For Christmas, Iris got a Lego Egyptian pyramid, complete with a spring-loaded sarcophagus that attacks grave-robbing archaeologists. When you build a Lego model of something, you have to start with the plastic bricks you have available, and the result will be recognizable but with an obvious Lego personality.

In other words, when you look at a Lego model, you say, “Wow, Legos!” When you look at a Greek sculpture, you don’t say, “Wow, marble!”

This is similar to the way the katakana writing system works in Japanese: it takes foreign words and uses Japanese building blocks to synthesize something that is obviously borrowed and obviously Japanese at the same time.

First, a quick recap of the two Japanese writing systems we’ve met so far.

Romaji means Japanese words written with the Roman alphabet. It is the one your Japanese teacher will try to get you to stop using, because it’s brain-rotting training wheels.

Hiragana is the syllable-based character set used to write certain simple words. It is also used by children and novice students of Japanese (hello!) to write everything. When you succeed in learning hiragana, which doesn’t take more than a week or two, you’ll feel like you’ve accomplished something. You have accomplished nothing. (How is this drill sergeant zen master routine working for me?)

Katakana is hiragana’s evil twin. The name reminds me of the classic 80s computer game Karateka, in which a black-belt karate master tries to learn the four Japanese writing systems and gets so mad he just goes around kicking people. Like hiragana, katakana has 46 characters, and each character represents a legal syllable in spoken Japanese. But katakana is used to write different stuff. No, really. It’s as if we had one English alphabet for writing some words and another, Dr. Seussian, alphabet for writing other words.

Katakana’s main raison d’être is writing foreign words and loanwords. (I don’t know if “raison d’être” is something they say in Japan, but if it is, it would be written in katakana.) When I took French, we learned that France has a snooty (even by French standards) government department charged with maintaining the purity of the language and defending it from foreign interlopers like le meeting and le weekend. This effort has mostly failed, of course, but there have been occasional successes, like “l’ordinateur” for “computer,” instead of, perhaps, “le computeur.”

Japan hasn’t even tried. Japanese is so full of English words, it reads like a bicultural ransom note. Those English words, along with some assorted French, German, Portuguese, and recently borrowed Chinese words, are rendered in katakana. For example, here’s how you say “computer” in Japanese:

コンピュータ

which is pronounced “konpyuuta.”

But katakana’s responsibilities don’t end there. My katakana workbook, Let’s Learn Katakana, explains that it’s also used for onomatopoetic words (Japanese has many of these), sometimes for the names of plants and animals, and most important: for writing domestic telegrams. (Let’s Learn Katakana was published in 1986.) Also, sometimes people write words in katakana just to look cool. You know, like how Americans use Comic Sans.

The good news is that katakana is very easy to spot. You will practically never look at a word and wonder whether you’re reading hiragana or katakana, because one is sinuous and the other is jagged. Here, side by side, are the hiragana and katakana for “ma”:

ま マ

If it’s cuddly, it’s hiragana. If it looks like ninja weaponry, it’s katakana.

Now, back to those English loanwords. It’s hard to fathom just how many thousands of English words have made their way into Japanese, from “aisu kurÄ«mu” (ice cream) to “zero.” A huge section of Let’s Learn Katakana is devoted to recognizing and writing these words, and I didn’t understand why until I started paying closer attention to Japanese packaging and labels. (Incidentally, Japanese packaging and labels are the world’s best; the Japanese invented frustration-free packaging back when Amazon was just a river.)

The other day, for example, I was doing some laundry and noticed a shirt with a label that said エディーバウアー. My brain clicked over into Japanese mode. (Is it a sign of progress when you can feel your brain switch gears like an automatic transmission?) Written in romaji, that would read: “ediibauaa.” Eddie Bauer!

Meanwhile, Iris peered at another label and saw ナイロン. That’s “nairon,” i.e., nylon.

