Japanese and Korean: testing the links
I think the classification of Japanese and Korean is the biggest remaining puzzle in philology. There are two issues: are these two languages related to each other, and are they part of any known family? The Japanese and Koreans call each other ‘close but distant neighbours,’ and this phrase also roughly sums up the relationship of their languages. You often read that the grammar systems of Japanese and Korean are very similar, but their native vocabularies are very different. This unusual paradox has deterred many language experts from embracing a genetic relationship. Others say that both languages are isolates (orphan languages), or are distant cousins of Turkish, Manchurian and other central Asian languages. On the Internet you still find references to links with Tamil or Malay. No theory has won real acceptance. Even the most elementary questions of affiliation remain unanswered.
I am a Japanese-to-English translator with basic Korean, and this question has become a bit of a hobby with me. But the literature in print and on-line tends to be either superficial or very abstruse—in other words, of little use, or inaccessible to the generalist. Now, I don’t want to take issue with the experts—I know nothing of Manchurian. Nor am I a grammarian, which is why I here use schoolboy European parts-of-speech terms, even though they do not fit Asian languages well. What I want to do in this little essay is try to add some basic comparative detail to this debate, for those without deep knowledge of either language. First I will illustrate the main resemblances of syntax, and then look at the native vocabularies, before finishing up with some thoughts on the impact of their unusual writing systems. Because this is a tricky field, there are bound to be gross simplification, omission, generalization and very likely a few big errors below, for which I apologize in advance.
The similarities of Japanese and Korean syntax are remarkable by any standards. To begin with what is usually considered a superficial point: Word order is almost the same, not just in broad SOV (subject-object-verb) terms but often down to the tiniest quirks of linguistic device. I would guess they are nearly as close as English and Danish at the everyday conversation level. But Danish is not only closely related to English; it seems the Vikings additionally Scandinavianised English word order. That Japanese and Korean have retained a similar degree of closeness down the centuries may not mean anything, as many linguists say, but it is certainly suggestive.
Syntax is the bolts and hinges—the tool kit, if you like—of sentence construction. Some of the basic features of Japanese and Korean are also found in other Asian languages such as Mandarin: lack of plural marker (usually), lack of relative clauses (all subclauses are backed up behind the word they modify), lack of “gender” or other verb classification, few “endings” or inflections generally, simple tense systems and (to the frustration of the translator) a tendency to omit words if they can be assumed. Compared with European languages, all three Asian languages make heavy use of particles.
But Japanese and Korean also share basic features that are unusual in other languages, starting with “topic/subject markers”, which are a bit like articles but have many other functions too. These are found in both Japanese (—wa and —ga ) and Korean (—[n]eun and —i,/ga), and one of them, —ga, is shared. These markers are the most characteristic feature of Japanese and Korean—no other language I know of replicates this system—and they dictate sentence structure. Because of this, we need to see very basically how they work. There isn’t space to discuss the specific roles of each of the two particles. This sentence just shows the two functions for which they are named, topic marking and identification of the subject (of an implied subclause in this case).
It’s not really difficult, just unfamiliar, but things do get much fiddlier in more complex sentences. (The “object marker” is something different and I’ll come back to it later.) The point to remember is that they underpin what is probably a unique system of sentence structuring.
The best-known distinctive trait of Japanese and Korean is the honorific system of verb endings. While Korean can also stick things in the middle of verbs, and Japanese at the front, both systems cover broadly the same bases in terms of register (technically, “formal polite,” “informal polite,” “plain” etc). The “neutral” ending, the one used in teach-yourself textbooks, is –‘mnida’ in Korean and –‘imasu’ in Japanese. In both languages there is a plain form ending in –‘yo,’ though it is blunter in Japanese, (and there is an exclamatory —‘yo!’ as well, as in ‘that’s expensive!’—‘bissandayo!’ in Korean, ‘takayo!’ in Japanese. Both also have special honorific forms for certain common verbs that are essentially completely different words—e.g. ‘kimasu’ and ‘mairimasu’ (lower and upper register for ‘come’) in Japanese. These show no resemblances (between Japanese and Korean), with one very basic exception. ‘Iru,’ to be in Japanese (for animate objects; ‘aru’ for inanimate objects) can be elevated to ‘gozaimasu’ in honorific usage, while its equivalent in Korean, ‘itda,’ inflates to ‘gessumnida.’ (The stem parts here are ‘goz’—- and ‘gess’—-; —‘imasu’ and —‘umnida’ are the endings). The plain form of the verb ‘to be,’ (‘itda’ and ‘iru/aru’) is –‘da’ in both languages (‘it’s a book’ is ‘chaegi—da’ in Korean, and ‘hon—da’ in Japanese).
