Wiktionary talk:About Korean/Romanization

Adjustments to modern pronunciation

“with necessary adjustments based on the phonetics of modern Korean. For example, the standard South Korean pronunciation of 한자 would be romanized hanja, not hancha as given in the 1939 paper, because the ㅈ is no longer tensed.”

The Pyojun's entry says otherwise. Do you have an example of somebody using an adjustment to McC-Rsr motivated by a pronunciation change? The only one I can think of is reflecting official DPRK standard pronunciation in DPRK-related Wikipedia articles. – Dustsucker 2008-07-31

"RR transliteration"

The "RR transliteration" (or erroneously, "RR transcription", from the template) is very confusing. It seems to be based on the RR transcriptions in Korean romanization#Examples, but with little explanation on this page, very few existing pages actually use it correctly. Hbrug 21:11, 6 November 2011 (UTC)

I might work on a Korean autotransliterator later (provided Template:str index is undeleted..) Hbrug 21:32, 6 November 2011 (UTC)

What is very confusing about it? Is it the system itself that's very confusing, or Wiktionary's text about it?

Of all systems for romanizing Korean with which I am familiar, it seems to be by far the simplest. The resulting romanizations should normally be unambiguous, although that doesn't necessarily extend to how the rules are worded.

If you need an explanation, it would be easy to reproduce the rules, but AFAIR there has been a discussion on one edition of Wikipedia about whether that might constitute copyright infringement, because the official rules were not stipulated under a permissive license, which lead to that content being erased from that edition of Wikipedia. I don't know anything about intellectual property law in South Korea and on different Wikimedia projects.

If few pages use this romanization system correctly, that's because earlier editors have made their edits before it was proposed here to use it, and more recent editors were (and are) not just unfamiliar with how it works, but unaware of this proposal and probably even of the system's existence – I've never seen it used anywhere except here on Wiktionary. Cheers, Dustsucker 04:35, 14 November 2011 (UTC)

I think what he's referring to is the fact that we have both "RR" and "transliteration". One of these is the Revised Romanization, but what is the other? It is not described anywhere either here or elsewhere. -- Liliana • 05:52, 14 November 2011 (UTC)

RR transliteration is the best in one-to-one correspondence of letter and sound with some exceptions due to phonological changes. BTW, Google Translate seems to be using RR, only it ignores a few small rules where letters should change. --Anatoli 10:27, 16 November 2011 (UTC)

Sorry, I forgot about Google Translate. So nobody would touch the transliteration for many years, until GT took it up, and that's about it as far as I know. I'm not familiar with the Korean internet, though, perhaps some sites use it in URIs and the like? Dustsucker 12:51, 16 November 2011 (UTC)

It is described elsewhere: English (see § 3.8), Korean (see 제3장 제8항)

You ask “One of these is the Revised Romanization, but what is the other?” – I guess they are two methods intended to complement each other, seeing how they are presented together. What to call each one of them and what to call both of them together is a good question. In August 2003, Menchi started the English Wikipedia article using the lemma “Revised Romanization of Korean”, and it seems to have stuck. I don't know where that name came from, my guess is the “revised” is from “개정” which is used quite a lot in this text that otherwise just calls it a “새 로마자 표기법”. This official ROK document refers to the system(s) numerous times but does not seem to use a particular name. Unicode calls it “KMOCT” here and here. Anyway, the provision in §3.8 clearly allows for transliteration, whereas the other rules amount to a transcription but not a transliteration system, so these two terms could be used to distinguish between them. HTH, Dustsucker 12:51, 16 November 2011 (UTC)

It's worth stressing again that relying on Google Translate alone is not enough to transliterate Korean into RR properly (and not just Korean). Noteably, syllable final g, d, b, need to change to k, t, p. Initial l becomes r between vowels (including h next to another vowel) and at the beginning of a word (foreign words mainly). Care should be taken with consonant clusters where phonetical changes happen, which are also reflected in the romanisation. --Anatoli 22:14, 16 November 2011 (UTC)

While I did not mean to suggest that Google could be used for romanization, I'm afraid you are mixing things up here. Remember there are two different systems found within the ROK's MOCT rules of 2000. One of them is for normal transcription – this is the one where ㄱ becomes g, k or ng, depending on standard pronunciation. The other (§ 3.8) is for transliteration – this is the one where ㄱ is always g, irrespective of pronunciation. Google seems to use the latter, and if it indeed does (I haven't checked), I don't see why it shouldn't be used for transliteration. Dustsucker 17:42, 17 November 2011 (UTC)

(I apologize for writing so much. Please ignore this post unless you think about implementing fully automated transliteration.)
The big disadvantage of transliteration is that it doesn’t tell you much about pronunciation. That can also be an advantage: You don’t need to know about a word’s pronunciation in order to romanize it. For example, for a compound word whose second element starts with a doen-sori-fiable consonant (ㅂㄷㄱㅅㅈ), it may be difficult to tell whether doen sori occurs or not, unless the term is found in dictionaries. Transliteration’s disregard for pronunciation enables automation as suggested by Hbrug above, except you’d still have to decide about other issues such as capitalisation, unless capitalisation should not be used at all – unfortunately, § 3.8 is silent about this, and the examples don’t include proper nouns.[1] For example, it is not clear how § 3.8 would transliterate personal names. According to § 3.4, normal transcription would be Hong Gildong for 홍길동. As for transliteration, if we take § 3.8’s demand that transliteration be “done according to Hangeul spelling” (“한글 표기를 대상으로 적는다”) to its extreme, we might end up with Honggildong, because based only on Hangul spelling, there is no reason to insert a space, or hong gildong, because Hangul spelling alone gives no reason to capitalise, or honggildong. If you think honggildong is a correct transliteration by official rules, automation should be straightforward, as long as text doesn't contain arabic numerals, hanja or possibly problematic punctuation. Dustsucker 17:42, 17 November 2011 (UTC)

Romanization of translations

Korean entries (as in "Korean terms with a dedicated page on Wiktionary") are usually transliterated using four systems: McCune-Reischauer, Revised Romanization, Yale and what is defined as "transliteration". My question is: has any one of these systems been agreed upon for use in transliterating Korean terms in the "Translations" section of a page? GianWiki (talk) 12:11, 16 November 2013 (UTC)