Dieter: The discussion below, about Romanization of Korean names and terms, reached a lull on Feb 2004. The participants seemd to have come to an agreement but I haven't seen any page where the agreement is outlined for future use.
Alex Weldon: A lot of these (the names at Korean Go Terms -dv) are poorly romanised. Most mistakes had to do with putting in the letter "h" where it doesn't belong (probably to try to clarify a vowel, but aspiration is important in Korean, and an "h" in the wrong place might give the wrong impression) or with the character "ryeol" (usually equivalent to R/L in English) appearing after another consonant (in which case it's actually pronounced as an "n," but here it had been romanised as an "l," which is misleading). I've fixed up the English spelling/pronunciation. (02/09/2004)
John F. There is a duty of care to the general reader when presenting Korean. I'm not convinced that the presentation here discharges that care.
Korean is a specially difficult case, of course, because of the unusual (relative to English) range of consonants and the sound changes that take place depending on position in a word - changes that are not apparent from the spelling. Korean may well hold the record for number of different romanisations.
But a romanisation suitable for someone learning Korean full-time (various university courses) is likely to be different from someone living in Korea though not really learning the language (various methods devised by the US military for their soldiers in Korea). That in turn is likely to be different from a guide-book for tourists, and there will be differences again for non-specialists (e.g. newspapers) who just need to refer to Korean in passing - and especially those who lack access to accents. Standards organisations and libraries have their own requirements. Overlaid on all of this are the multiple official and unofficial systems used in Korea, which seem to depend more on feng shui than common sense.
I don't know the best answer. For GoGoD I have adopted McCune?-Reischauer, and various others such as Go World have copied that. I have explained the reasons elsewhere. It is a reasoned case, but not a watertight one. There may be a better choice. But I think it has to be explained. The method above lacks any explanation so far. If the justification is to be that it helps with the pronunciation, which seems to be the implication of the comment on 'h', I'd find that hard to accept - how does a non-specialist pronounce 'eo'? How does a non-specialist who is willing to take some trouble to look up words (e.g. deg here) work out from kon-nan that he's really supposed to look up kon-lan (and that's one of the easier ones)? If this is a method used at X university, what relevance is that to the general reader? Mc-R is accepted as a standard by the Korean government and the major standards organisations and libraries. It makes look-up easy and can be used with or without accents. Its only real drawback seems to be a weakness in showing pronunciations. But given that no other system is really much better, that doesn't seem to be much of a handicap.
Alex Weldon: Okay. I should have added a pronunciation guide, and will do that here. The system I use is the "standard" romanisation, that is, the system that is used for road signs, etc. in Korea. The advantage is exactly that, that it's standardised, whereas if people improvise and try to spell things phonetically in an ad hoc sort of way, it leads to ambiguity.
As for your point about difficulties in looking things up, I would agree, except that this page also gives the hangul (which should actually be spelled hangeul, in this system) spellings, so there is no ambiguity, unless the person in question is using a romanised dictionary, which is a bad idea to begin with, since many consonnants are supressed at the end of words unless there's a suffix, at which point they reappear, etc.
Without further ado, here is the pronunciation key for standard romanised Korean.
All the consonnants are as they appear, with the exception that k, p, t and ch are all aspirated (breathe out when you say them) and g, b, d and j are actually halfway between g/k, b/p, d/t and j/ch when appearing at the beginning of a word or after a vowel.
Anyway... that's the system used on street-signs, and it gets my vote. However, you're free to make a counter-proposal if you don't like this way, and we can start a discussion. Once everyone agrees on one way, we should have a page for the pronunciation key (linked to from any page that uses Korean words), and everyone should stick to it.
My intention in "fixing" the romanisation here wasn't to impose my system on everyone... simply, there were about three competing systems in use on this page, and I took it upon myself to at least put them all in the same system.
John F. I had already made my counter-proposal in effect. Can I refer you to the page Korean Names, Alex? There I give just some examples of the variety of romanised names met with in the English go literature. Names are the most common Korean terms any western go player is going to meet, and that seems to me the best starting point.
What you refer to as the standardised system is far from that. The Korean government proposed it, but the Korean people haven't accepted it. If you look at official government publications you will see Mc-R. E.g. if you get tourist literature from the KTO you will see mostly Mc-R, though it has to be said the hotel names will usually be in one of the idiosyncratic variants.
We don't have to follow the Koreans, anyway. We use Hepburn for Japanese, but the Japanese have their own rather different system.
None of this is easy!
Alex Weldon: Honestly, I don't care how it's done as long as it's consistent. I looked at Korean Names, as you asked, and the vowels, which we both agree are the hard part, are exactly as I have them, so I'm not sure what your point is. The only difference between the system used on Korean Names and what I consider "standard" is that whoever created that page uses the older style of consonants. That is (with that system on the left, and mine on the right):
If you want to change all the romanisation to that system, I'd be more than happy to stick with that in future. My main objection was the haphazard insertion of the letter "h". Like, I'd prefer to see the word for apple romanised as sagwa, not sahgwah, since there are no hyeots (hangeul equivalent of H) in the Korean spelling.
John F. The system is not 1 to 1, I'm afraid. If it was, maek above would be maeg (which you do see), and also the odd cases like matbogi would come out at mas-po-ki if the hyphenated form of McC-R is used, which is what we have used for names (partly to distinguish them from Chinese). Of course, Cho Hun-hyeon is not going to listen to either of and use either that form or Jo Hunhyun, (though other Koreans might!).
The McC-R system as used by GoGoD has been taken up by quite a few other people, but is still far from universally accepted. Still, one thing we can claim in its favour is our consistent use of it. From what I have seen, most other databases or sites, for example, have used a mix of various systems, often with the result that the same player appears in several different guises. It may well be that McC-R is on its last legs and that the Korean government will one day command support for its so far very variable efforts to impose a standard. At least when that day comes, those of us who use a consistent system can write a simple script to change over in an instant.