Katakana and Japanese loanwords make me a little queasy, because if you’re an English speaker committed to speaking Japanese well, you’ll spend a considerable amount of time essentially speaking English with a stereotypical accent. I asked my teacher, Toshiko Smith of the Seattle Japanese Language School, whether I could tell people in Japan that I’m from “Seattle,” instead of having to say シアトル (Shiatoru). Sorry, said sensei: if you want to be understood, use the Japanese pronunciation. Indeed, in Japan, Ms. Smith would be known as スミスさん (Sumisu-san).

Because, you see, Western names are also written in katakana. Here’s mine:

マシュー

There’s no “th” sound in Japanese, so it’s pronounced “MashÅ«.” I’m not sure how I feel about this. Is it too late for me to choose an alias like ターミネーターさん？

In hiragana, the simplest of the Japanese writing systems, most characters represent a consonant-vowel combination—a syllable. In English, you can smash letters together in unlikely combinations like “strengths” (nine letters, one vowel!) and it’s perfectly legit. In Japanese, however, certain simple sounds are just plain illegal. For example, “wa” is a common syllable in Japanese. “Wu,” however, doesn’t exist. I wonder what people scream at Japanese pep rallies?

Because Japanese is built from these syllables rather than letters, most Japanese words go consonant-vowel-consonant-vowel, like when they made you sit girl-boy-girl-boy in elementary school. (Seriously, everything about learning Japanese makes me think about elementary school. When is nap time, again?)

This also explains why there are 46 hiragana characters, even though the range of allowable sounds in Japanese is much more limited than in English. There’s no character for “k”; instead, there are five characters for ka, ki, ku, ke, and ko:

かきくけこ

Learning a character set involves at least three different skills: remembering the sound (and perhaps also the meaning, if the character represents a whole word) that goes with the character; remembering the character that goes with the sound; and producing the character by hand with a pen (or calligraphy brush).

Because I’m over thirty years out from learning the three R’s, however, I’d forgotten than writing involves any skill other than mechanical reproduction. (Hmm, when I put it that way, it sounds like robot sex.) It turns out that, for me at least, learning how to read the character is pretty easy; learning how to write the character is almost as easy; and making my brain cough up the character that goes with a particular sound is stupefyingly hard.

This is what Dan Schmidt was getting at in a comment on my previous post:

One possible issue with Byrne is that from looking at a few sample pages it looks like it is more about recognizing hiragana than being able to write it. But as a tourist that’s probably all you need anyway.

Byrne is the author of the hiragana mnemonics book that helped me quickly learn to read all of the hiragana. Exactly as Dan warned, however, I still struggle to go from sound to character—that is, I can see the character ほ and know immediately that it’s pronounced “ho,” but if I need to write a word with “ho” in it, I often get stuck remembering which character makes that sound.

This is not a big deal for hiragana or katakana, which have a few dozen characters each; sooner or later I’ll remember them all. It’s a serious problem for kanji, the Chinese characters used to write the bulk of all Japanese words. Kanji are more complex than hiragana or katakana, and there are a hell of a lot more of them.

James Heisig, who we will meet again soon, warns about this problem in his book Remembering the Kanji. The very title of the book hints at it: there are over 2,000 kanji to remember. It is one thing to recognize a kanji and know what it means or how to pronounce it. I can recognize about a dozen kanji at this point, the most complicated of which is the character for “eat,” which looks like this:

食

It’s quite another thing to be able to write that character from memory, or even to remember what the character looks like when all you have is the concept “to eat.” It’s like the difference between recognizing that you’re eating lasagna, and making lasagna from scratch without a recipe. Here’s how Heisig puts it:

If you try to shortcut the process by merely learning to recognize the characters for their meaning without worrying about their writing, you will find that you have missed one bird with two stones, when you could have bagged two with one. Let me repeat: study only from key word to kanji; the reverse will take care of itself.

He’s onto me! Oh, well, if I don’t make every single beginner mistake, what will I have to write about?

Next up: The secret connection between katakana and Legos.

Japanese has four writing systems.