Another unusual feature of Japanese and Korean is treating adjectives as verbs: you don’t say “the house was big”, but rather “the house bigged.”
Now let’s look at five striking similarities in syntactical areas more familiar to Europeans. Both languages have only one noun “case” ending (particle), the objective (–‘o’ in Japanese and –‘[r]eul’ in Korean; see above example about liking girls with long hair). The yes-no question marker is the same in both: –‘ka?’ in Japanese and ‘–kka?’ in Korean. Korean has a richer spread of tenses than Japanese, which has only simple past, but the very common –‘tta’ plain past form is identical in both. The simple negative in Japanese is –‘nai,’ put on the end of the verb; in Korean it is ‘an’— , put at the front of the verb (‘I’m not going’ is ‘ika-nai’ in Japanese, versus ‘an-gayo’ in Korean). Curiously, these two features—‘t’ (or ‘d’) sounds in past tense endings and ‘n’ sounds in negative endings—are shared by many European tongues too. I’d guess too that the adverbial endings, -ge in Korean and –ku in Japanese, might be linked.
Japanese and Korean tend to put words for ‘although,’ ‘because,’ ‘while’ etc. at the end of the clause rather than the beginning. Several show resemblances: ‘although’ can be —‘noni’ in Japanese and —‘na’ in Korean, ‘since’ or ‘as’ can be ‘node’ in Japanese and ‘nikka’ in Korean. ‘Because’ can be ‘tameni’ in Japanese; Korean has ‘ttaemune.’
Both languages have multiple personal pronouns (depending mainly on register), but prefer to avoid use of ‘you,’ which is considered too direct. There are a number of lookalikes: ‘We’ can be ‘ware’ in Japanese, and is usually ‘uri’ in Korean. He is ‘kare’ in Japanese and ‘geu’ in Korean (‘karera’ and ‘geudeul’ are ‘they’). Paralleling this are the words for ‘this’, ‘kono’ and ‘kore’ in Japanese and again ‘geu’ in Korean (Korean also shares with Japanese, but not Mandarin, the feature of having two subtly different ‘there’ words). Locational words (European ‘prepositions’) show many resemblances. Two are almost the same: (Japanese)—‘ue’ / (Korean)—‘wie’ (‘on, above’), and —‘e’ / —‘e’ (‘towards’). Others are similar—‘mae’ / ’ap[e]’ (‘in front of,’ assuming a shift from ‘m’ to ‘p,’ which I’ll touch on later), ‘ato’ / ’dwie’ (‘behind’), ‘naka’ / ’ane’ (‘inside”) and ‘yoko’ / ’yeop[e]’ (‘across from,’ ‘next to’).
There are also a number of shared linguistic devices that are more difficult to explain, such as verb endings that express surprise or doubt and particles with no English equivalent. I give only one concrete example here, the “reported speech particle.” This is added to the end of statement to suggest something like the bracketed (‘I hear that’) in ‘(I hear that) he has been promoted.’ In Japanese, this particle is ‘to,’ and in Korean it is ‘rago;’ for some reason both seem related to the words for ‘along with’ in both languages (‘to’ in Japanese and ‘hago’ in Korean).
