every person, not to have duplicates.  If we follow these guidelines, we will end up with lots of duplicates because people will be using both Roman and non-Roman alphabets when creating profiles for the same person.  This situation would never be resolved because the search engines cannot distinguish between two different alphabets. It is also unrealistic to hope that non-native speakers (of whichever language) would be able to effectively search for a surname if it is written in a different alphabet.  I could not do it, and I bet many people will be the same.

I think if someone is coming to Wikitree, then chances are they will be familiar and comfortable in working with the Roman Alphabet.  I suggest that when it comes to creating profiles, and LNABs in particular, that the Roman Alphabet spelling is used since this will be the alphabet that vast majority of Wikitree members will be able to use.    I am not suggesting we should use the Anglicised spelling, I am proposing the closest approximation to the spelling of the names in the original language but using Roman letters.  

I would then insert the name using the "non-Roman" alphabet letters in the nicknames field.  This way the profile is still recording the original spelling, but this will also enable the search engines to capture the widest possible pool of profiles.

Please understand, I am not against the use of other languages, I am just trying to think of practicalities.  For me personally as an Arborist, there is no way I can search for duplicates checking against other alphabets.  I think we need a basic rule that the majority of Wikitree users can work with, otherwise the inevitable outcome is a large number of inadvertent duplicate profiles.  

Or we need many more Arborists who are capable of working in more than one alphabet :-)

---

This suggestion has come up before, and as always I want to voice my strong opposition to it. If we switched to using the Latin alphabet for everything, we would be giving up on ever being international in any real sense of the word.

---

I am only suggesting this for the LNAB when it comes to creating profiles.  I agree it is important to be international, and to incorporate the original language and spelling of someone's surname, but since the search engines cannot cope with more than one alphabet, we should settle on one to facilitate searching.  Otherwise we end up with lots of duplicates which is not the aim of Wikitree.

Could we solve the problem - at least until the technology catches up - by asking people to create profiles using the Latin alphabet but then recording the correct, original spelling, in the Alternative Names field?

---

There is also the issue of Spanish names which seem to float down from the mother rather than the father.

There is also the minor problem in Russian of males and females having different name endings

????? Александрович (Romanov) Романов		John Atkinson
?????? Александрович (Romanov) Романов
????? Александровна (Romanov) Романова
??????? Александрович (Romanov) Романов	14, 1850 St Petersburg, Russia

 

---

It is my opinion, and not a very popular one it seems, that the only practical way to run an international forum like Wikitree is to agree on a common language, a lingua franca, and to standardize its use, whether English or French or Latin or Hieroglyphics. The use of multiple languages, while it might be appealing, is just not practical.

---

[This is supposed to be an answer to the comment Leigh Murrin. I clicked "reply" under that comment, but it doesn't show up in my post.]

I am having problems you are trying to address, and I fully agree with your ideas. But I find that the current form to input people is friendly neither for Roman-first nor for original-first input. Moreover, I just discovered that after changing by surname from Romanised "Karapetyan" to proper "Карапетян" I ended up also changing my ID. Since I also did it for my father, but corrected his surname first, I now have Karapetyan-1 = Карапетян-1 pointing to myself and Karapetyan-2=Карапетян-1 pointing to my father. This is as confusing as it can be.

A more robust approach would be to do it along these lines:

Surname at birth (Romanised):
Surname at birth (native spelling):
First name (Romanised):
First name (native spelling):

And so on.

---

Konstantin, I entered my profile before I entered my father's so the number in my WikiTree ID is lower than the number in his. It happens all the time, and it's no big deal. (Although I will grant that it look weirder when you're 1 and 2 in the series.)

Let me say that I'm thrilled to see a Russian speaker on WikiTree. Just yesterday, I was looking through Wikipedia's List of space travellers by name, trying to compile a list of deceased cosmonauts/astronauts/taikonauts who have profiles on WikiTree. It was very frustrating to see how few cosmonauts have profiles on WikiTree, whether in Cyrillic or transliterated to Roman. I hope to see more Russian speakers on WikiTree, so deficits like that can be addressed.

---

My problem on the use of cyrillic is that my husband's grandmother was born in pre-soviet Russia and we only have three in-accurate and differing transliterations.  I do not have the knowledge to work out what would be correct.  For instance, she has sch, bur should that be Shch?  She is the only person with her surname and I cannot believe this is the case in reality.  I would be grateful for any help

Answer 1

The use of non-Latin alphabets is guided by the same rule as the use of various languages: Use their conventions instead of ours. We use the language (and alphabet) that makes sense for the person the profile is about.

