The impression one gets about Japanese writing systems—note the plural—is that the Japanese went to a restaurant and couldn’t decide on a meal, so they ordered multiple meals for everybody. It’s a case of overkill on an epic scale.

To use a fictitious example for comparison, let’s make believe the Koreans not only invented their hangul alphabet, but ALSO kept on using Chinese characters, and ALSO used a Romanized alphabet—and then mixed-and-matched ALL THREE systems at once, to a greater or lesser extent, every time they wrote a sentence. The Korean example is fictitious, but the Japanese do it all the time—it’s business as usual.

There’s not too much for us to do with Japanese, since we’ve already done almost everything we need. We’ve already discussed Chinese characters, Romanization, and syllabaries, all of which involve Japanese. There’s very little more to say.

I can put it this way: I have bad news; I have good news; I have so-so news. The bad news is that the backbone of Japanese writing remains Chinese characters. The good news is that Romanization is widely used. The so-so news is that there’s a syllabary system in use in between. Actually, there are two related syllabaries, but because of their similarity I would NOT put the total number of Japanese writing systems at four, but rather at a mere three—if that’s cause for any relief.

Kanji   What the Chinese call hànzi the Japanese call kanji, which translates, not surprisingly, as “Han characters”. However, there are adaptations, including some new kanji characters created in Japan for Japanese use, some kanji drawn differently, and some kanji used with different meanings from the original. Kanji entered Japan, where no writing system had existed before, from China via Korea in the 5C, probably by missionaries, but also likely by merchant traffic.

Kanji also have pictograms (shōkei-moji), ideograms (shiji-moji), compound ideograms (kaii-moji) as well as the other categories of more complex forms. I’ve reviewed what I can find, and have found a few more interesting ones listed under Japanese kanji that I’m sure work equally as well as Chinese hànzi.

Let’s start by extending the meaning of some old friends, 日 sun and 月 moon. I now find that, in the context of measurements of time, 日 also means day and 月 means month, as in “Tomorrow is the first 日 of the 月. There is a lot of logic in this. Also, in the context of the days of the week, the symbols in question refer to two of the days out of seven—I’ll let you guess which, as in “Call me on 日 or on 月”. Exquisite.

Another old friend is 山 mountain. I don’t think that this character, 石 stone, is particularly easy, or worth trying to remember. But I do like what you get when you combine the two, showing stone underneath a (compressed) mountain:

Here’s an obvious pair:

This one also seems obvious—I see water flowing between two banks:

Look at this pair. In the first one, do I see a person (人) under a “ceiling”? And in the second character are drops falling from that “ceiling” obscuring the original view?

Finally, I think this is my favorite of the “new” Chinese characters I found among the kanji, not only because it includes three old friends, 山, 上, and 下, but because of the simple, yet concise logic. The mountain-climbing narrative I wrote in 2009/31 notwithstanding, most people going up and down over a mountain do so over a:

Still, it’s a ponderous system, and after the war, in 1946, the Japanese government worked on simplification of kanji to make it easier for kids to learn, and for more ease in general in reading periodicals and literature. The number of characters was reduced and variant and obscure forms were discouraged. I have to say that this sounds like what the Chinese government did in the 1950’s with simplified Chinese, as well as instituting pinyin to replace Wade-Giles.

But the arguments for reform or even elimination of kanji continue: (1) there are too many kanji; (2) they are difficult to learn, to read, and to write; (3) the Roman (Latin) alphabet is used internationally, and using kanji separates Japan from the rest of the world (except China); (4) processing kanji on a word processor and a computer (which IS possible, I understand) is more time-consuming; (5) text blending kanji and kana (syllabary symbols) is inefficient compared to text using kana exclusively or rōmaji (Romanized letters) exclusively.

Rōmaji   Romanized Japanese script is called rōmaji. It’s the word “Roman” without the N, but plus JI. The macron over the O indicates a long vowel, not a tone.

I never realized until now how parallel the situation is with that of China. Both have tried to simplify the Chinese characters and both have tried to simplify Romanization, but with different success. In China, pinyin has successfully replaced Wade-Giles. That’s not exactly the case in Japan.

The traditional Romanization is based on the updated system of an American missionary, James Curtis Hepburn, who published a Japanese-English dictionary in 1887 using his system of Romanization. The Hepburn Romanization System is called in Japanese the Hebon-shiki Rōmaji, with an interesting variation on his name. It is by far the most widely used Romanization system. Its defect is that it’s based on the phonology of English, not Japanese, and therefore varies in accuracy. For this reason there has been some opposition to it in Japan.

In 1937, an ordinance was issued by the Japanese Cabinet for a Romanization system that was more faithful to Japanese called Kunrei-shiki Rōmaji, or Cabinet-Ordered Rōmaji. That ordinance was overturned after the war during the Occupation of Japan, but it was then reissued in 1954. Although a revised Hepburn was codified and proposed in 1989, it was rejected in favor of Kunrei.