These then are the main syntactical similarities that strike me after ten years of translating Japanese and three years of on-and-off private study of Korean. I could lengthen the list, but I think I’ve said enough to suggest that there must be some sort of relationship between these two languages. For reasons I have never fully understood, syntactical resemblances are not considered significant by many linguists. My own impression, of course, is that in this case they are significant. This is much more than the broad resemblance of grammar between Mandarin and either Korean or Japanese. In many, perhaps most, areas of syntax, Japanese and Korean mirror each other in a way that otherwise only occurs with close genetic affiliates; the shared use of the unusual topic/subject marker system alone strongly suggests affiliation. (By the way, it is a myth that Chinese is “closer” to English than it is to Japanese or Korean; there is only one significant grammatical similarity with English, albeit a very important one for translators: word order).
Where linguists seek truth is words. So what about the native lexicons of Japanese and Korean? Before plunging into this, a few qualifications. Firstly, we must ignore the huge stock of Chinese loan words shared by both (though it is interesting and relevant to compare the phonetic treatment of Chinese words in both tongues). Secondly, neither Japanese or Korean has what westerners understand as an established etymology for native words—if they did, I wouldn’t be writing this. So there is no easy way of referring back to ‘proto-languages’, or of drawing family-wide comparisons as you can with, say, Romance or Turkic languages. Thirdly, Japanese and Korean are written phonetically in different scripts that represent somewhat different sounds, a source of distortion that I worsen by turning both into Roman script. Which leads to point four: although Japanese is simple to romanize and sounds much as the letters suggest, Korean is not and does not. Roman script cannot easily accommodate Korean quirks of aspiration and voicing (listen to the difference in English between Clay’s bout, Clay’s pout and clay spout to get an idea of the distinctions), so that ‘g’ ‘b’ ‘j’ and ‘d’ are often more like ‘k’ ‘p’ ‘ch’ and ‘t’. Also, some vowels in the latest official Korean romanization are phonetically misleading, especially the very common ‘eo’ and ‘eu’, which are not ‘ee-oh’ and ‘ee-ooh’ but simply variants of ‘o’ and ‘u.’ (In fact, romanizing Korean generally is a trail of banana skins and I have probably slipped up a few times). All of these factors blur comparisons of individual words when written in Roman script.
I said at the top that Japanese and Korean have very different native vocabularies, by which I meant there are few cognates. In fact, it is all too easy to find candidate cognates. For example, looking at column two on page 46 of the Periplus Pocket Korean Dictionary, I notice that ‘tteonada,’ depart, looks a bit like ‘deru,’ ‘leave,’ in Japanese, when you compare the stems which are ‘tteona’— and ‘de’—. A few entries down I see ‘tteoreotteurida,’ ‘drop,’ which reminds me of ‘taoru,’ ‘to fall’ in Japanese. And what about ‘tteoreojida,’ ‘finished, used up’—doesn’t that resemble ‘tariru,’ ‘be enough’ in Japanese? Moving down the column, there is ‘ttulta,’ ‘pierce,’ which calls to mind ‘tsuranuku,’ ‘pierce, go through.’ More speculatively, I see ‘tusu,’ ‘pitcher,’ and think of ‘taru,’ Japanese for ‘barrel.’ Reviewing, I spot ‘ttatteuthada,’ to be warm; the stem, ‘ttatteut,’ closely echoes the Japanese word for warm, ‘atatakai.’ And so on. You can amuse yourself for hours in this way, confident that at least a few of your matches probably are hits (I’m particularly attracted by ‘ttatteuthada’ / ‘atatakai’ for some reason). Problem is, there is just no way of knowing.
Nonetheless, certain putative cognate pairs have won a degree of acceptance over the years, presumably because they are mostly nature and body words, the “test beds” favoured by European comparative philologists. Some examples are (Korean followed by Japanese): ‘mul’ and ‘mizu’ (‘water’), ‘mom’ and ‘mi’ (‘body’), ‘seom’ and ‘shima’ (‘island’), ‘iri’ and ‘inu’ (‘wolf’ and ‘dog’), ‘gom’ and ‘kuma’ (‘bear’), ‘ip’ (‘mouth’) and ‘iu’ (‘say’), ‘geot’ and ‘koto’ (thing, and also a grammatical particle in both languages), and ‘nat’ and ‘nata’ (sickle, blade).