Let me say that again: Japanese has four writing systems. If you want to read and write it fluently, you have to learn four writing systems. This is like being told that if you want to pass the driving test, you will have to build a car from scratch, and that car will have to pass California emissions standards.

Luckily, of the four systems, you get one for free if you’re literate in English. This one is called romaji, and it’s used to write Japanese names and other words using the characters of the Roman alphabet. When you see “konnichiwa,” that’s romaji. It’s blessedly common to see place names written in romaji on signs in Japan, but it’s certainly the least used of the four systems.

When you were a kid, did you ever create a secret code where you replaced letters with simple shapes and could thereby pass notes to your friends without, well, I’m not sure who we thought was going to spy on our mail. Parents? Cops? Rival ninjas? In your home-cooked code, maybe A turned into a circle and B a bunch of wavy lines.

That cipher you created is something like hiragana. Hiragana, for reasons I’ll get to soon, isn’t literally an alphabet, but it is literally and figuratively loopy.

Let’s meet hiragana. Here’s an actual character:

ぬ

That one is pronounced “nu,” and it’s one of my favorites. Hiragana books tend to teach mnemonics for memorizing the characters. The one I used, Hiragana Mnemonics by Bob Byrne (actually, I used the [app](http://itunes.apple.com/us/app/dr.-mokus-hiragana-mnemonics/id387585135?mt=8)), says “nu” looks like a pair of chopsticks trying to capture an errant NOOdle. “Nu” for “noodle.” Not bad. This is one of the easier ones to remember.

As with any mnemonics, the dumber they are, the better they work. I initially dismissed Byrne’s mnemonics and said to myself, “I don’t need a crutch to learn a few characters,” and then found that, gosh, it’s hard to forget that the “mu” character (ã‚€) looks like a farting cow.

During the two weeks or so that it took me to learn the 46 characters of hiragana, I found myself thinking over and over that it’s funny how our English alphabet is so logical and straightforward, and this Japanese one is a hodgepodge of squiggly goofs. More specifically, I reasoned, our letters look like the sounds they make. I mean, an M looks like it should make the “M” sound, right? Whereas in hiragana, this thing:

ま

is supposed to sound like “ma.” Ridiculous!

Intellectually, I know that this is absurd. Our glyphs are just as arbitrary as theirs, although when I take a break from Japanese practice, shake off the wiggly weirdness of hiragana, and look back at the Roman alphabet, I start to notice some oddities, like the fact that you form your mouth into an O to speak the letter O.

My friend Neil has been to Japan several times, so I bragged to him that I’d learned all my hiragana. He’d learned the characters, too. “But I never got to the point where it looked like words,” he added.

Way to spoil my moment, dude. Fluent readers, in English or any other language, don’t read words letter by letter. We don’t sip. We gulp down whole words and sometimes whole phrases as single units. But hiragana doesn’t look like words to me yet. It still looks like code.

A couple of years ago, when Iris was in kindergarten, I spent an afternoon every week volunteering in her classroom. Seattle public schools use a “writer’s workshop” curriculum, where students in every grade are expected to write every day. This includes kindergarteners, who have a wide range of writing abilities, ranging from “can write” to “wouldn’t recognize the letter P if it peed on them.” I spent a lot of time helping kids try to recall which letter makes, say, the “K” sound. I did my best to empathize with the difficulty of the task, but remembering back to a time when I didn’t know the alphabet was simply beyond me.

Well, that nightmare about being back in school has turned real. Recently I woke up at 4am and couldn’t get back to sleep until I remembered the hiragana that makes the sound “ho.” (It’s ほ.) When I read hiragana, I sound exactly like a tentative five-year-old sounding out a word. (Incidentally, I frequently have the nightmare where I forgot to go to class all year and now have to take the final exam. I do not, however, sit bolt upright in bed, dripping with sweat, when I wake up from it.)

Even if I get to the point where hiragana looks like words, I have another problem: most words in Japanese aren’t written in hiragana. Four writing systems!

I realize I’ve said little about how hiragana actually works, and how and when you use it to construct words; that oversight will soon be remedied at length.