This sounds just like Wade-Giles versus pinyin, but in Japan, there’s been an odd quirk. Although Hepburn lacks de jure status, it’s the de facto system of choice. Even the Japanese government mandates passports to be issued in Hepburn, and the Ministry of Transport uses Hepburn on road signs and in rail stations. Although Wade-Giles has died a peaceful death in China, Hepburn is still going strong in Japan.

Kana   In between the traditionalism of kanji and the internationalism of rōmaji lies a home-grown system. It’s syllabary-based, and will therefore remind you of the home-grown Korean system of hangul, except that kana are not written in syllable blocks but in a straight line. Still, there’s another complication. There are two kana systems, similar to each other: hiragana is the primary kana system, and katakana is the supplementary kana system. The two kana systems, both developed in the 9C, have 46 characters each. Here is the Kana Syllabary. The two extra combinations shown in red, WI and WE, are not in common use anymore. In each box, the hiragana is on the left and the katakana is on the right.

Let’s confirm that Japanese works well with a syllabary. Japanese, like other, totally unrelated languages we’ve seen, particularly Hawaiian, is a LA-LA language, with the simplest of syllables. The basic form is CV, or one consonant (first) and one vowel (last). Hawaiian does it in ho.no.lu.lu and Japanese does it in sa.yo.na.ra. You can also have just a vowel (V), as in Hawaiian a.lo.ha or Japanese o.sa.ka. (Both also have diphthongs, like mau.i and to.kyo, but let’s not concern ourselves about that now.) Still it struck me that Japanese has one standout difference. If you don’t see it, look at the chart. A Japanese syllable can optionally end in an N, either as CV(n) or V(n). Look at the Shinto (shin.to) religion; at Shinkansen, the Japanese name for the Bullet Trains (shin.kan.sen); at Asahi Shimbun, one of the leading newspapers (shim.bun). (Whoops! We need further explanation. I would say the M you see in Shimbun is the fault of Hepburn Rōmaji. It leans toward English, where an N before a B will appear as an M—say “pinball” and it will come out PIM.ball. I will go out on a limb and wonder if Kunrei might keep it as an N.) In any case, you will note that, to accommodate this quirk of ending a syllable in an N, there is a separate line in the syllabary so you can add an N where needed.

Check it out—try finding the four hiragana symbols to write “sayonara”.

Using Blended Systems   We haven’t yet said how all this comes together. Actually, it’s not really as complicated as it seems.

Kanji remain the backbone of the language. Kanji are used for nouns, adjective stems, and verb stems. The problem with them, though, is similar to why Arabic script was inadequate for Turkish. While kanji (hànzi) fits Chinese well, they are inadequate for Japanese. Japanese has grammatical endings for verbs and adjectives, and kanji fall flat in being able to represent that. But hiragana can. Therefore, a text in kanji will be sprinkled with hiragana to take care of grammatical endings.

I picture it this way. We can write “seven” as the logogram 7, but if we want to write “seventh”, the logogram itself isn’t enough, so we write 7th, adding alphabetic symbols to the logogram. Similarly, if you have a pair of fours, you can write “fours” as 4s. I think we can safely say hiragana work that way along with kanji.

Also, there are some words that there simply are no kanji for, or for which the kanji are overly difficult, so they’re written in hiragana. Hiragana is also used for the Japanese equivalent of small connecting words like “on” and “to”. (Hiragana is also the basis for Cyrillizing Japanese.)

The reader will consider what we’ve said so far and realize that hiragana is the string that ties the entire kanji system together. Perhaps we should call the system in daily use kanji-hiragana, but that’s just my term. However, there are special occasions. Not only can the whole language be written in hiragana alone, it’s actually done in certain cases. Foreigners can be eased into Japanese with texts just in hiragana, and so are native schoolchildren, who also learn hiragana first. As a matter of fact, children’s books are often written just in hiragana. I wonder if the kids enjoying a simple reader in hiragana realize what they’ll be up against when kanji are introduced. But once we go back to the full kanji-hiragana, we have to now see how rōmaji and katakana fit into the picture.

Since World War Two, all Japanese have learned rōmaji regularly in elementary school. Although it can be used to write the language, in practice it’s just usually used for signs and computer input. It’s also used on business cards that need an international slant, and company names and brand names, both domestic and foreign, are often in rōmaji.

But in practice, rōmaji only supplement the kanji-hiragana system in occasional use with things like company and brand names for international use, like Toyota and Yamaha. Rōmaji will also appear within a kanji-hiragana text to express acronyms like NATO, CD, DVD, or units like m (meter) km (kilometer) l (liter). Arabic numerals are also used.