To ensure you're not missing duplicates, the best idea is to search for both, because someone may have created a profile using the Latin alphabet where they shouldn't have because they didn't know this rule. While the search will sometimes catch little differences like accents (eg. searching with e vs é), searching for an anglicised Russian name will not bring up results using the original Russian.

Also note that names in additional languages can be added in the nicknames field and other last names field as appropriate, eg. if a Russian spent part of their life in Germany and was sometimes known by a German version of their name.

Comments

Hi Lianne, I guess the only challenge with this approach is I have no idea how to spell a Russian surname in Cyrillic (for example), or a Greek surname in Greek letters, so I would not be able to find the duplicate.  

I think this convention to use "their spelling instead of ours" makes sense, and is workable, when Latin characters are used, and I do this all the time myself when working with profiles.  I don't think this approach works when we start mixing alphabets, however.    As you say, "searching for an anglicised Russian name will not bring up results using the original Russian"

What would be the suggestion here?  I personally think we need more practical guidance on this one as a community.

Leigh

---

My grandfather is from former Yugoslavia ==> his name is Петровиħ  which makes no sense for me but for all the relatives in Yugoslavia ==>

1. I would like to see Momcilo

2. People in Serbia are more interested in Петровиħ

I visited Serbia and found his birth certificate in the Cryllian Alphabet that made no sense for me BUT when I visited the churchyard where people didnt speak english this document was magic and helped me a lot

==> Use all spelling interesting and use the alternative field available right now. In the best of all worlds we should maybe have another field where we can set that this is the name written in a specific alphabet

Youtube Movie when I got the birth cert translated https://www.youtube.com/watch?v=681w8GdfW2w

The Birth cert
http://www.wikitree.com/photo/jpg/Petrovic-32-3

His profile http://www.wikitree.com/index.php?title=Petrovic-32

---

Am I missing something here?  I was under the impression that searches include entries in "other nicknames" and "preferred name" fields with the entry in "proper first name" and entries in "current last name" and "other last names" fields with the LNAB.

As long as that impression is correct, then it seems that the solution is very simple - enter the proper first name, preferred name, LNAB, and current last name in accordance with the "use their convention" specification and enter all other possible spellings and Latin alphabet versions as other nicknames and other last names fields.

Answer 2

We currently have profiles in 7 non-Latin scripts: Arabic الأَبْجَدِيَّة العَرَبِيَّة‎‎‎, Chinese 汉字 or 漢字, Greek Ελληνικό αλφάβητο, Hindi देवनागरी, Japanese 漢字, Farsi الفبای فارسی , and Russian русский алфавит. Without some common denominator, some lingua franca, truely international trees would become unintelligible for the overwhelming majority of users. Most non-Latin scripts have officially recognized Latinized transcriptions which can be put in the Other Names field resulting in a profile such as 毛泽东 aka Mao Zedong, or Влади́мир Ильи́ч Улья́нов aka Vladimir Ilyich (Ulyanov) Lenin. That would make such a profile accessible but not solve the problem of searchability. The more I think about it the more I come back to what seems to me the only solution, though most likely technically difficult to implement: additional searchable fields, non-displayed on the profile, only on the edit page, allowing the input of names in different scripts.

Comments

Thanks Helmut - I think this is a good suggestion.

---

I agree that extra fields are needed. It's not really working for me to place Romanized versions of Japanese names in the Other Last Names and Other Nicknames fields. The Proper First Name field doesn't render properly in search results. It would be more helpful to have optional fields for Romanization, perhaps that only display when other writing systems are detected in the name fields.

---

For better or for worse, many international organizations (I am familiar with the International Standards Organization - ISO and its Moving Picture Experts Group - MPEG) have faced this problem and have ended up adopting English as their standard language for the publication of standards that apply to the entire world. While it may not be necessary to go this far, I believe that when a script other than Roman is used then a Romanized alternative must be provided as well or we will end up with an unworkable Tower of Babel. Admittedly, there are many alternative methods of Romanization for different languages, but perhaps we could adopt one for each of the seven scripts mentioned by Helmut and add others as needed.

Answer 3

Leigh,

My first thought is that, while this is a legitimate problem, it's a relatively small one.  In other words, I really don't think we have many duplicate profiles where one is in the Roman alphabet and the other is in a non-Roman alphabet.  I don't have any numbers at my disposal, but I'd guess that to be very, very rare.

