[codyrobbins]

Having an alphabet, as it does, Korean indeed isn’t hard to learn to read and write. The difficulty is learning how to pronounce it properly—it has a number of distinct consonants that are treated as the same consonant in most other languages and that therefore are extremely difficult for non-native speakers to distinguish and produce properly.

For example, it has three distinct unvoiced velar stops [1] which are all considered the same sound in English: normal [k], aspirated [kʰ], and faucalized [k͈].

We use both the normal and aspirated [2] unvoiced velar stops in English as positional allophones [3], meaning we consider them the same sound but they’re actually different and we distinguish between them unconsciously depending on where the sound occurs in a word. For example, /k/ is aspirated in word-initial position in words like cam but unaspirated after word-initial /s/ in words like scam. If you listen closely enough you realize that you turbulently expire a lot of air after the c in cam but don’t do so in scam. Despite using both phones [4] in English, you consider them more or less the same hard k sound—if you pronounce cam with an unaspirated /k/ it may sound slightly odd but you’ll still consider it the same word cam. The problem is that in Korean doing so would result in two different words—so it makes it hard for English speakers to perceive distinct words in Korean that differ this way, and to pronounce the correct word that they intend when talking.

This is true for many languages because phoneme inventories never overlap, but it’s especially difficult for English learners of Korean because Korean has a third /k/ sound—the faucalized [5] one—and it doesn’t occur at all in English. I know phonologists—people who are trained in the study of human speech categories—who have an extremely difficult time discerning between [k] and [k͈].

Peter Ladefoged, one of the world’s foremost phoneticians, has a fantastic website to accompany his phonetics textbook and it has recordings of all the sounds known to be produced in human languages. You can listen to the distinctions between the Korean consonants here:

http://phonetics.ucla.edu/appendix/languages/korean/korean.h...

The /k/ sounds I’ve been discussing are the third row down (weight of measure, rope, and large). As an English speaker try to hear a difference between how the consonant is pronounced—it’s extremely difficult to notice.

Our sensitivity to these phonemic distinctions develops at an extremely early age: by one month old infants already begin to stop distinguishing between different sounds that their language slots into the same category as the same “sound” [6]. Infants don’t even babble at this age never mind produce or understand adult speech, but they have already stopped noticing certain differences in speech—that’s how deeply ingrained the way we perceive and produce speech is, and why it’s almost impossible to speak a non-native language without an accent.

(I am only tangentially familiar with Korean phonology so I might be getting some of the details slightly wrong here—please correct me if I’ve misstated anything.)

###

[1] http://en.wikipedia.org/wiki/Unvoiced_velar_stop

[2] http://en.wikipedia.org/wiki/Aspirated_consonant

[3] http://en.wikipedia.org/wiki/Allophone

[4] http://en.wikipedia.org/wiki/Phone_(phonetics)

[5] http://en.wikipedia.org/wiki/Faucalized_voice

[6] http://www.sciencemag.org/content/171/3968/303 & http://home.fau.edu/lewkowic/web/Eimas%20infant%20speech%20d...

[helpbygrace]

As I am Korean, I also feels same difficulty for English sound between 'r' and 'l'. 'f' and 'p'. because in Korean, 'r' and 'l' have same sound 'ㄹ'

[codyrobbins]

Yes, that’s exactly the same problem illustrated the other way around. Japanese speakers also fail to make the /l/ and /r/ distinction and it’s a rather stereotypical feature of a Korean or Japanese accent in English. Japanese happens to have sounds for both /f/ and /p/ unlike Korean, but not for /f/ and /h/. So you’ll hear Japanese speakers of English having a hard time producing those two sounds distinctly just like Korean speakers have a hard time with /f/ and /p/.

The Korean consonantal distinctions are particular tricky, though, because it’s one consonant in English being not two sounds in Korean but three.