That just leaves us with the other syllabary, katakana. Note this use in English of italics for a foreign phrase: “It was a cause célèbre.” I think the best way to describe how katakana supplement the system is to say it’s the equivalent of italics. It’s used for foreign words, foreign names, loan words that aren’t fully assimilated, company names domestically (Toyota, Yamaha). It’s used to emphasize words. It’s used for technical and scientific words, such as names of plants, animals, and minerals. It’s used for new Japanese words, such as karaoke and pachinko. Chinese loan words entered early on, and are, logically, written in kanji, but there are words in katakana from Portuguese, Dutch, German (medicine and mountaineering), French, Italian (music and food), and English.

[A digression: English is the main provider of most modern loan words, all written in katakana. The words, not surprisingly, are modified to fit Japanese. Curtain appears as kaaten (note final N), elevator is elebeetaa, girl is gaaru. (Note that L is replaced by R, since Japanese has no L—the same thing happened in Hawaiian; also, as in Hawaiian, sounds are made to fit the CV structure, such as in the last syllable, -RU.)

English words are abbreviated in a way not done in English: supermarket is suupaa; department store is depaato; word processor is waapuro. Meanings can vary—mansion appears as manshon, but means condominium.

Most unique are the exclusive Japanese creations, such as a “salaryman” for a company worker and “walkman” for a type of portable radio, but most surprising, “walkman” then actually entered English as a regular word!]

To describe the whole Japanese writing system as it is used on a daily basis, I’d like to put it in this formula: kanji-hiragana (+katagana; + rōmaji). But under special circumstances (children and foreigners learning) it can all be done entirely in hiragana or (for computer entry) rōmaji. Kanji by themselves are grossly insufficient to express Japanese, and the katagana syllabary is not used by itself in isolation from other writing.

Finally, we’ll present a demonstration of the combined system in daily operation. Here, taken from Wikipedia, is a headline from the Asahi Shimbun from 2004. Four-way color coding will ease identification: kanji = red, hiragana = blue, katakana = green, rōmaji and Arabic numerals = black:

Although it also could theoretically be rendered entirely in hiragana, this is the same headline transliterated entirely into rōmaji:

And finally, the same headline translated into English:

Let’s now see what we have. The kanji, here in red, are visibly the most complex symbols—remember the few we’ve learned are “the easiest of the complex”. They cover basics, such as the nouns, but then only the stems of the adjectives and verbs, being inadequate to do more, as Japanese, (but not Chinese) requires. I am just curious about the kanji that looks like one mountain on top of another, but know nothing about it.

Here appearing in blue, the hiragana in a mixed text are primarily the maidservants of the kanji and provide the adjective and verb endings. You can see them here following the kanji, where needed. They are much simpler in style and rather cursive.

In green here are the katakana, here first representing a foreign name, then the loan word “marathon”. They look less cursive, easier to identify, and less “uptight” than the hiragana.

The only rōmaji in the mixed text is the international symbol for meter, but you can get a better look at rōmaji in the full transliteration. Arabic numerals are also mixed in, but a comment is necessary. We are used to the next higher step after “thousand” as being “million”, and consider ten thousand as simply a multiple of thousand. But I remember reading that in Japanese, ten-thousand is considered a separate unit. Therefore 1 of this unit, as shown above, is 10,000, while 2 of this unit is 20,000, and so on.

When we reviewed Hawaiian, we noted how “Merry Christmas” ended up appearing as “Meli Kalikimaka” because of the simplicity of the CV syllable structure of Hawaiian (2008/22), as well as R appearing as L since Hawaiian didn’t have an R. It would behoove us to look at how the foreignisms are adapted to Japanese, and we’ll find very much the same thing.

Many languages allow clusters of consonants. Just imagine a Hawaiian or Japanese confronted with the difficulty of pronouncing an English word like “strips”, with five consonants and only one vowel! Where are all the other vowels? That word could theoretically have vowels added to be pronounced su.tu.ri.pu.su. That’s just theoretical, and an extreme example, but let’s review how Japanese wraps its syllabic structure around the foreign words above.

The word meter has three consonants, but only two vowels, and ends up as mē.to.ru, with one vowel added to complete the CV(n) pattern. Once Japanese adjusts the non-Japanese TH to S, Marathon fits Japanese quite perfectly as ma.ra.son. Even the N fits in as being an optional possibility at the end of that syllable.

But let’s review more carefully what happens to poor Radcliffe’s name. First of all, the word is normally spelled rather flamboyantly, so let’s strip it down to its essentials, rad.klif. Now we realize that Japanese, not having an L, will substitute R, so we have rad.krif. The rest is easy. To fit the pattern, vowels are needed, so rad becomes ra.do (the choice of which vowel seems arbitrary). Now krif is highly consonant-rich, but vowel-poor, and is fleshed out as ku.ri.fu, so our friend Radcliffe ends up as Radokurifu. Say it fast, and it will indeed sound very Japanese, and with luck, even a bit recognizable.