One solution might be to make an exception and allow duplicates which are transliterations into the Roman alphabet.  These could then be identified by a prominent template that links to the profile in the original alphabet.

Other templates might be created to identify profiles which are transliterations (i.e. non-Roman words rendered in the Roman alphabet, or vice versa) and have no "companion" profile in the other alphabet.  That way, someone who knows how to do the transliteration correctly could add one in the future.

I should mention that transliteration is done according to strict rules, and you can't just "wing" a transliteration.  Doing this right takes some expertise.

But as I said, I seriously doubt this is a large-scale problem.

Comments

Thanks Fred, my thought is that this issue will continue to grow as Wikitree expands.  

Take Greece.  Current population is 11 million people.  If every single living Greek had a profile on Wikitree (which we would hope would be the case!!!), then if we allow both alphabets, each person will have two profiles, one in Greek letters, one in Latin, meaning 22 million profiles rather than 11 million.   This does not make sense when we are trying to create one World Family Tree.

Are you suggesting that someone creates a program that allows us to convert Latin spellings into the original alphabet through a template?

---

Maybe I wasn't as clear as I hoped.  No, I wasn't talking about an automatic transliteration template (that would be an impressive trick!).  I just meant one of these colorful boxes at the top of the bio section that might read something like:

"This profile is a transliteration into the Roman alphabet from the original Greek.  To see the Greek original CLICK HERE"

With the link taking you to the Greek alphabet profile.

Server capacity is an issue we'd have to ask the tech folks from Team about,  But I seriously doubt this would pose a big problem.  I recently bought a 4TB external hard drive at a surprisingly reasonable price.  Bandwidth would be the next issue I can think of, but again, it's "Ask the techies."

Answer 4

What I would like to see is some application of soundex search, or a soundex field.   Not sure if only one is needed per profile, or one per name entry. (As we are talking about having an original language script entry and a romanized script entry - not to mention this applies to given names as well as surnames.)

Soundex converts a name to a code on a phonetic basis.

It was used in immigration contexts and especially where either different writing systems were used or where members of the population were unlikely to be able to write their name - at least in a script intelligible to a foreign official processing their entry into a new country.

I don't know if there are particular criticisms of soundex but it has been a very useful tool for my research in other contexts.

I daresay it even helps within an English language context - as many names are equivalent to each other phonetically, but searches here are spelling specific.

Comments

The problem with Soundex is that it is based on English and does a very poor job with many other languages.

Answer 5

I consider myself a globalist with respect to the desire for WikiTree to be a truly global site that works for everyone, and of course, not knowing non-Roman scripts gets in the way of that.  

While I don't think anyone has legislated it, English has become the de facto international language, as French used to be.  In the Indonesia Project we are experimenting with profiles which are intentionally bilingual, using Indonesian first as the language which the individual profiled would have used, and English below that as the international language to make the profile globally useful.  

The main fly in the ointment I can see is that if WikiTree adopted this standard, it would need to use international English as the standard, not American English, and set spell-check to show you as being in error if you used color and labor spellings rather than colour and labour!  

The more immediate globalist challenge is to re-rig the computer so that the surname (LNAB) can display wherever the individual's own culture would have placed it, whether at the back (American style) in front (East Asian style) or in the middle.

Answer 6

One thing to keep in mind is wikitree is not isolated. We should not solve a problem that has already been solved or is part of a standard. I downloaded Gramps, the open source genealogy software, which can communicate (import/export) family trees with other software and wikitree using GEDCOM format. What I saw there was, Gramps have a notion of alternative names. So, a birth name is not a field, it is a type; For the same person, you can have several name records, each with an entirely new name card. When I enter a wife in wikitree and export it with GEDCOM and import it into Gramps, the Married last name is NOT a current last name field. It is an entirely new name card, typed as married name, not birth name (think of it as a hierarchy profile->name card->fields). This issue is very similar, we need a new name card for the name of same person in a different alphabet. So it seems to me, if wikitree adopted the standard that is already out there, we would not have this problem. No I understand creating a new name record is more tedious, when you just need a new field. For another alphabet though, you need several fields, you need prefix, first name, middle name, last name, pretty much all of it. So my suggestion is, allow creation of multiple name cards for the same person/profile (as possible in the standard) That would probably solve the search problem too. Because all alphabets are searchable if the birth names are given. Today we are putting the other alphabet name (incorrectly) in the other last name field, and it is not found in the search there.