[bane]

Correct me if I'm wrong, but this mis/near-mapping of phoneme sets can easily result in fantastic corruptions of words as they migrate through various language "lenses" then when reintroduced to their origins sound nothing like the original word!

e.g. Moving through a couple cultures... r->l->n and b->p->f or b->v->w might turn "rollerblade" into "lolaweulad"

[codyrobbins]

Actually, that’s more or less how the phonology of languages change internally over long periods of time [1]: sets of phonemes merge, split, appear, and disappear. The particular forces driving these changes are varied and complicated, but particular sounds that have adjacencies on various axes twist and enjoin and diverge, dancing about according to constraints that produce surface complexity like in a multibody system. Something about the way I’ve always envisioned phonological change makes me think of Voronoi diagrams [2].

###

[1] http://en.wikipedia.org/wiki/Phonological_change

[2] http://mbostock.github.com/d3/ex/voronoi.html

[pault]

Brazilians pronounce "vicks vapo rub" like "vicky-vapo-ruby". That one had me in tears the first time I heard it.

[guard-of-terra]

Which is, in fact, strange. Most languages do not have the v/w distinction, but very few non-native english speakers confuse v with w, especially in write. But Koreans and Japanese tend to confuse r and l even in write.

[gurkendoktor]

I am afraid this will be a trivial question now, but does that mean that Korean pronunciation is much harder to learn than Japanese? I often heard that they are very similar languages, and they can certainly sound similar to the uninitiated.

[codyrobbins]

Japanese and Korean are actually not phylogenically related. In fact they’re both generally considered to be language isolates, which means they don’t seem to be related to any other languages. This may not be strictly true and there are proposals that there could be a link through a hypothetical language family called Altaic—they’re usually considered unrelated, though.

The “difficulty” of learning any particular language in relation to another is a bit of a meaningless comparison in my opinion. Ultimately, it appears to be the case that all human languages operate according to a common underlying linguistic mechanism—the seeming differences between particular languages are in fact probably actually just different surface realizations of the underlying mechanism with particular options twiddled in various combinations coupled with an idiosyncratic lexicon (i.e., a relatively unique but ultimately arbitrary vocabulary). The nature of the underlying mechanism is the subject of debate, as well as the list of the particular options available, but it seems pretty clear that all languages are utilizing a common underlying cognitive architecture.

That being said, it would seem to me if you define a relative acquisition difficulty metric as something like

  D(x, y) = a*O(x, y) + b*P(x, y) + c*L(y) - d*C(x, y)

  O(x, y) = cardinality of the set of differences between the language “settings” of x and y
  P(x, y) = cardinality of the set of phonemic distinctions present in y but not x
  L(x)    = cardinality of the lexicon of x
  C(x, y) = cardinality of the set of cognates shared by the lexicons of x and y

where a, b, c, and d are scaling factors, then that might give you some semi-reasonable rough correlation in the form of a “distance” between two languages along various axes. The exact definition of O(x, y) is far from clear, however. And even just having knocked this out I’m realizing that there’s almost an infinite number of other things you could throw in here that you would arguably need to know to claim to have “learned” a language: phonological processes, knowledge of and relative productivity of morphological processes, pragmatics, etc. This is why I’m of the mind that any such measure is of dubious utility, because you would need to take into account a lot of factors—most of which we still don’t fully understand.

There’s probably work done on this somewhere but second language acquisition isn’t my area of expertise.

[dsrguru]

First of all, Korean and Japanese are very similar. They're in different language families because they are not believed to have been evolutionarily related (more closely than other languages in other families), but they are in fact incredibly similar. This might be partly a result of tons of historical contact between Korea and Japan and/or of both sending scholars to China in ancient times, but the mechanism of their coevolution had to be more involved than that. I don't know enough about it, and I'm not sure if linguists do either. However, the fact remains that the two languages in their current forms are closely related, which is what gurkendoktor was asking about in regards to learning time. While the closed class words and inflectional suffixes don't have related pronunciation, the general structure of the languages are ridiculously similar. Word order is identical in many sentences, both use particles or postpositions to mark the function of nouns, both use topics instead of subjects, both allow you to omit the topic if it can be inferred through context, both have a respect hierarchy built into the grammar, both have tons of pronouns and related categories of family words, both have the adversative passive of Chinese, both are agglutinative in the sense that they allow you to add a noun after a verb phrase to form a relative clause that modifies the noun, both have lots of similarly-pronounced Chinese-derived open class words, both have the rare alveolo-palatal fricatives and affricates in their sound inventories as is found in Mandarin, both make the /h/ consonant a voiceless palatal fricative before [i] or [j], etc. Those are just the similarities I can think of off the top of my head. I don't know much Korean, but I'm sure there are a lot more similarities.

As for pronunciation difficulty, I can tell you that from the standpoint of a native speaker of American English who can also articulate every consonant and vowel in Mandarin Chinese, Japanese phonology is far easier for me than Korean phonology. Japanese really only has one consonant whose place of articulation doesn't exist in English or Mandarin (a voiceless bilabial fricative, which is essentially an /f/ but with two lips instead of the bottom lip against the upper teeth) and five monophthong vowels (their position in the mouth stays the same for the duration of the sound), most of which are pretty standard among languages whose vowels are monophthongs. While Japanese does have pitch accent, so does Korean, and unlike Chinese, pitch almost never affects the meaning of a word in the Japanese spoken in Tokyo (there are a couple exceptions) or in the Korean spoken in Seoul. If you want to sound native, you'll have to spend about the same amount of time in either language learning the pitch contour of each word.

Korean on the other hand has more vowels than Japanese, one of which is now pronounced as a diphthong that is very difficult for most English speakers to pronounce correctly (although it's easy for anyone who knows French and easy enough for anyone who can pronounce /y/). The real issue is that Korean has those tense consonants whose specific manner of articulation doesn't exist in any other language. It's the one part of Korean phonology I do not pronounce correctly. Interestingly, I just came across this on Wikipedia:

'An alternative analysis[2] proposes that the "tensed" series of sounds are (fundamentally) regular voiceless, unaspirated consonants; that the "laxed" sounds are voiced consonants which become devoiced initially; and that the primary distinguishing feature between word-initial "laxed" and "tensed" consonants is that initial laxed sounds cause the following vowel to assume a low-to-high pitch contour – a feature reportedly associated with voiced consonants in many Asian languages – whereas tensed (and also aspirated) consonants are associated with a uniformly high pitch.'

So maybe there is hope for us non-Koreans! But yeah, Korean pronunciation is more difficult for most native speakers of English. Also, the grammar requires a little more memorization than Japanese does since there are more inflectional forms in Korean. On the other hand, the Korean writing system is far simpler than that of Japanese since Koreans don't really use Chinese characters anymore at all. It might also take you a few hours or days longer to learn the Japanese syllabaries than the Korean alphabet. As this thread shows, it's incredibly easy to learn to read and write Hangeul.

[sleepydog]

Yes, Korean is much more difficult to pronounce. Japanese pronunciation is extremely simple: 5 vowel sounds, no dipthongs, and a handful of consonants. The only struggle for me as an English speaker was learning not to blend the sounds together or slur. Otherwise it's a piece of cake.

The biggest difficulty in speaking Japanese, after wrapping your head around the grammar, is learning "The right way" to say something according to your standing with the person you're talking to. There's several layers of social cues that you have to navigate to speak fluently.

[codyrobbins]

Japanese has a relatively small phoneme inventory, and luckily for English speakers it’s mostly a subset of the English inventory, but there are a few allophonic variations that we don’t have in English.

Pronunciation is not necessarily so simple, though: native speakers of every language employ a wide and relatively complicated set of phonological processes unconsciously. Here’s a partial list of the Japanese ones:

http://en.wikipedia.org/wiki/Japanese_phonology#Phonological...

Unless you’re unconsciously making these pronunciation changes on the fly, you’re going to have a discernible accent. I cringe when I hear English speakers of Japanese who haven’t internalized the simpler processes pronounce words like 人「ひと」as hee-toe and です as deh-sue. I’ve spoken Japanese for ten years and I still pick up something new every time I listen to a native speaker.

Regarding your other point, I always see people mention how complicated it is to learn to use the correct levels of politeness and honorifics in Japanese. But we encode largely the same information in English regarding relative social positions, familiarity with the conversational partner, etc. It’s really not any more complicated than it is in English, and in fact you could argue it’s less so because we primarily convey politeness with indirection and idiomatically. For example, if a waitress hasn’t brought your change at a restaurant after a few minutes and you say ‘Bring me my change now’ it would be incredibly rude. Instead you would usually say something along the lines of ‘Excuse me, I know you’re busy, but if it’s not too much trouble would you mind bringing our change over when you get a chance?’ You start by asking the waitress to excuse you for your imposition, go on to express that you understand the demands on her time, and then ask her (not command her) if she would mind bringing the change (not if she will actually bring it), when she has a chance to do so (not necessarily right this very second)—all despite the fact that you’re the customer and ostensibly she’s waiting on you. Depending on your tone of voice, the implication is that you really do want the change right now so you can leave but that whole song and dance conveys politeness and respect.

In any event, I have never had a problem speaking to practically everyone in Japanese using plain language. The only time I typically use polite language is with absolute strangers or people who seem in a bad mood, and even then I mostly speak in plain form and if your tone of voice is warm it serves to enact some closeness between you that people seem to appreciate. Very occasionally I’ve met a Japanese that seems to be a bit insulted if you speak too familiarly with them, but, honestly, I can usually chalk that up to that particular person just being a bit of a tight ass. If they persist in speaking to you in polite forms then you just switch back to the polite register to make them comfortable. Also, if you speak only in polite language then Japanese have a tendency to think that you only have a rudimentary understanding of the language. I have never used keigo, ever—if you’re in a business setting it might come in handy, but even Japanese themselves have to be instructed in how to use it when they enter corporate jobs. I can’t find the study now, but I was reading a paper that showed that native Japanese speakers have extremely warm attitudes towards foreigners who use keigo in business settings regardless of how proficient their use actually is or how many mistakes they make. If I remember correctly it was more or less irrelevant how well you used it—the Japanese had a high opinion of you simply for attempting to do so.

[sgronblo]

Yes that is true. Korean has many more similar sounding vowels and consonants and more special pronunciation changes whereas if you can pronounce all the hiragana by themselves you can basically pronounce any Japanese word.

[philsnow]

Dr Ladefoged passed away a few years ago; just thought I'd mention it to you because you said that he "has [sic] a fantastic website".

[codyrobbins]

Yes, I’m aware that he died, but since it’s still up I was thinking that arguably he still has it? I actually was going to say ‘had’ but I felt that would incorrectly imply that it doesn’t exist anymore. Then I was going to put ‘late’ in front of his name but that sounded kind of stuffy. I guess I should have passivized the sentence since if the subject no longer exists it’s ambiguous whether a “possession” relationship can exist between the subject and the direct object any longer.

[poodougnut]

> if you pronounce cam with an unaspirated /k/ it may sound slightly odd but you’ll still consider it the same word cam

/g/ is devoiced word-initially, so you're more likely to think that the word is 'gam'. [1]

Even though the Korean 'normal' [k] is not marked for aspiration, it is too aspirated; it's just that the voice onset time is shorter. [k͈], on the other hand, is unaspirated.

[1] http://linguisticmystic.com/2011/12/31/scuse-me-while-i-mix-...

[michaelmior]

Great addition. Anyone who can say "I am only tangentially familiar with Korean phonology" is likely to be leaps and bounds ahead of the majority of the HN community.

[losvedir]

> Our sensitivity to these phonemic distinctions develops at an extremely early age: by one month old infants already begin to stop distinguishing between different sounds that their language slots into the same category as the same “sound” [6]. Infants don’t even babble at this age never mind produce or understand adult speech, but they have already stopped noticing certain differences in speech—that’s how deeply ingrained the way we perceive and produce speech is, and why it’s almost impossible to speak a non-native language without an accent.

I've heard this before but am confused by the fact that children can learn a native accent (which presumably includes those ignored sounds) for a language they're exposed to many years after infancy.

Does that mean that although babies will ignore it by one month old, they're capable of reversing that if the need arises?

[jcromartie]

As an example, consider the bajillion nearly-equivalent romanizations of this dish: http://ko.wikipedia.org/wiki/%EB%96%A1%EB%B3%B6%EC%9D%B4