It's not a question of imposing one particular system. It's a question of consistency. The method above is not yet quite consistent in two ways: examples such as maek and the fact that the romanisation is not actually a romanisation of the hangeul but a representation of the pronunciation, which is fine so long as that is made clear.
McC-R is not perfect either, of course, and in the various GoGoD dictionaries I try to have my cake and eat it. I use McC-R as the base, but I also list variants from other systems, and add pronunciation notes where necessary!
Alex Weldon: Yeah, it's not a simple issue, and what we're really talking about is not just the system used, but the whole purpose behind giving an approximate romanisation of the words. The trouble with changing Hangeul into romanised letters (and, indeed, vice versa, perhaps even more so) is that there's something to be said for always representing a given character with the same letter(s), so it's clear which one it is... but on the other hand, isn't the whole point so that people who can't read Hangeul know how to pronounce the words? Spelling matbogi as "maspoki" is more misleading than helpful.
Perhaps the solution is to provide both, eg. (맛보기) : Mas-po-ki (pron. MAT-bo-GI)
In that case, though, both the spelling romanisation and the pronunciation romanisation should be consistent throughout. No "maht-bow-gee".
John F. Thanks, Alex. We now seem to be very close on this knotty topic. It was more than a couple of years ago that I researched the topic of Korean government romanisations. It may be time for another look - do you know much about this? Last time I hoofed around libraries, embassies, etc. This time I think I'd rather be an armchair browser :)
John F. An update. I did some further research. A little bit of background may help first. The Korean government chose McCune?-Reischauer romanisation in 1984 and used it in textbooks, maps, roadsigns, etc. Major western institutions also used it, so it had a lot going for it. But Korean individuals totally ignored it when writing their names. In July 2000 the Korean government changed their minds and came up with a new system (the one Alex used first). This is in place already in textbooks, and has to be used on roadsigns, etc by 2005. Though compulsory, this system allows names to be exempt, but the goverment undertakes to make strong efforts to get names to conform. Ordinary Koreans have no intention of conforming with their own names. Indeed, many have objected vociferously to the new system even for road signs. Worse, perhaps, is that western institutions seem strongly disinclined to follow the Korean government and will apparently stick with McC-R.
So we seem to have gone from a dog's breakfast to a dog's regurgitation of its breakfast.
The main questions now seem to be: to what degree is the Korean government really pursuing its plan to persuade people to standardise their names, and to what degree anyone is taking any notice? As far as I can see, the answer is close to zilch. Anyone know any better?
En passant, it is well known that Lee is a very common non-standard option for Yi, but I came across some figures on this for the first time. 95% of the Yis use Lee, and the rest choose one of Rhee, Li, Yi, I, Lie, Leigh and Ri. They would have to use I in the new system (Mc-R has Yi, which is actually an exception in that system). Those who like Park for Pak (which must be disconcerting to Americans as there is no r in the sound), would have to become Bak. Cho would become Jo, and since we get lots of people thinking Go Seigen is Mr Seigen, I imagine we'd get even more people thinking Jo Hun-hyeon has a personal name Joe.
For those who wonder why we stick with Yi Ch'ang-ho when he writes Lee Changho, it's because of the desirability of standardisation. If we knew how all Korean pros, past and present and future, choose to romanise their names we could take these as the standard. But in practice we know only a tiny percentage, and even those we know often appear in different versions, being changed rather in the way that some people change hairstyles.
For the time being I'm sticking with McC-R as the main option for names, but as ever I will continue to list all recorded variants in the names dictionary.
Alex Weldon: I don't know much of the politics or history behind it. What I can say is that I was living in Korea up until six months ago, and all the road signs I ever saw were in the new system, while most people I knew still spelled their names however they felt like it (Park, Lee, Choi, etc.). So, the government has in fact managed to change all the road signs, but people seem understandably reluctant to change the spelling of their names.
John F. If I understand correctly, the politics of it is that the latest change was undertaken by Ministry of Culture and Tourism, who set up a small board that excluded even other government ministries. But it also refused to canvas opinion from western authorities, even language professors working in Korea. The Ministry of Education went along with it, probably in return for MOCT supporting them in 1984 - that's the way the Oriental civil services work. Since they excluded a lot of people they've upset a lot of people.
Actually, the system proposed doesn't seem at all bad. As to the road signs, the rural areas haven't caught up, I read, but apparently all road signs are changed anyway every five years. So by giving them till Dec 2005 they reckon they can avoid any extra cost for making this change, if the normal rotation is followed. Textbooks have already changed over, but I suspect this is not specially significant. They can't be using romanisation that much in Korean schools, can they? Indeed, schools, along with companies and individuals were specially exempt from making the changes in their names. I don't really understand why schools were singled out like this, but it seems to be something to do with alumni groups being important in Korea.
Alex Weldon: Well, for schools it might also have to do with accreditation in other countries. "Sorry, we do acknowledge degrees from Chosun University, but it says here that you graduated from Joseon University."
ViciousMan: For Korean names of people, organizations, and companies, we ought to use the most common romanization in use and/or preferred regarding the person and/or organization, regardless of the conformity to either M-R or RR. The fact of the matter is that Koreans do not aim to conform to either; Kim Jong-il would be Gim Jeong-il in RR and Kim Chŏngil in M-R. "Kim Jong-il", by far, is the most common form of his name. With Hyundai... Hyeondae in RR and Hyŏndae in M-R. Get the picture?
BTW, for Chosun University, use "Chosun University."
ViciousMan: For Go terms, use RR (which we already do, for the most part)