The usual objection to putative cognates is that they could be borrowings. In the case of the animal and farming words above, this is quite possible, as Korean technologies influenced Japan round about the fifth century. But for words as basic as ‘thing’ and ‘water,’ it seems less likely. Either way, the main argument against borrowings is whether systematic sound correspondences exist, across a range of words. How do Japanese and Korean fare in this respect?
Again, the overall picture is suggestive, but the leads are fragmentary. There is some evidence for at least one sound shift, involving Korean initial ‘p/b’ and Japanese ‘h’ (a shift found within Japanese too). It is seen in pairs such as ‘bul’ and ‘hi’ (‘fire’), ‘baem’ and ‘hebi’ (‘snake’) ‘byeol’ and ‘hoshi’ (‘star’), ‘byeo,’ and ‘ho’ (‘rice ear’), ‘bom’ and ‘haru’ (‘spring’) and ‘bari’ and ‘hae’ (‘the fly’). ‘M’ and ‘p/b’ also show a good deal of interchangeability between and within Japanese and Korean (eg. in Japanese the character ‘bun,’ learning, can be read ‘mon’). There is also the sequence ‘sul,’ ‘sol,’ ‘dal,’ ‘gul’ versus Japanese ‘sake,’ ‘sugi,’ ‘tsuki,’ ‘kaki’ (‘alcohol,’ ‘pine,’ ‘moon/month’ and ‘oyster’); also, perhaps, —‘dul’ and —‘tachi,’ the plural markers. But there is also this set: ‘nal,’ ‘dol,’ ‘seoli’ versus ‘nama,’ ‘tama,’ ‘shimo’ (‘raw,’ ‘stone/ball’ and ‘frost’), and—an odd one, this—‘dari,’ ‘dari,’ ‘darri-‘ versus ‘hashi,’ ‘ashi,’ ‘hashi’- (‘bridge,’ ‘foot’ and ‘run’). Interesting pairs are ‘ttae,’ ‘dae’ versus ‘toki,’ ‘take’ (‘time,’ ‘bamboo’); ‘gureum,’ ‘ssireum’ versus ‘kumo,’ ‘sumo’ (‘cloud,’ ‘wrestling’); ‘chupda,’ ‘jopda,’ versus ‘samui,’ ‘semai’ (‘cold,’ ‘narrow’) as well as ‘ap’ and ‘mae’ (‘in front of’); and ‘eonje,’ ‘meonjeo’ versus ‘itsu,’ ‘mazu’ (‘when?’ ‘first’). Finally, compare ‘eodi,’ ‘eotteo-’ and ‘eoneu’ with ‘doko,’ ‘dou’ and ‘dono’ (‘where,’ ‘how’ and ‘which’).
So much for the comparative cherry-picking. Yet it bears repeating that, despite the above, the majority of words in Japanese and Korean look and sound very different (even the shared Chinese words often sound different, mainly because Japanese has lopped off all those –ng endings). So how do we reconcile the similarity of grammar with the divergent lexicons?
I can’t offer any new theories, but would make a few points. First, if Japanese and Korean are related—and that is my belief—they separated at a very early date, and are the only two significant survivors of whatever branch of whatever language family they belonged to. This important point bears repeating: the absence of other family members makes it impossible to confirm cognates and sound shifts by cross-referencing. If Indo-European did not exist, I suspect it would be quite difficult to conclusively link English with, say, Albanian, or even its traditional neighbour Welsh, although all are now known to be IE tongues.
If asked to assess the relationship of Japanese and Korean in Indo-European terms, I would perhaps cite English and Russian as a rough parallel. They must have parted ways at least two thousand years ago. In the absence of solid historical documentation, I doubt that “proto-Koreans” invaded Japan (or vice versa) and imposed their language as the Romans did in Gaul; more likely they were two branches of one people who settled eons ago on opposite sides of a sea and grew apart—a bit like Finns and Hungarians, perhaps. That’s little more than gut-feeling, though.
The other point I want to look at is the role of East Asian writing systems, particularly in the first millennium. In western European countries, the use of a single phonetic script must have helped stabilize pronunciation and moderated phonetic divergence among languages more or less since the Romans faded away. Nothing like this happened in East Asia, for the main script was for centuries Chinese. Chinese characters are semantic symbols, and exert little influence on pronunciation, so that even many of the ‘dialects’ of Chinese have evolved into mutually unintelligible languages. In these unusual circumstances—most of the world’s other major scripts have long been phonetic—it is possible to see one reason why Japanese and Korean words, lacking the binding force of a common phonetic script, might have drifted far apart over time.
Unfortunately this idea cannot be tested as we know virtually nothing of early spoken Japanese and Korean. This is of course another byproduct of use of Chinese. For centuries, Japanese and Koreans wrote in Chinese, or contrived to make Chinese characters express their languages phonetically and carry their grammar, an insanely complex business that must have barred all but a tiny elite from anything approaching literacy. Scholars have sweated blood trying to figure out Japanese and Korean phonetic and grammatical rules based on the way characters were used, but to little avail. All we can really do is speculate.
So here is one last speculation. For both the Japanese and Koreans, one of the consequences of adoption of Chinese writing was that they learnt to think of language in terms of syllables, not letters (a Chinese character cannot be reduced to its constituent ‘letters’). So when the Japanese came to inventing hiragana, the Japanese phonetic writing system, they duly used syllables—usually a consonant followed by a vowel (‘to-yo-ta,’ ‘ya-ma-ha,’ ‘mi-tsu-bi-shi’). In other words, the Japanese writing system forces words into a fixed set of syllables, each of which must finish on a vowel (or an ‘n’). It seems to be the consensus that Japanese has in fact always been a syllabic language, with syllables always ending on a vowel. But is it possible that the Chinese-derived syllabic straitjacket has flattened what was once a richer and more complex phonetic system?
Korean faced the same problem as Japanese, but found a different solution. Hangeul, the native phonetic writing system, were invented in the 1400s (an extraordinarily late date when you think about it) and are both a syllabic and alphabetic writing system. Hangeul characters are structured in the same way as Chinese characters—single squarish units written from top to bottom, left to right. But hangeul are not ideogrammes, but little phonetic packages containing their constituent ‘letters’ bundled up inside them—imagine ‘cat’ with the ‘ca’ written above the ‘t’. It is possible that, because of this quality, hangeul were able to preserve sounds in Korean that have been lost in Japanese. It is certainly striking how much richer phonetically Korean is today than Japanese.
The interesting thing is, the core sounds of Japanese or Korean are similar, as anybody who has heard Japanese or Koreans speak English can probably testify—their accents can be hard to distinguish. Neither have a ‘European’ f or v, or a ‘European (rolled etc.)’ r sound (both have a sort of combined r and l sound—the one that gives rise to all the flied lice jokes), or any true consonant clusters. They both also have ‘tensed’ doubled consonants (distinctive but hard to explain), such as in the shared –‘tta’ ending mentioned above. Words have “level” stress in both, and their intonation, in particular the way sentences are pronounced phrase by phrase, is also similar. As with syntax, it looks as though phonetically too both languages worked from same tool kit.
I said at the beginning that the classification of Japanese and Korean is the biggest remaining puzzle in modern philology. In fact, I should have written ‘ought to be’, as the big questions (unlike the minutiae) have received little informed attention really. Outside the region, I suppose this is mainly due to eurocentricism and the sheer inaccessibility of Asian languages without years of study. Within the region, the major obstacles are cultural. Neither the Japanese or the Koreans are comfortable with the idea of a genetic link between their languages, and what it would imply, because both peoples like to think of themselves as “unique”. It doesn’t help either that they don’t like each other very much. So, as far as I can make out, few people anywhere are seriously researching the genetic affiliation of their languages.
I have tried to be as comprehensive and accurate as possible in this short essay, but its purpose is not really to educate. There isn’t much new or original here. I just want to publicise the issues and get others interested. So, if you have deeper or supplementary knowledge, please correct my mistakes, or contribute comment, at firstname.lastname@example.org. I’m particularly interested in information on Mongolian and other North Asian languages.
Please see some ads as well as other content from TranslationDirectory.com: