Japanese, That Cryptic Beauty

[japanese-cryptic-beauty]

The sunken cost fallacy of learning Japanese

In the intro to the book "Remembering the Kanji", the author argues convincingly that for non-Japanese the approach of learning Japanese (or rather: kanji) like Japanese school children do is flawed. And one of the ideas he brings forth is that such learning approaches are colored by the bias of "that's how I did it, after all" exposed by speakers of Japanese.

Well, same to the Japanese learning community as found on the internet!

Try to look on the internet for approaches of learning to speak Japanese only and learn Japanese primarily by speaking, and you will find a chorus of braying naysayers that tell you that this is impossible.

You know what this reminds me of?

Of the endlessly repeated arguments why Japan can't let go of the kanji themselves. Artificially constructed, biased arguments.

What's the typical argument for the kanji? That they serve to disambiguate the many homonyms in the language and are therefore essential to writing Japanese.

They are not to speaking Japanese.

You will find examples brought forward in favor of the hypothesis that Japanese cannot be disambiguated if written merely in kana (Hiragana and Katakana), and these are usually highly artificial. They are probably completely unintelligible to the native Japanese speaker, too, if spoken out loud without prior explanation. (Spoken Japanese has additional clues to the meaning of a word through the pitch melody, though, disambiguating some homonyms.)

As regards speaking Japanese, not a single Japanese person has learned to speak Japanese in school. You learn it the way almost all people in the world learn to speak - by imitating their parents and surroundings. But the same is true for reading and writing. Not a single Japanese speaker had to pick up the basic vocabulary and grammar of the language that way. You learn by imitation.

I'm not trying to argue that imitation is somehow a perfect way to learn a language. But I would say that learning with Romaji or Kana alone would probably give you more progress to the actual spoken Japanese language if you simultaneously learn grammar and vocabulary than focusing on Japan's overly complicated, dated, and inefficient writing system.

It won't get you reading manga anytime soon if that's your goal. Or many street signs, for that matter. It all depends on your goals. But if your goal is to parse the language, follow along with anime, or just hold a conversation, I'd say we have been barking up the wrong tree a long time.

When I understand a spoken Japanese sentence, I notice two things about it:

I happened to know most or all of the vocabulary.

I have some awareness of the grammar involved.

The second part is necessary because as foreign speakers we start with "decoding a language" instead of "imitating a language." Both processes take place until a certain familiarity is reached, then the actual constructs start becoming second nature.

By the way, in China you have a similar problem with the Hanzi and the fact that young Chinese use Pinyin to quick-type their language, just as Japanese use Hiragana. Both are ways to sound out words and be given the written form in an automated way, often with prediction of which word might make most sense (or is most common, at least). This has actual impact on literacy and how many working kanji/hanzi you retain for independent use.

It also highlights that knowing the language first is the actual key, not the kanji. It is just that the internet has been bamboozled by a set of learning institutions whose primary mode of teaching is through writing which is convenient to the medium itself.

[japanese-cryptic-beauty]

The nightmare of Visual Novels

So, I found this one website recommending Visual Novels (VN) for reading immersion. (Not to be confused with "graphic novels" which is, at its heart, a fancy term for comic book...) These are basically computer games where you follow a sort of branching story. A very typical one is the "dating sim" where you interact with a variety of love interests and try to, well, "romance" them - by choosing actions they may like, playing the game multiple times to achieve different endings, etc.

Funnily enough I know the concept mostly from the world of anime and manga, one of these meta things where there are now stories being written about VN - which themselves are stories and a way of story-telling. For example, "The World Only God Knows", where a dating sim addict is called upon to romance "actual" girls to drive hellish spirits out of them. Or the hilarious "Romantic Killer", where a gaming addict is forced into living a dating sim scenario.

But this is not about dating sims or the contents of particular VNs, this is about learning Japanese. I came across a website recommending learning Japanese through VNs, and the arguments were convincing:

You get the dialogue displayed on-screen which allows for reading it carefully.

You typically also get a voice-over, so you get also additional listening immersion, something harder to do with, let's say, anime. (Because VNs pause after a certain point and give you time until you continue. Also, depending on streaming site, you don't get Japanese subtitles / CC captions. Like on major anime streaming site Crunchyroll which typically has none of that.)

You can enlist certain programs for capturing text from the most common VN engines to actually use a dictionary with it.

So, sounds great, right?

The Shit You Can't Buy These Days

Welcome to the world of modern capitalism, where the prevention of spending money advances faster than the ability to actually buy stuff.

Huh?

This starts when you try buying these things. The geniuses that try to sell us things have decided that it's not good if we can simply buy stuff. No, this content is surely for that person in that market and can not be had anywhere else. Fans of manga may already be aware of this - some English translations of manga can only be bought in English-speaking countries because obviously those could be the only people ever interested in them...

So, if you live outside of Japan, you obviously have no interest, and hence get no way, to buy Japanese language VN. You may be able to buy an overpriced translated version on some console somewhere, but we won't let you lay your fingers on our precious software, no.

Hurdle #1: Money, money, money

So, there's a site called "DL Site" where you can legally buy Japanese VNs and, surprise, download them. This site also sells for example manga.

One problem: Payment providers, at least the usual Western gang thereof, doesn't like this website. For some reason, payment providers or major credit card companies nowadays like to stay as far away as possible from anything that could be classified as "adult entertainment" - and frankly, many VNs have that vibe. Not all of them. But these payment providers either limit themselves to a squeaky clean portion of the store (manga only, though) or are not available at all.

Mind you, I wasn't hunting for smut. I merely was trying to buy a VN.

I then went on an odyssey for finding a way to pay. Most options were Japan only - either obviously so or more subtly. Credit cards only work if issued in Japan. At some point I got myself a LINE (the Japanese chat app) account to use LINE Pay - only to find out that I cannot transfer funds into it for no apparent reason.

I was about to give up when I found the solution for the payment problem, which is as obscure as it gets. The payment providers don't want to be associated with the content, but there's a companion site to DL Site where you can buy coupons. These coupons can be redeemed into points that are equivalent to Yen and which you can pay with on DL Site, circumventing the absence of Western payment providers. (It actually works just fine, but it does look a bit fishy at first...)

Problem #2: Nah-nah-naaaah-nah

By which I mean: You didn't think it was so easy, right?

Now that you legally own something you paid for and downloaded it unto your Windows computer, can you play it?

No!

For you see, the publishers of these games have quite some opinions as to who can play their games. "Japan only", that's what. Seriously.

What used to be the dreaded copy protection in decades past is now some silly gating of content behind dubious checks that are made to ensure that the Japanese content can only be enjoyed in Japan.

I don't even pretend to understand why!

Seriously, the software in questions may check:

Your timezone

Your language settings

Your input settings

Your locale (date, time format, etc)

And only if these are sufficiently "Japanese" it may permit you to actually play the game you own (after all, you entered your serial).

The lengths to which a game company goes just to prevent you from playing their game, not for pirating reasons, but simply for ensuring that you are Japanese, is absurd.

At first I tried to run the game inside a virtual environment, "Virtual Box", but no matter how hard I tried, it wouldn't start there. It was quite a demanding little game... for something that plays some music and shows mostly static images and text.

But... do you really want to change all your content, input, language, and time settings to Japanese to start a game? Do you?

There's a way to simplify this a bit. You can have a second account on your Windows computer with all these settings in place, I called mine simply "Japanese". But frankly, that's also shit. Because to use a separate account you have to sign out of your main account, which closes all your apps, browser windows, etc. (Yes, Windows isn't exactly a true multi-user system in its consumer version.)

So this is a big hassle to just play a game for half an hour. (After all, you're brushing up your Japanese, you aren't fluent, so you'll probably tire of it at some point.)

Conclusion

Trying to lay your hands on a Japanese VN as a gaikokujin (foreigner) is a major hassle. Frankly, after managing to make the game run, I didn't actually play it because it had been so tiresome to actually make it work and there were major hoops to jump through just to get it to start.

Now, if you have a second computer, say, a laptop, this can all be navigated, but I spent hours on two separate days just to get to the point where the game was bought, downloaded, installed, and worked. The requirement to mangle some fundamental settings on my computer to just start many of these games doesn't sit well with me.

I never even installed the support for dictionaries because I felt I was done with it.

[japanese-cryptic-beauty]

Phonetic composition for Wanikani

Remember how I mentioned the book "The Kanji Code" in my lengthy article about on'yomi ("Chinese") readings of kanji?

Turns out the benefits of the book can be had for Wanikani users through the magic of "userscript."

What's the benefit

So, phonetic composition is how a great many of the kanji you're (supposedly) trying to learn are made. These particular kanji are called 形声 or "keisei" - "form-voice" kanji.

For the "form," they contain a part (often referred to as "radical") that gives you the meaning/semantic category of the kanji, like what the concept is mentally filed under.

For the "voice," they contain another part that hints you at how they are pronounced. It is commonly believed that kanji are only pictorials linked to concepts, but for keisei kanji this is not the case. A big part may not pertain to the meaning at all. Spotting the "voice" part can help you recall the on'yomi sounds of kanji much easier and systematically.

Yes, this is specifically for the "Chinese" readings. The Japanese readings are words from a language that had nothing to do with the formation of the original hanzi script, after all.

(Examples are in the original article.)

How it works

Websites are rendered by your browser, but how your browser does this can be extended and changed by scripts that rely on the Javascript language. Nowadays browsers like Chrome or Firefox allow you to install "extensions" that do various jobs for you - blocking ads, returning the ability to copy/paste to pages that block it, etc.

Wanikani, the kanji-learning website, incorporated early on support for so-called userscripts, both a blessing and a curse. Curse because whenever they want to change something on the website, a very vocal userscript culture will complain about breaking their stuff. Blessing because it allows you to extend what WK can do.

How it's done

You can basically follow the guide here, but I must admit I failed to do so because I think something's missing.

What you do is this:

Install Tampermonkey (guide) - this is the basic engine.

Install WK Open Framework (guide) - this is Wanikani's own script support.

Install Keisei support (guide)

Now, if this worked, you should see something like this on various pages - for kanji readings and radicals:

This will tell you if a phonetic component is present in your kanji or if it is itself present in another kanji as a phonetic component. The readings printed in bold are readings still used for that kanji, others listed are merely present in phonetic components but not in the original kanji. Because history.

It also shows you cases where the link is weakened because the pronunciation changed - in the example you see this marked in red. Basically there is a grading system (so far I've seen "heaven", "above", "middle", and "below").

A Caveat

There's only one caveat. Tampermonkey is very powerful. You need to trust the scripts you install, because it could potentially alter what you see on websites or enter/submit stuff for you.

On Chrome there is a good solution for this. When you click on the "Extensions" icon in the top right, you can select "This can read and change site data" from the three dots behind Tampermonkey.

I recommend doing it like this:

Select "On all Sites" during the installing of the WK Open Framework and the "Keisei" support script. (Else the install links have a habit of not working on Chrome.)

Then change it back to "When you click the extension".

Now go back to a tab with Wanikani open. Go back to "Extensions -> This can read and change site data" and select "On wanikani.com".

Congrats! Now Tampermonkey is only enabled on Wanikani by default.

Alternatively, use it in a separate browser. I haven't found an easy way to restrict permissions on Firefox, for example. But if you use it in a separate browser that you only use for Wanikani, you should be good.

Kanpai to the people who made this possible!!

[japanese-cryptic-beauty]

Does it fit you?

There's a lot of advice for learning Japanese out there, and sometimes it just pisses me off.

Now, it's always good to look at why something irritates you in case there's a grain of truth that you're not seeing... or wanting to see. I mean, some things are unavoidable - like there being quite some work needed to learn Japanese. Especially if you want to see some results. But...

Maybe an example will help. IIRC, the YouTuber was called Mochi. I just browsed through her tips what is needed to learn Japanese, but if you want me to distill it for you, it was essentially a brute force approach, not very different from how Japanese students learn it. Using her app for Space Repetition (SRS) was basically the only nod to anything that has happened in the "learning Japanese for Westerners" space at all. I bowed out at "learn to write kanji" as a needed step.

Now, undoubtedly, writing kanji a few times can give some insight into kanji themselves, to learn concepts like stroke order. It isn't especially hard, either. And if you have problems telling certain kanji apart, this might help fix them into your brain. But in general, from my own experience, I disagree.

You see, I started out by learning to write 1,000 kanji. But now I have largely forgotten all the connections between meaning and kanji built through this exercise already. Maybe the effort to learn kanji that I've seen before through this method is lessened, but I still learn them all over with Wanikani.

I'm not claiming she said that only learning to write kanji was enough. That would be misrepresenting what she said. But if I were to paraphrase it another way, she basically said "Learn it how I did. Trust me, I'm Japanese."

Don't trust the Japanese

I've been talking about SRS a lot, and so you might think I'm a big proponent for Spaced Repetition. To that I would say... it depends.

You see, SRS by itself is only a slightly better way to do rote learning so that you are more likely to retain it in the long-term. So if somebody tries to sell you their approach by saying they use SRS (and nothing else), all they offer you is to schedule your rote learning for you. ("Anki", as I recently learned, the name of the most popular free SRS software, may simply to mean "memorize" - because that's the Japanese kanji/word for it.)

If you want to rote learn Japanese, be my guest.

I mean, it's a valid approach. It's not the fastest or most intuitive, but it's a big part of how Japanese learn the most vexing parts of, well, Japanese, aka "reading and writing Japanese."

But the general majority of Japanese dispensing advice on learning their language on the internet seem frankly uninspired. If you're looking for strategies matched to your own way of retaining information, it's quite hit and miss. (Wanikani has a Japanese founder, though.)

I have written kanji, I have trained them over two different types of SRS. But at times they don't come to mind directly, and then I'm glad if I can recall the mnemonic story I've associated with them. Because I might not recall the meaning of the kanji directly, or its readings, but maybe any of the other ways I stored them will come to mind. But that's not the effect of SRS (or SRS alone), but of an alternate learning strategy, not simply relying rote memorization.

What is your goal? What are your priorities?

But Mochi is not the only one. NihongoDekita is someone I like to watch YouTube shorts of. She wins the prize for making the lamest argument why you should learn to read Japanese (= kanji + hiragana + katakana) natively. She puts an English sentence in Japanese letters and reads out loud and says "that's not English, that's Japanese" as her argument why Romaji is not good enough for learning Japanese.

But... while Japanese is incapable to produce the sounds English requires, you can read Romaji and produce exactly the sounds Japanese requires. (You would make yourself aware of how to pronounce it properly, but the same is true for learning Hiragana as a Westerner - or any language in the world transcribed to Roman characters.) Talk about a badly made argument.

Your learning of Japanese would certainly be incomplete if you didn't learn how to read it. But then your learning of Japanese is also incomplete if you don't learn how to write it. Or if you don't learn Kansai dialect. Or if you don't learn conversational variations how contemporary Japanese change Japanese from "proper" to day-to-day speech. Etc.

But there's one question to cut through all of this: What do you want to achieve?

Do you want to learn to read manga originals? Learning to read is not optional for that. Do you want to learn basics of face-to-face conversation? Romaji is a viable way, then. Same for understanding what characters in anime say (if you are willing to ignore all the displayed kanji). None of these require being able to print kanji. And so it goes.

Learning kanji would be a bad investment of time if your primary goal or priority is to learn to converse fast. It will majorly slow you down. Ironically, it will slow you even more down if you learn less! Learning 100-200 of the most common kanji is harder because some of them are more complex ones and since you wouldn't know the radicals they are composed from from other kanji, you have less to anchor them in your brain with.

Wanikani has been criticized for not preparing you for the JLPT N5 and N4 tests, but its goal is teaching kanji in a good way for building a bottom-up understanding, not get you to the goal of the most common the fastest. You go from building blocks to bigger structures, and you build from what you know. Which helps retention in a different way. Simply using SRS would not have the same effect for retaining the kanji because they would look still alien to you. Frankly, they would, even if you learn to print them without a structured approach.

Return On Investment

The reality is, your learning time is likely limited. So might be your patience before you see results that reward you for your progress.

Yes, if you learn kanji, learning to print them can anchor them in your brain better. But if you want to learn to print them from scratch, there's an additional investment of time. In the Heisig method, you would arrange your mnemonic story so that you can actually give yourself hints what to print in what order. That's a more complex requirement for a mnemonic story than merely coming up with a story that helps you spotting the elements in any order to remember what kanji this is. Kanji elements can become quite crowded, and can appear left/right, above, under, etc. Then you need to retain some memory as to how to arrange the elements in the given space so that you can actually write the character properly.

The same time might be used to learn more kanji and vocab, or to practice grammar, tackle some graded readers... And it requires you have time to sit down and do it at a desk or table, pretty much. Commute times, in comparison, might be usable only for an app or index/flash cards, or reading.

Each learning strategy needs to evaluated for how much time it uses, how effective it is, and what kind of structure it builds for the future.

So, for example. If your goal is to take the JLPT N5 test ASAP because you need it professionally, Wanikani is not your best choice. Yes, by now I've covered practically all of them. I referenced two separate lists, and they basically say I have 99% of them. But they did not come in that order. And when I go over Bunpro's N5 vocab deck, there are lots of kanji in there that are in neither list.

I, myself, find that okay, since my goal is a complete learning of the Japanese language, not a test, but if you really have to, you might rightfully get frustrated by such "you will eventually get there" approaches. (Heisig is way worse, after all.)

There's no one true way to learn Japanese.

There are always compromises you make. And if you're taking classes, the compromises your teachers make. Reevaluate them over time and compare them to your goals. You can approximate what works best for you as you learn about the learning of Japanese itself.

Over time, unless you do courses in a fixed system, you will have to structure your Japanese learning yourself. And even then, if you apply your Japanese in some way in the real world, you need to come up with strategies to come up with the missing parts that class didn't cover yet. So being aware of your goals and priorities and comparing your current approach with them once in a while can give you a valuable sense of direction.

[japanese-cryptic-beauty]

On'yomi, or it's all Greek to me

(I'll describe something really useful for memorizing the on'yomi readings of many kanji at the end of the article, so you might also want to skip ahead. It's the part with all the kanji in it, can't miss it, really.)

When you learn Japanese, you learn many facts that sometimes make you wish you didn't. You happily beaver away at learning to write kanji only to pick up on the fact, that Japanese has two predominant ways to read almost any kanji: on'yomi and kun'yomi.

You could on'yomi "the Chinese readings" and kun'yomi "the Japanese readings" and you would be both right and wrong. So very wrong.

Well, not entirely. The notion is actually useful. On'yomi are readings derived from the evolving Chinese language of different time periods but pronounced wrong. (Remember, Japanese doesn't emulate Chinese pitch and has a limited set of sounds.) And the kun'yomi preserve various readings from various parts of Japan that were the sounds of the original Japanese language predating the Chinese writing system. (Get the deets here.)

Wait. If you have native Japanese sounds, why do that? Who on Earth would do such a thing!?

Right. It's us! (I'm German.) German uses words derived from Latin and Greek a lot in spite of perfectly usable German words meaning the same or similar things. So I'm the one to talk, huh?

"die Intention" - the intention. Perfectly German word? "das Vorhaben," or better, "die Absicht."

"die Prätention" - basically to give yourself airs. Perfectly German word? "die Anmaßung."

"die Restoration" - the restoration. Perfectly German word? "die Wiederherstellung."

Latin was the language of the Roman empire, but this impacted more the cultures abutting the Mediterranean and their "Romanic" languages - Spanish, Italian (of course), and French, for example. The basic stock of the language took in a lot of Latin words, they are at the core.

Germans on the other hand, thanks to "Herman the German", stayed mostly independent from the Roman empire and played part in eventually bringing it down. These proto-Germans had their own language to start with, and probably some form of sour kraut. So why all the Latin and Greek?

Latin is the language of many academic professions, especially law (Roman law is the historic origin of most western law) and medicine (partially with the intention you don't understand what your doctor is saying about you!). Universities in the middle ages and beyond relied on Latin as the language of exchange. Newton wrote about gravity in Latin. (And lots of other things. A busy bee, Newton was.)

These pesky academics then went on the coin all kinds of complicated words for newfangled technology that the ancient Greeks and Romans surely had envisioned at some point. Now wait. That was Jules Verne.

Television - far-away seeing. Oh wait, no, that's telescope. My bad.

Automobile - it moves by itself.

Mobile phone - movable sound.

Arithmetic - something about numbers and counting.

Algorithm... Ha! Gotcha! Sounds Greek, right? It's derived from the far more accessible Arabic...

My examples are English but exactly the same words exist in German with minor spelling differences - "Television" oder "Fernsehen", "Automobil" or simply "Auto," "Mobiltelefon" or simply "Handy" (say what?), "Arithmetik" and "Algorithmus."

So, we're a mess, huh? Germans can't speak German without resorting to Latin, Greek, English ("Computer", "Handy"), a bit of French, Italian, Yiddish... and we mangled most of it. We can't help it. Most of it we translated to our own sounds. (That will show our neighbors and compatriots what we think about their invasive words. We'll use them but make you cringe.)

Don't get me started on English. A Celtic language mostly overwritten by Anglo-Saxon Germanic invaders with Frisian partially overwritten by originally Norse French (Norman) invaders, stuffed to the brim with Latin and not even the same root word for an animal and its meat. (Eventually, veganism might take care of that one.)

Deep breath. It's fine now.

Compound words

Now that was a lengthy excursion amounting to nothing, right?

Well, get used to it.

If you use a single kanji word, you typically use the kun'yomi sound. Because there was already a Japanese word representing that concept when writing arrived. Also, this is true of most verbs, adjectives, and adverbs, and some compound words (featuring multiple kanji). The tell-tale sign is typically the presence of some hiragana when writing it.

There are adjectives that don't have hiragana, like "dislike," one of the meanest little effers in the Japanese language. Its sounding ("kirei") ends in -i, but it isn't an i-adjective. It is an adjective, but it doesn't have any hiragana attached, either. It's a na-adjective that ends in -i. There are good reasons to dislike "dislike."

But in general the presence of hiragana puts you firmly in the territory of the kun'yomi readings. So a lot of common Japanese is unsurprisingly native sounds, just as English has native words for house, car/cart/carriage, like/likable, so does Japanese ("ie", "kuruma", "suki").

The category where you however will be inundated with on'yomi readings is the one of the vast majority of compound words - many of these complex concepts were imported from China. And so was the way to sound them out. Somewhat. Kind of.

In other words, Chinese fills the same role for Japanese that Latin fills (and, to some extent Greek, also) for European languages.

A useful shorthand

So, if you have to learn two or more readings for most kanji, how do you keep track? And do kanji make any sense?

Reading "Remembering the Kanji" originally raised my hopes that kanji were logical only to shatter those very same hopes over time. But it simply approached its subject matter from a limited view. At least the first book. (But that's how you're supposed to memorize the kanji, and I do call bullshit on that.)

Only the most simply kanji are pictograms. They depict the sun and moon for day and month, or the mouth, they clearly express 1, 2, and 3, etc. So I originally thought it might be useful to get a book about their origin to remember them.

Frankly, kanji (or more correctly, hanzi) changed so much since their time as petroglyphs they've become unrecognizable. Don't bother, really.

But that doesn't mean there is no sense in kanji at all, especially when it comes to the Chinese readings.

More complex kanji were built from simpler kanji - or from alternative forms of writing these kanji. People nowadays mostly learn them through memorization, but the radicals (as the parts of kanji are called) contain clues to their meaning and to their on'yomi reading.

You could go and use this as your method for memorizing the kanji, in fact. The book "The Kanji Code" by Natalie Hamilton lays out one such method. But you can have the benefit a bit easier, I think.

So, if you have a complex kanji, one part might give you the category of its meaning, and another part the sound. You can call them "signal primitives" if you want.

Here's an example:

謝 means "apologize" and its on'yomi reading is "sha."

言 is a hint to its meaning category, it means "say."

射 is a hint to its on'yomi reading, it reads... "sha!"

Here's another example:

養 means "foster" or "bring up" and its on'yomi reading is "you."

食 is a hint to its meaning category, it means "eat" or "food."

羊 is a hint to its on'yomi reading, it also reads "you."

So the original combinations might seem nonsensical at first:

"Apology" (謝) isn't a combination of "say" (言) and "shoot" (射).

"Foster" (養) isn't a combination of "food" (食) and "sheep" (羊).

But if take both these kanji as a way to tie together a broad meaning category with a sound, we do get a good result.

In fact, look at this example:

寺 is the kanji for "Buddhist temple" and its reading is "ji."

痔 means "hemorrhoids", meaning radical is "sickness" (疒) and pronounced ... "ji." (Hey, they were probably the first culture to spend long days sitting and writing.)

時 means "time" or "hour", meaning "sun/day" (日), and pronounced (you guessed it) ... "ji." (Time is indeed tied to the sun.)

持 means "hold", meaning "hand" (扌), and, yes, "ji."

峙 means "(to) tower" (like "towering over someone"), meaning "mountain" (山), and again, "ji."

Congratulations, you just learned four additional kanji in one go by looking at their components (which you do have to know) and spotting which part is the "meaning category" and which part a clue to its on'yomi sound. (Also, none of these words do in the least relate to the concept of a Buddhist temple.)

If you're interested in learning more about this method, have a look at the book. You might find it up your alley.

But even if you don't use the book, you can always look at each kanji you learn and check if the reading is "hidden" somewhere as part of the kanji already. This reinforces your memory of the original kanji's reading and helps you make sense of complex kanji.

[japanese-cryptic-beauty]

Building the inner parser

When I was a kid, probably elementary school age, I tried one day to understand what actually the difference was between English and German. I had had some minor exposure, but for example my mother, always my primary caregiver, didn't really speak it. I tried coming up with English words, but this trial-error approach was almost entirely error. I found it hard to understand in what way English was different. At times it seemed like a badly pronounced German, but I couldn't figure out the difference.

Had I had somebody knowledgeable in my environment, I might have learned early on how many English words of Germanic root origin are in fact one shift of spelling away from German. I had kind of intuited this myself, probably with the uncanny way children pick up languages anyway from exposure.

But more interestingly in hindsight - what else had I been lacking?

How do you learn a language?

Over the course of starting out to learn Japanese I had a few intuitions, nothing groundbreaking or new. Let's just say it shifted my own, personal priorities, what I considered helpful and what I needed to do.

Naturally, through my experience in school, I gravitated towards finding a textbook that would suit me. As I mentioned before, I very quickly rejected Romaji-based books for their focus on business Japanese (I certainly don't dream of working in Japan...) and eventually came up with "Japanese from Zero" - at least it seemed to align with where I wanted to go.

I quickly became dissatisfied with its approach of throwing "Hajimemashite" style sentences at me without explaining anything about what that really means. It left me equally puzzled as the nonsensical "How do you do" exchange in English, partially because they wouldn't go into the grammar and origin of those things in particular, and partially because it sat there like an artifact, an atom, not further broken down for real understanding. It creates an illusion of understanding by throwing common constructs at you, just like a phrase book for a tourist would.

And then came my detour into kanji.

A genius move, obviously.

Exposing yourself

But even before that, I had already decided I would watch lots of anime for simply exposing myself to the language. I consider the ready availability of anime one of the biggest boons for learning Japanese. My English, especially conversant, spoken English (American style) really took off with the availability of multiple language audio on DVD. No longer did I have to buy a separate VHS of a title in English - for extra money, as they often were "imports." I could just buy the regular. And even better, they came with subtitles.

I was a huge Futurama nerd back then, and ironically I owe a big debt of learning how to speak English fluently to imitating and quoting a cartoon, albeit a smart one. Watching it over and over not only improved my English through exposure to actual speakers, it made me puzzle over idioms and cultural intricacies... and made me spot, over time, an enormous amount of translation errors in the German dub, sometimes completely reversing the meaning.

So, I watched a lot of anime this last year, always enabling Japanese audio, of course. Later on I learned this is called immersion. Just as I used my natural interest in available English material, I could watch stuff in original language that I actually found attractive - something which engages me naturally. Nothing worse than making yourself watch something just for learning, really. (To my mind.)

After all, I had noticed regularities in Japanese from watching Sword Art Online, even if they were rather minor ones and partially owed to the lack sonic variety in the Japanese language. And from exposing/immersing myself, I had some intuitive insights over time. I noticed how our inner parser works and what can help it function.

What is a parser?

While computers, deep down are only working on numbers or even bits or minor voltage differences, humans have always needed layers of abstraction to actually make good use of that. Nowadays we talk about "programming languages," but in order for the computer to understand these "languages," they need to be broken down into parts and then translated. The breaking down part is done by something called a parser.

There are many parsers. In the early days of computers "text adventures" were a common genre of games. You would be presented with text describing your situation and you had to type your actions as response. Again, analyzing what you typed fell to something called a parser, and the best text adventures were known for their advanced parsers, "understanding" language like "give banana to three-headed monkey" and pointing out things it didn't understand in a way so that you could improve your instructions accordingly.

This didn't happen in a vacuum, though. Computer science had actually looked to linguistics, or rather its most formal part, for inspiration. It took part of what linguists use to understand language itself, its basic building blocks, and derived lessons for how to make a computer understand what we want it to do.

The first thing that happens when a computer breaks down a "stream of characters" is that it tries to find usable tokens in it. This is, IIRC, called the "lexical stage." And when faced with Japanese, I noticed I would have to develop a similar thing within myself in order to break down spoken Japanese.

What helps and what doesn't

Anime Japanese will repeat certain idioms, phrases, and words endlessly. You cannot miss it. I marvel however at the endless variety that translators resort to when bringing them to English. Certainly not helpful for the learner!

Given how often they appear, certain words like "hontouni" ("really"), "ano..." ("errr..."), "des(u)" (helper verb, often used in constructs like "I am") are impossible to miss. But the rest largely remains an unbroken stream of syllables interspersed with tiny islands of understanding.

Now there are several ways to go from there, and two primarily present themselves. You can try to learn more vocabulary or you could learn more grammar to fill in these gaps.

I opted for vocab first (as part of my Wanikani approach). The main reason was is that each vocab word constitutes one token that you can pick from the stream of syllables. And Japanese aids with this. A lot of vocab is itself not modified but spoken as is in a sentence, especially nouns. Even vocab that is modified, like verbs and adjectives, is modified at the end, so you can at least recognize the stem of the word.

Grammar is, in my opinion, not as useful for this purpose initially. One reason is that a lot of Japanese grammar relies on particles, little helpers made of one or more syllables, that tell you what role a word or part of sentence plays in the whole, like who does what to whom. They're comparatively easy to spot in written Japanese, due to the dreaded kanji. Particles are part of the hiragana that appears between the concepts largely represented by kanji.

But in spoken Japanese, the same sounds like "wa" (topic), "ga" (subject), "mo" (adding another subject), "to" ("and"), "ni", etc, are really hard to spot because these sounds are so common in Japanese words as well. If you can't tell at least the individual words apart - tokenize them - then you have real problems spotting the particles that modify them at the end. And that's where vocab is useful.

Adding in some structure

While Japanese is not as structured and simple as English, understanding the structure can help at some point. English favors us by adhering to a strict "subject verb object" structure for the basic sentence design. Japanese is less forgiving, but at least the verb is at the end of a sentence that has one, and a lot of Japanese sentences are merely that - a verb. Knowing this can help spot verbs.

It's always good to know what the "doing" part is to break a sentence down, after all. To that end, I would say, knowing the conjugations of a verb is more useful initially for building our parser of spoken Japanese than particles.

Being able to spot patterns like "-anai" and "-imasen" helps us identify verbs and its meaningful stems early on, and this breaks the language apart further. Since these constructs naturally repeat over and over they become easy to spot, so I recommend to do the basic verb forms next. And there aren't that many common ones, either.

Think of it as playing a game of elimination. Each word of vocab you actually know eliminates those tokens from the stream if you recognize it. Each verb ending helps you identify one of the tokens that keeps changing in a sentence, but also helps you spotting what is being done and where the sentence likely ends. Combining these two eliminations, over time, will break down the stream of spoken Japanese considerably.

The great weakness of immersion

One big problem is that we don't get the structural information about what is being said on screen when we hear the spoken word. Subtitles in Japanese would certainly help, as we could clearly see the distinction between meaningful concepts (kanji) and grammar and helper constructs (hiragana) while following along. However, we would likely no longer understand the show we're watching, so this has to wait for later. (Also, a quick check tells me that Netflix typically offers Japanese subtitles and Crunchyroll does not. Might be wrong, but that's what it looks like when checking two shows.)

This is where we leave the media immersion and basically go to the typical "teaching grammar" approach. A good website like Bunpro provides plenty of examples that help immerse yourself in a somewhat artificial and (optionally) slowed down stream of Japanese along with the written word. Kind of like the textbook approach, but with audio. I found that Bunpro came naturally to me after all this mixture of immersion and training kanji and hiragana through Wanikani.

It comes as no surprise that a mixture of approaches is needed, but in the end it all goes towards building that parser, I feel. More vocab identifies more tokens. Specific grammar constructs help breaking down the stream further towards its parts. And yes, knowing idioms helps, too, but I still would like to be taught them properly.

Unsurprisingly, for adults Japanese can't be learned through immersion only, especially if you can't bombard a Japanese adult with questions like children do. But it can be a huge help in building that parser and a general feel for the language, its pronunciation, how the written word relates to the spoken word, and for learning the melody and rhythm of Japanese. And each construct you learn will eventually hook into that parser, until one day, you begin to understand more and more of Japanese in realtime.

In hindsight, this process I just described did for me a lot more than school for becoming fluent in spoken English. Just like reading built my understanding and vocab, nothing beats immersion ultimately for learning to first break down and then understand a language reliably when hearing it.

But now I have to get into reading. Because putting the horse before the cart does have some problems to say the least.

[japanese-cryptic-beauty]

In the Land of Electric Learning Aids Part 3: Bunpro

In the last installment I discussed the well-known kanji study website Wanikani. Have you followed that link? It's funny. It's the About Us page describing what Wanikani is. Guess what? It's out of date.

One would think a website would keep the page where it talks about itself and what it is about up to date...

Anyway, as I mentioned in that article, Wanikani does a good job of teaching you kanji and vocab, but it doesn't teach you grammar. (And it does a comparatively weak job when teaching you words that have a bit of grammar tied to them.)

This is where Bunpro comes in.

Ooh, ahhh, what is it good for?

The main focus of Bunpro is to teach you grammar, but it also teaches vocab, and, to a degree, reading comprehension.

I only found Bunpro because someone on the Wanikani forums slighted it by saying that they think SRS is unneeded for learning grammar. Yes, Bunpro teaches Japanese grammar using the Spaced Repetition method I've talked about before in the previous articles.

The default way is very similar Wanikani - you get a set of grammar items to study, and after you've moved through the set, you are quizzed on them and have to enter an answer.

The beauty of Bunpro is its default way to quiz you, though. It presents you with "mad libs."

Or in other words, a "phrasal template." You merely fill in the missing item(s). Bunpro will also provide you with clues and hints if there are any ambiguities.

Take for example the ko-so-a-do words sore, kore, are, and dore. You will be quizzed with various sentences where you have to fill in your here, there, over there, and where. Hints will tell you how close the respective thing is to the speaker - because kore is close to the speaker, sore is close to the listener, are is apart from either, and dore is to find out where something is.

Part of what's great about this approach - except for some initial frustration - is that it also incorporates more unusual examples and usages - like dore can stand for "let's see" as well.

Ordered but not rigid - Paths and Decks

Where Bunpro definitely is different from Wanikani is its high degree of possible customization. Do you prefer to use the mad libs or would you rather see answers and read them in reviews? Do you want to be quizzed (required to enter text) or do you want to self-evaluate like in Anki? Bunpro lets you choose.

But beyond that, Wanikani presents its content as an ordered dependency tree.

The items you initially unlock in WK will in turn unlock further items. New batches of items are further unlocked only if levels are passed (when you pass your initial reviews on almost all kanji of that level). There's some flexibility there, though. Let's say you don't get a particular kanji to Guru level. Then its vocab words won't unlock, but in general this does not affect the overall progression much - unless you're really stuck in a level with too many missed kanji.

Something like that is called a Path in Bunpro, and Bunpro's default way of teaching grammar is called Lessons. If you follow Bunpro's Lessons you will be taught the grammar lessons in the default order that the site staff thinks is best. But you're also free to use the order of popular textbooks like "Genki". In other words, you can use Bunpro as a supplemental tool for following that course as well.

Please also note that even paths are suggestions only. You can go in any order, nothing is locked from you. So Bunpro allows you to grammar for the final N1 test right from the beginning. Going in default order merely safeguards you from seeing stuff too early that you might not have learned yet. Furthermore you can mark things as mastered that you already learned elsewhere.

Bunpro also teaches vocab through a feature called Decks. This is a beta feature that's also available for grammar. Basically, by default a deck is an unordered bag of items you can draw stuff from to learn, or have a fixed amount drawn for you as your next review batch. What makes this super-cool is that you can order vocab according to how much it is used - in anime, or in general, or on Netflix, etc. So you learn the most useful stuff first - for your own preferred use.

Cram, cram, cram!

One of the most useful features on Bunpro I came across recently is the "Cram" function.

When you learn a new item, you are quizzed on it once, then it goes into your SRS review queue. For some items this is fine, but there are more complex grammar points that warrant more initial practice, like let's say the conjugation of u-verbs, and for that you can create a Cram session.

To do so, you select simply which items you want to practice and how many quiz questions you want on each (or all available!) and then start. So if this was the negation of u-verbs, you would now be asked to put various u-verbs into their casual-negative form or their polite-negative form, with a great variety of verbs to stimulate your brain cells.

I did that specifically for the u-verbs because they contain a little property that requires some practice - they shift sounds from one syllable to another: -mu becomes -manai and -mimasen, -tsu becomes -tanai and -chimasen. Add to this that you have to master this in hiragana, where each means a shift of syllable and an exception for verbs ending in plain -u, and you have something that (IMO) definitely warrants some cramming to drive home.

Furigana and Wanikani

In the article on Wanikani I promised that I'd demonstrate how it is a cool feature that WK exports its information as web service.

Furigana are annotations that show you how to pronounce a given (set of) kanji.

By default, Bunpro shows furigana to help you read. But you can switch them off... or go Wanikani. By registering a Wanikani API key Bunpro can safely access your WK progress and hide furigana for all the words and kanji that you already know. (You can also optionally allow them to be displayed when hovering over.)

This makes a lot of sense when combined with the way Bunpro quizzes grammar. When a kanji word is required, you enter it in hiragana (how it is sounded out) and Bunpro will insert the kanji for you when displaying the answer. (This is close to how kanji typing on a phone works.)

To make this even more cool, you can also import the WK vocab you have learned but not mastered yet into Bunpro for extra practice.

The Pros

Bunpro uses SRS for helping you retain important grammar points.

Mads libs / phrase templates allow you to study grammar in the context of actual sentences.

Example sentences are read by a real Japanese speaker, allowing for immersion and being exposed to actual pitch accent.

Lots of examples, included in the quizzes, making for a great variety of questions.

Furigana help you read kanji, but what you see can be synchronized with your Wanikani learning.

Grammar can be learned in various orders, including that of popular textbooks.

Vocab can be studied in various orders, based on how common it is in various media.

Vocab progress can also be imported from Wanikani.

Cram function allows for additional practice outside the regular SRS cycles.

Highly customizable, including how you review your material.

Unlike Wanikani, Bunpro is organized in a way that helps you study for the JLPT Japanese Language Proficiency Test by level both for grammar and vocab.

There's a section for reading which gives you example text to read for each set of grammar passed. (You can even change the reading direction.)

Like Wanikani, Bunpro uses gamification to help motivate you. As you progress, you earn XP towards gaining levels and various badges.

Each grammar item also contains a section with resources for further reading on the topic, both online as links and offline as references to pages in various books. You can mark those you did peruse as "read."

Bunpro has an undo function for situations where you think you simply made a typo. (If you end abusing it, you're only cheating yourself, though...)

Dictionary definitions and references to grammar and vocab are available as context help when you click on example sentences.

There are forums available to interact with other users.

Bunpro's dashboard is actually quite good and updates itself without reloading the page.

The Cons

Bunpro is clearly a site in active development. While I personally didn't encounter outright bugs, I did find various omissions and other issues - like missing audio and broken links.

This is also true for the Wanikani integration on vocab progress. It marked some words I didn't know as done and a few words I do know were omitted. (On average, it works really well, though. And vocab is still beta.)

There are also content issues with the vocab cards, as they may quiz you on meanings you haven't learned with the card itself - by which I mean, meanings that are not listed at the top of the card. They are in the examples, though. In order to see what I mean, please compare the vocab meanings of "suru" with its example here.

When it comes to the reading section, the amount of material per 15 grammar items wouldn't even fill a page taken together.

Bunpro is a service you pay for. It costs about half as much as Wanikani. And yes, it also offers a lifetime option to pay only once, which also goes on sale around New Year's.

As far as I see, Wanikani progress is not tracked permanently, but you have to sync it manually by pushing a button every once in a while even after registering the API key.

Clicking on words has different results in different places. Clicking on example sentences brings often context help. In Reading sections you get grammar help. In other places it only makes the furigana visible. It's hard to know what to expect when clicking a thing.

Caveats

Well, as you can see with the Cons section, Bunpro is... evolving. This is both good and bad. You can expect more out of Bunpro in the future, and I expect that stuff to be well thought out. But you will also see all these small mistakes that come with that. I personally thought the situation was acceptable, but your mileage may vary.

Conclusion

Instead of a summary, let me share an anecdote of my own experience with Bunpro so far.

So, I was looking at the N5 vocab deck and wondered where to start. I didn't really want to learn more kanji-oriented words... because if I had capacity to spare on that, I would do that to further my current Wanikani progress! So I picked the katakana foreign loan words.

Ignorance is bliss (they say), and so I found myself marking word after word thinking "This isn't so hard!" until I "suddenly" had 100 words in my vocab review queue. Add hubris to ignorance. Yes, katakana words are basically badly pronounced / spelled English words for the most part, but that means you still have to memorize how to spell each of them. (Somehow in my brain I thought this would come naturally, as it all seemed "obvious" at first glance.)

Now I had a review queue with a hundred items and when quizzed on them, I failed most of them. (I personally "blame" part of it on the difference between tsu, zu, su, and their ilk for a lot of them, because which of them is used to emulate an English sound is kind of arbitrary.) And fail I did. How to remedy this?

I wondered if I should kick most of them back out of review, but that seemed, like, well, work.

So I made it into an experiment. Would simply encountering the words over and over in SRS reviews improve the situation by itself even if the amount was huge? (After all, they are kinda familiar.)

How Bunpro quizzes you is that it does a first pass over all items in queue, and then it asks you again for the ones you failed before. Successful items will advance in their SRS level, other items won't. So the items I had the most trouble with came back almost immediately in Reviews, while items with greater rate of success would not, quickly separating the tricky ones from the rest.

Within a few review cycles I had the material internalized. And my vocab study advanced by a hundred words! Ah, the artificial joys of progress bars and percentages! My reward center doesn't care, I still was happy.

Finally, I wrote a lot of positive stuff about Bunpro, and the website does not endorse me in any way. In fact, I pay them. Well, I paid them once. Because, again, I became a big fan in a short time and I do recommend Bunpro wholeheartedly. I'm looking forward to what else the site will provide in the future.

Further installments in this series:

Remembering the Kanji / Anki

Wanikani

[japanese-cryptic-beauty]

In the Land of Electric Learning Aids Part 2: Wanikani

In the first post in this series, I laid out the RTK method. I used it to learn to write 1,000 kanji before getting fed up, basically. And a little happy accident came to pass... I looked at my Anki index cards to find that they had each additional cross-referencing information recorded on them, like which level of the Japanese Language Proficiency Test they belonged to and... a "Wanikani" level??

So I did a little research, and I found the Wanikani website. Wanikani is both very similar to "Remembering the Kanji" and very different. The reason I found myself attracted to it was that it employed a method similar to Heisig's while ignoring his approach of ignoring readings for such a long stretch.

Through Wanikani you learn...

Radicals: Again, these are building blocks for constructing more complex kanji. Wanikani's naming of them is purely tailored to Wanikani's own mnemonic stories to help you remember them.

Kanji: Well, duh... Each kanji is tied to a mnemonic story to help remember it by. There are additional hints on how to reinforce the story with various senses to help retain it better. Additionally each kanji will be taught with its most common reading. This is typically the most common "on'yomi" (Chinese-ish) reading. Again this reading is reinforced through a mnemonic, often tying into the mnemonic for the kanji itself.

Vocabulary: Wanikani presents you with a great variety of vocabulary words. Currently 6,000 vocab words are taught alongside 2,000 kanji. Vocab words are always introduced after the respective kanji have been learned.

All of these are presented through Spaced Repetition (SRS), meaning you will be repeatedly quizzed on them once you completed the initial lesson. Quizzes require you to type answers. For all three kinds mentioned above, you will be presented with the Japanese version and provide the English meaning you've learned. For most of the content you will also be presented with the Japanese version and you need to provide how it is sounded out by typing in hiragana. (You are supported in this by a system that allows you to type hiragana through a regular Roman letter keyboard.) Hiragana-only words and radicals are naturally excluded from this second quiz; the sound of hiragana is, well, obvious. And radicals taken by themselves as building blocks have no sound.

Wanikani's SRS system is fixed and not configurable. The website will quiz you on your lessons in fixed intervals, and assign each item a grade of mastery - Apprentice, Guru, Master, Enlightened, and Burned. Apprentice items show up a lot - they are constantly in your face for most of a week until you pass into the Guru stage. (If you are successful retaining them, that is.) The spaces between items become bigger - until eventually, after months, you "burn" the item and it's taken out of repetition. However, at each stage wrong answers might get you downgraded again to an earlier stage.

The longer you use Wanikani, the more your review queue will become a mix. Mostly items you're trying to raise from Apprentice to Guru. Then occasionally items you want to push to the next higher level. Eventually also items you might burn altogether. You can control in which way your review and lesson queues are shuffled, but over time I became very comfortable with the default way.

The lesson queue, by the way, is how new material is presented to you. In the default way you move level by level, unlocking first new radicals, then new kanji, then new vocab. However, it's not exactly layered - if you learn new radicals (make it to Guru) you actually unlock new kanji, and with new kanji you unlock their associated vocab words. So by applying a different order to your queue you would see more kanji and vocab earlier. But... eventually I found that slogging through a big heap of vocab actually drives the kanji home, so I no longer begrudge that. The more vocab I've been learning, the more I appreciate it for how it builds my knowledge of actual Japanese beyond the individual kanji.

The Pros

The biggest is: Somebody put a lot of thought into this. Readings are not simply presented to be learned, each reading is tied into a mnemonic story, and the most useful readings are taught right away. The remaining, rather incidental readings are taught as exceptions with vocab that uses them. This becomes apparent with very common kanji over time.

Comes with a complete SRS right out of the box, usable right away.

There are mnemonic stories for almost anything, and they are memorable enough. There is also a certain consistency among them, and some sounds are tied to recurring characters - "kou" to "Kouichi" (the site founder) and "ken" to "Ken the Samurai."

For the "Chinese" readings, compound words as vocab drive them home over and over (almost all words comprised of multiple kanji are read with the "Chinese" on'yomi readings).

For the "Japanese" readings, single-kanji words and words containing hiragana help you retain those. (Almost all verbs and adjectives and single kanji words are read using the "Japanese" kun'yomi readings.)

The mixture of both readings helps you retain the most common readings for a kanji easily.

All vocab readings can be listened to, featuring two voices - male and female. These are actual people that have been recorded, so you would actually hear correct pitch accent and variations in pronunciations and pitch between different speakers. These can be set to auto-play.

Wanikani is often genuinely funny, just as the Tofugu Japan blog from which it originally was birthed. (Tofugu is a great resource for digging deeper into the Japanese language and its peculiarities while also being entertaining.)

Wanikani makes good use of gamification. It's clearly divided into levels, you get a funny email each time you clear a level, and you can see your progress displayed clearly at all times. This helps motivation. There are 60 levels to clear in total.

Wanikani allows other websites to safely access your progress as a web service if you enable that. I will show you in further installments how that is super useful when discussing other websites.

It's a really good way to train your hiragana. Since you enter all sounds in hiragana, you will retain them much better. My hiragana atrophied a bit from disuse before using Wanikani - no more!

The information is cross-referenced in many helpful ways. Words link to the kanji constituting them. For each kanji you can find which vocab uses them and which of those you have learned and/or unlocked already. Other little info bits have been added to vocab cards, like what kind of grammatical category a word is. Color is employed to highlight key points (meaning, reading, underlying kanji meaning).

It allows you to register synonyms if they are more helpful to your own memorization. This can be helpful if you already memorized a large stock of names for radicals from RTK and find some of Wanikani's choices counterintuitive. Be warned, however, that these will still be used throughout the mnemonic stories.

There is a community forum available to ask questions, but also a support itself. What you get out of them can be mixed, but it's better to have them. Yet another way for the world to know your opinion!

(I'm not endorsed in any way by Wanikani. In fact, it's me giving them money.)

The Cons

You cannot register your existing progress, by which I mean you cannot say "I know this already." Which would be annoying for people who learned some kanji already through other tools, books, or courses.

Teaching hiragana and katakana words was added late and its usefulness is... mixed. Since they are not displayed in context and with examples, ko-so-a-do words (where, there, this, that and such words follow a pattern in Japanese) for example are not taught in a very helpful manner. There are too few katakana words present to be really be helpful for training it.

Given its considerable success and having been around for so long, Wanikani evolves quite slowly and not always successfully.

Some of the information on the website is just not useful, like the "critical condition" items. There are words I have gotten wrong almost all the time and yet they never qualify for this category.

The organization of information for lookup is... not good. Searching works well, but most of the tabs on the top menu of the webpage organize information in unhelpful way. Try finding out which kanji you're currently learning at Apprentice level and you will see what I mean.

While Wanikani is clever in being forgiving about subtle ways in which your English answers were wrong (typos, and many variations of saying the same), it's absolutely unforgiving for Japanese. If you make a typo, your best bet usually is navigate back to Home screen - unless your mistake has been counted already. What it does nicely is accepting other readings that it's not looking for and advising you what it actually wants from you, like when you enter on'yomi and kun'yomi was asked for.

Wanikani costs money. Anything after level 3 is charged for, with you either paying monthly, annual, or you can pay a lump sum once for a "lifetime membership." That one actually goes on sale once a year in December. You're looking at several hundred dollars if you stick with Wanikani through the whole thing.

This is especially irksome if you actually have more time to spend per week and could go faster - which Wanikani doesn't really support. Yes, you can speed up your progress to a rather challenging degree, but if you are learning Japanese full time, you might find this limiting. It probably takes about two years to clear Wanikani for a lot of people.

Wanikani does two things - it teaches kanji and vocab. (And some useful terms for what things are called.) It does not teach you grammar or provide you with suitable in-context reading material. The context sentences it provides have neither audio nor do you have typically the required vocab to read them in their entirety. In fact, I have been ignoring them completely so far.

The Caveats

There is not much to say here. But if you dislike something about Wanikani, it's unlikely to change. The feedback thread in the forums is full of responses from people complaining about how useless it is to give feedback because it is ignored. I cannot say how I will see this in one or two years, but it's not a good sign.

Conclusion

Given all I said, I wholeheartedly recommend Wanikani and have bought a lifetime membership after being a customer for roughly 5 months. (I did it on sale, though.)

If you really want to learn how to read Japanese, this is likely a good choice, at least if you speak English well. While I cannot say this with the confidence of somebody who contrasted a lot of options, I certainly find my own experience with Wanikani to be worth my money.

Further installments in this series:

Remembering the Kanji / Anki

Bunpro

[japanese-cryptic-beauty]

In the Land of Electric Learning Aids Part 1: Remembering the Kanji / Anki

So, in my first post I laid down my personal journey of learning Japanese, mentioning a lot of resources, but in this one I want to focus on what they do (and maybe don't do) for you. If, like me, you prefer to be glued to a computer screen and be reminded by a machine to do your vocab and kanji reviews, I'll shortly dive into each option, its pros and cons and, maybe, caveats.

Remembering the Kanji

I'm rather sure this is the book that "started it all." Its premise is that in order to learn the Chinese characters which the Japanese call kanji, you need an approach that helps you remember how to tie their meaning to their construction. Therefore it presents you with the idea to learn a mnemonic story for each character that is built around one of its actual meanings and then learn to write the character itself according to this story.

For the first third of the book you are given these stories. They either help you to...

Memorize a radical: You learn to draw a simple character which is often repeated in constructing more complex kanji, a building block. One of the most simple ones is simply a square, and you learn to associate it with "mouth" since kanji generally avoid round shapes. Whenever this mouth shape appears in a kanji, the story will somehow mention a mouth. (The names of radicals are often optimized for making good stories and not necessarily tied to the actual meaning of the kanji.)

Construct a kanji: Here radicals are woven together by a story that ideally helps you to remember not only which radicals appear, but also the order in which they appear in. Since you're trying to draw the kanji yourself, knowing this order is necessary. The story is derived from a keyword which is one of the core meanings of the kanji.

One example of constructing a kanji would be the kanji for the keyword "I" composed out of the radicals "five" and "mouth." The mnemonic story counts the number of orifices in the head as five, and since the senses reside in the head, somehow this should make us remember that "five" above "mouth" is how we write "I."

Already, this early example presented me with the limitations of any mnemonic story - can you wrap your head around it? In this example, my own mnemonic story always was: "The five senses self." It's actually not a complex kanji to remember by any measure to begin with. But the original story didn't do it for me.

Before we move on to more practical aspects of this approach, let me say, that in my opinion, the intro chapter of this book is one of the great examples of deception, whether intentionally or unintentionally, as the author probably believes what he wrote to a great extent. I initially bought the rationale presented for the method hook, line, and sinker, but the longer I worked with the book, the more convinced I became that it was, indeed, bogus. (The rationale, not the method.)

What I consider a straight deception is stating that the intent behind providing less and less premade stories for the learner to work with is to supposedly help you in making your own. Not the fact that this was getting to be a lot of work and would have made the book considerably longer, no siree. While I appreciate the fact that finding a suitable keyword for each radical and a primary meaning for each kanji indeed was a lot of work done for us, I don't buy this part at all. You don't need to make hundreds and over a thousand mnemonic stories yourself to be prepared for making your own in the future when encountering a new, unknown kanji you want to learn. Bollocks, I say.

The other misleading advice why this method must be followed this way and this way only frankly seems to be rooted in self-aggrandization, even if subtly expressed. I'll not go into it.

On to more practical things.

How to work with Remembering the Kanji

There are three problems you might want to solve when working with RTK:

Keeping your learning organized

Coming up with stories

Drilling what you have learned

The answer to all three of these points is, by common consensus, Anki. Anki is, at its heart, a free software for preparing and rehearsing flash/index cards. Each deck of cards can be used for learning about a particular topic, and many premade decks exist, for example to help you cram for an important exam, at least standardized ones.

Anki helps you with 3. (drilling kanji) by employing the SRS - the Spaced Repetition System. It will present you with things you want to learn by spacing them out over time. Initially you see a card more often, and the more successfully you recall it, the less often you see it. This system supposedly helps with transfer from short to mid to long term memory.

Whatever is "on the front" of your index card is basically your question or your hook asking you to remember what you need to know about a topic or a word. In our case, this will be a keyword representing a kanji. You then proceed to draw that kanji, then flip the card, and self-evaluate how hard it was to recall the kanji correctly. You basically grade yourself.

Where Anki really helps with RTK is that many premade decks exist. These contain premade stories that people made for their own study, and typically also a lot of secondary information that got shared within the RTK community on Anki - annotations, the stroke order for each kanji, what level of the Japanese Language Proficiency Test JLPT this kanji belongs to, etc.

Since almost everyone working with RTK also starts using Anki at some point, they are typically seen as one approach these days.

The Pros

Pre-made Anki decks help you organize your learning.

Pre-made Anki decks provide mnemonic stories where the book does not.

SRS helps retain information.

Writing kanji provides a deep link between the key word and the kanji itself.

Mnemonic stories can engage other senses, like for example the visual imagination, which can aid in learning. You can employ whatever approach retains information best for you - associations, picturing it, etc.

You can synchronize your learning between various computers through Anki Web.

It's mostly free, though you would typically buy the books.

There exists a big and helpful Anki community.

The Cons

RTK "by the book" expects you to learn how to write 2,200 kanji before starting on how to read them / sound them out. (It personally took me 5-6 months to make it to the 1,000 mark.)

This bears repeating: You will not learn a single Japanese sound associated with these kanji until you start with volume 2!

It will not contribute in any other way to your learning of the Japanese language. This is kanji only.

Anki is very customizable, and finding a SRS pattern that suits you personally is not trivial and takes experience.

Anki comes with its own approach and vocabulary which can make it hard to assess if a solution for your own problem exists.

Self-grading is also not trivial, and you need to learn how to assess yourself in a way so that Anki's SRS algorithm will space out your lessons in a useful way.

The Caveats

I find that premade decks have one huge caveat. You get an unfiltered view into people's minds, and many of the decks so meticulously made have been provided by what seems male teenagers. There will be a share of mnemonic stories you would shake your head at, and maybe even be offended by. Your mileage may vary.

As I mentioned above, I find the rationales given in Heisig's intro to his own method and book at best unintentionally misleading. His generalization from his own experience is in hindsight neither convincing nor my own experience.

Whether learning how to sound out kanji "en bloc" is a good approach... I find it hard to say. This will make you aware of common patterns, definitely. But I find the organization of volume 2 hard to follow. It additionally presents every reading in katakana, only to resort to hiragana when writing out examples. The whole way volume 2 is presented as opposed to the clarity of volume 1 is off-putting and overly dense/compressed. Personally, finding this after wading through all of volume 1 would have been severely disheartening.

I found a good way to remember my own mnemonics, only to be constantly "chided" at each new sub-chapter by the author of how important it is to follow exactly his approach. It took me a while to build the confidence to ignore Heisig and trust my own instincts about how I learn instead of Heisig forcing his approach on me. (If you can force anything on somebody by writing a book, that is.)

Conclusion

Heisig does present a workable method in "Remembering the Kanji" if that is your only goal. His approach is to front-load your brain with the inherent knowledge what the individual kanji mean to put you on equal footing with, let's say, the Chinese when learning Japanese. Whether this is a good approach - I have my doubts. It does beat pure rote learning of kanji, though, especially at the snail pace courses dish them out.

Only the most disciplined or stubborn will likely walk this path. Also, it's almost entirely free, so if you're on a budget, what choice do you have?

In further installments I will present you with other electronic learning tools I employ in a similar fashion. Stay tuned!

Further installments in the series:

Wanikani

Bunpro

[japanese-cryptic-beauty]

Post 2: Correcting Post 1

Or: The journey of Japanese is a journey of folly.

So, I wrote that Japanese has pitch but it's usually not important for meaning, you just sound unnatural.

... Which is something you could say ... Kinda ... sorta ...

Only the Universe has been bombarding me with the importance and prevalence of pitch ever since. So I want to qualify the statement, at least.

Check out this interesting post I found here. Now, if I read this slightly cryptic answer right (it talks about "word-accent" which I assume can mean both "pitch" - Japanese, Chinese - and "stress" - English), pitch is significant to distinguish words with same sounds (homonyms) as follows: Chinese - 71%, Japanese - 13%, English - 0.47%.

Unsurprisingly, in a language (Chinese) using (in its standard pronunciation/main dialect) 5 pitches to differentiate words, pitches are "damn important" (71%). But 13% is actually also quite significant. Your mileage may differ.

Personal bias in skimming information

Now, for me, myself, this is not so surprising. I may have come across the information that pitch exists in Japanese probably twice but my brain may have chosen to willfully ignoring it.

(Textbooks may mention it but then omit it to not overload beginners, I guess.)

Thing is, I have a horrible history with learning how to stress words in other languages. These little accent-y things and squiggles on top of words? They're my nemesis. So I guess I wanted there to be no equivalent of that in Japanese.

I had three years of Ancient Greek in school, an investment of my time that taught me a valuable lesson: Don't learn Ancient Greek. I'm kidding - or am I? - the valuable lesson was to do my choices in a less knee-jerk way. In hindsight, nothing of value would have come of learning Ancient Greek even if I did well. Your opinion may differ. Fine. I really think it's a colossal waste of time unless you make a job choice requiring it. The best it could have done for me was get better at the learning itself or the learning of languages in general. But it failed in the most important thing that language does: connecting people. If it fails at that, it barely has any right to remotely exist as a subject. Nobody likes you Ancient Greek, go away.

Besides the letters, Greek gave me major problems with the stress accents. Put it on a syllable and I would never manage to pronounce the word. I always put the stress wrong. My brain hasn't unlocked how that works. I notice the same in Spanish. I may emulate a speaker (not that I speak Spanish, but when I repeat after someone), but even then I might not even hear the difference.

I know I've aggravated my Swedish teacher for not hearing the difference between his "ooh" and "uh" sounds he was making for the letter "u." (Ironically a song based entirely on the sound "U" is now playing on shuffle.) Hopefully continued exposure and awareness of the problem might help...

There's some good news, though

While my chosen learning method of WaniKani doesn't display stress, it features two speakers, Kyoko and Kenichi that say the readings. I wondered on occasion if they were real, but they do use pitch when pronouncing words. (And they might slightly vary between each other.)

So there's a neat little feature:

You pick your speaker and chose to let pronunciation play out every time the reading is featured - in case of reviews after you transcribed it yourself first.

I have to admit I was in the habit of not playing it when I could do multi-kanji word readings well, but this exposes me to pitch and pronunciation peculiarities I might otherwise ignore.

More exposure!

[EDIT:] Probably more good news, depending where you come from

So, in researching this I came across this article. It provides two pieces of context that I want to quote .

A reply to a letter to the editor of a manga magazine printed in Romaji (without accent marked) quoted in point 5: "When two or three words sound exactly alike except for pitch accent, context is going to resolve the ambiguity virtually 100 percent of the time. In practical terms, accent is probably the least important aspect of Japanese pronunciation no matter what your level of language skill."

In other words, even bad pitch accent will be understood almost always. This is where I first will employ the most horrible, stereotypical example that's ALWAYS trotted out to justify anything: the chopsticks/bridge example.

Chopsticks and bridge both transcribe to "hashi" (and to the same Hiragana). They only vary in rising and falling pitch. Now you could argue that makes pitch important. But as the above quote states - context usually resolves that quite reliably.

I mean, having browsed reddit and Quora for a while, you will inevitably see this being trotted out, and people ask the rhetorical question: Don't you think it's important to know if the bridge is burning or a set of chopsticks?

Rhetorical questions are, of course, only questions in the most technical sense. They're usually just people being smug or actually making biased or even bad faith arguments.

As mentioned above, people are not dumb. Japanese is already incredibly context-sensitive. You constantly have to keep track of things said before - because omission is common. You can leave out the subject because you mentioned it before as "the topic." Good luck with translating that, AI...

You usually can infer what is being said. But propagators of the burning chopstick dilemma are trying hard to make you think their way. Or are they even trying? It's so tired an example. If there were lots of these, surely people would quote them, too? Who cares what's going on with the damn bridge, anyway?!?

Okay, moving on.

I found this from point 6 a very useful and probably true assessment: "People without hearing impairments can mimic the melody of language, but they can hardly interpret visual accent markers into the oral/aural domain without special training because visual and auditory stimuli are processed very differently in the human brain. In all likelihood, the author of the above-mentioned letter simply feels more comfortable visually with accent markers. But using such markers to speak Japanese creates pronunciations that are worse than a crude synthesizer."

Well, I was notoriously bad at it in one language already. I guess I will hold my horses on trying to make that my method. Thankfully, audio material is available in enormous quantities, so we can learn from the melody of native speakers. And that is good news.

[japanese-cryptic-beauty]

Japanese... where to start...

It happened all so simply, so naively. I had started watching "Sword Art Online" the anime, and on a hunch, I decided to watch it with English subtitles along the original Japanese audio. (I had not done that in a while.) And at some point, not the first time in my life, I thought: Japanese doesn't sound so hard.

(And that I love the sound.)

Right. I was young and naive. So yesteryear. Literally. But eventually I acted on this naive, misguided notion and bought myself, after short thought, two books - the first installment of "Japanese From Zero" and, for some reason, "Remembering The Kanji". (It was actually recommended by several buyer comments/reviews.)

Quickly, through the book, internet resources, and YouTube videos, I realized Japanese did not have one written language, but three: Hiragana, Katakana, and Kanji. (Technically, it has four: Romaji, its Roman letter transliteration.)

Overwhelmed with this abundance of several exotic writing systems I initially, in fact, tried to avoid all of Japan's native writing systems and learn it from Romaji. Good old Roman letters! But I quickly realized that there was barely any market for Romaji products, and the books attempting to teach basic Japanese with Romaji all seemed to be geared towards businessmen. It had about the appeal of sucking on the end of a tablecloth.

So I decided to start with the first "Japanese From Zero book" which gradually introduced the Hiragana syllabary, a system of 46 basic syllables (and some combinations thereof) that produce all the basic sounds the Japanese language consists of. The reason Japanese can be written in this fashion is that it is actually one of the languages with the least amount of basic sounds to fashion words from. (Which made it sound deceptively simple to my ear...)

Hiragana can be mostly used to render all of these sounds. It is taught to kids as a first entry into writing to start out with, it's generally used to transcribe Japanese phonetically (more further down), and it plays a role in the grammar of the written Japanese language.

So I threw myself into it. I learned to write the 46 Hiragana letters (and read them), and endlessly reproduced them, tested myself with flash cards, etc. But I also had ulterior motives... I hoped that if I learned Hiragana well, I might get away with not having to engage with the more elaborate actual writing system of Japanese. It was an attempt at a horse trade - if I do this, I don't have to do that, right?

And I progressed in the book. But it began to sink in... It's good and absolutely vital to know Hiragana, but knowing Hiragana will not be sufficient. (Well, unless I'm content with reading books made for young children.)

How written Japanese roughly works

Japan early on imported the Chinese writing system, including its letters, the Hanzi - eventually known in Japanese as Kanji ("kanji" in Japanese literally means "Chinese Letter"). There are 10,000s of them, but in actual use the Japanese written language resorts to roughly 2,200 of them (the amount a high schooler would learn), and maybe 3,000 in academic settings.

Each of the letters has its own meaning and by now several ways to sound it. But using Chinese letters clashed with one fact - the Japanese had already their own language and grammar, and the Chinese grammar wouldn't work for them. Japan kept importing Chinese culture, including Buddhist and Confucian texts, but eventually Hiragana and Katakana developed to aid the Japanese speakers. Hiragana supposedly as a simple writing language for noblewomen's diaries which eventually evolved to novels. And Katakana filling the exact same space but developed by Buddhist monks to annotate Chinese Buddhist texts. Both evolved into full 46 syllable languages that you can render Japanese with.

In order to remedy the fact that Japanese grammar clashes with Chinese writing, the Japanese eventually started to combine Hiragana with the Kanji. Kanji would denote the concepts, but Hiragana letters would fill in all the Japanese grammar to build sentences. So most verbs involve one or more Kanji to define their meaning, but Hiragana is used to add verb endings for the various forms the word can take - whether the action is in the present or past, if it's continuous, an imperative, and how polite the speaker is, etc.

食 means the concept of "eat, food." It can also be used as the noun for "eating" or "food."

食べるmeans "to eat." Here two Hiragana letters are added to form the verb. Different verbs can be formed from the same Kanji. They seem to thankfully all end in a "u" sound.

(Never mind that 食 is sounded out completely differently in these two cases. Food can be called "shoku" whereas to eat is pronounced "taberu", the kanji being the "ta" part and "be" and "ru" being the two Hiragana that follow.)

Hiragana fills in all the missing parts that Chinese didn't cover for the Japanese language. Hiragana letters will tell you which word plays what role in the sentence (who am I doing something to? what is being talked about?), they differentiate between different words, adjectives, and adverbs, and sometimes form the connective tissue in compound words.

It comes all together now, right?

By the way... Katakana also has uses for foreign loan words. Because one clearly needed another writing system for that shit. Don't get me started.

Remembering The Kanji: The Next Delusion

So one day (maybe two months after starting at most) I decided that halfassing wasn't going to cut it. I wanted to take the bull by the horns. I decided that if I wanted to learn Japanese, Kanji weren't optional.

I arrived at this somewhat crazy notion by the detour of this book: "A Guide to Japanese Grammar: A Japanese approach to learning Japanese grammar". The author was doing a good job of summing up Japanese grammar, but he kept insisting that the proper way to learn Japanese was using Kanji right from the beginning, eat or die.

The first rule of the Japanese Learning Club is... Never, ever, trust somebody who learned the Japanese language the way the Japanese are taught it. My experience is limited, but sometimes I think people just like to share their own flawed methodology around because they can't think of anything better. (This seems to be a common gripe you find about Japanese classes in general on the internet.)

But I needed a game plan... and I had by random chance bought a book to just do that: "Remembering The Kanji" (RTK or "Heisig method"). It was developed by a man who learned the Japanese kanji in record time before learning Japanese itself. His reasoning was that Westerners notoriously have a hard time learning Japanese, but Chinese and other people that use or used the Chinese Hanzi have a head start. They have a whole library of those "kanji" (or a lo of them) in their head and their meanings. They just have then to grapple the intricacies of the Japanese language itself.

So I changed approach.

The first volume follows the concept that the kanji are built up from basic symbols called "radicals." More complex kanji are formed from more simple kanji and radicals. And they are more easily learned if you actually order them that way instead of a haphazard order.

Or, to put it another way: You don't want to start with kanji involving 48 brush strokes when you can start with those that involve 1 or 3 and build from them. Japanese school kids have a lot of time, repeat endlessly, and spend an amount of time on this with their brains at their most pliable stage. Let's face it - I'm neither blessed nor cursed by these circumstances.

I was convinced by the system and the logic behind it.

The basic methodology breaks down to: Learn a concept. Learn a mnemonic story that helps you remember how to produce the kanji in writing while tying it to the concept. Use flash cards (Anki is the free app well-suited for that) to drill myself into repeatedly learning to write these kanji. The author of the book insists that going from concept to writing them is best initially.

I learned to write 1,000 Kanji from keywords in 6 months.

But once I crossed over the 900 kanji mark, things came to a crawl. When initially I had learned maybe 20 kanji a day, then 10, now it was 5 a day. I was feeling more worn every day.

Yes, there was the satisfaction of recognizing a lot of the Kanji displayed in anime. But just as often as not, I still didn't. (Remember, 2,200 are commonly used.)

I had written whole notepads full of kanji, but I was increasingly having doubts, especially when the stretch from 900 to 1,000 really started to go on and on.

I once had clung to the notion that Japanese could maybe be read into English without even knowing Japanese - something which is partially true. But it was less and less appealing to me. I realized that I was not making good use of my time, because I didn't build vocabulary. Because I still didn't know how Japanese words sound.

So, while I was predominantly watching anime at the time I couldn't parse those sentences into "tokens" because that's hard to do when recognizing only a select few words. I noticed that words I knew I could already spot in the stream of sounds, and if I knew more words, this would happen more often, until I can "tokenize" the Japanese language into separate blocks reliably that my brain can process in turn.

And the "Heisig method" was not helping with that. When does reading kanji into sounds happen? Volume 2. The advice is seriously to learn 2,200 kanji first and then learn readings next. This was absolutely disheartening. I couldn't imagine slogging through another 1,200 kanji before even starting on sounds. This method is fine for monomaniacs, but for me it was seriously wasting time I could also use to build my own inner parser of spoken Japanese.

My goal had been to understand spoken Japanese first and foremost. I felt like I had lost my way and wasn't getting the results I actually wanted.

So I changed plan.

The current plan: WaniKani

The sunken cost fallacy is a wonderful thing. I had learned many things about the Japanese language at this point that, had I known them at the beginning, likely I wouldn't have started to learn it. In fact, it's widely regarded as one of the hardest languages to learn. Silly me.

But I had spent six months getting 1,000 kanji down. I was no longer deterred by the idea I might not be able to distinguish them. And I wanted something to show for it. There's also a certain beauty to all of it. Even if there also isn't.

And yes, I had spent a lot of time, and when you spend a lot of time on anything, you start to identify with it. And that makes you spend more time on it.

Inside the Anki index cards set for the Heisig method you also find references to where these Kanji are used - are they part of a level of the Japanese Language Proficiency Test JLPT? It will be noted on the card. And something else was mentioned. A "WaniKani level..."

Turns out, WaniKani is a website based on a similar spaced repetition approach like Anki itself. But WaniKani does a few things that the Heisig method doesn't. It starts with the combination of sounds and kanji right off the bat. And it teaches you vocabulary built from these kanji.

So I started over!

Yep, I started over at the 1,000 mark to try out a different system. Wasn't I just complaining about monomaniacs??

The first 3 of the 60 levels of the WaniKani website are free. I am now 5 months in and approaching level 12 (the 20% mark). I have learned about 400 kanji and 1,100 word of vocabulary in the process.

Wait. Didn't I originally learn 900 kanji in roughly the same time period...? Yes and no.

When using the Heisig/RTK method, I learned to tie concepts to kanji and write them. But I wasn't learning Japanese, really. Now I'm learning how to read these letters into sounds, remember the concepts behind them, learn the multiple sounds behind each (!), and many of the actual compound words formed from them.

Does that make a difference? Oh yes!

自転車 means "bicycle" and is pronounced "jitensha."

With the Heisig method I was able to see the letters in an anime, translate them to "self 自-revolve 転-car 車" and figure out that this might mean "bicycle." But I simply didn't know that these specific letters form that word. And I couldn't sound it out, nor would I have picked it up in a stream of Japanese audio.

So if anybody told me now that RTK is the way to go to really learn Japanese? I would probably cuss them out. Yes, some people can frontload their system like that. It might work for them. But it's a very poor approach for people who like to associate things in their thinking. (Don't get me wrong, it still beats many other approaches to learning Japanese and especially the kanji where people fail to outright make much progress. Or so it is claimed on the interwebs.)

I not only learn the most common sounds first, I learn to use them. I build words, and these words I might spot in turn when listening to Japanese. It gives me an avenue to practice Hiragana as well, even Katakana. So these 400 kanji? They have become building blocks for nouns, adjectives, adverbs, and verbs. And these 1,100 vocab words glue them into my mind.

But you can't stop there. Or I can't.

Where to go from there

You know, when I think I have understood how Japanese works, it throws me a curveball. I thought, for example, that unlike Chinese, Japanese is an unpitched language. (In Chinese five distinct pitches are part of the sound and alter meaning.) I was consoling myself that the endless parade of similar sounds in Japanese at least meant I didn't have to tackle pitch.

Turns out I was wrong.

Japanese has two-pitch melodies, but only on occasion it actually has impact on meaning. You will, for the most part, just sound unnatural. Hooray!!

[[EDIT: I want to qualify this statement and do so here.]]

Sunken costs, deliver me from this insanity. No? That's not how you work? Okay. Moving on.

There are many crazy things about Japanese, including the many different sounds associated with letters. But I'm grappling that part.

Grammar is another thing. Thankfully, a fair amount of Japanese grammar actually comes down to comparatively few rules and forms. I mean, technically Japanese only has present and past. (Say whaaaat.) Or so I've come to understand. If I learn it's much more complicated at some point - sure, sunken costs, dig the hole deeper.

Reading means looking up things

But I want to start reading my manga in Japanese. (To get practice with the language itself and reading.) This poses some problems right off the bat. Let's say you need to write down a kanji for looking it up, how do you do it?

The most inconvenient method is drawing it with the mouse into Google Translate. You can do it, but it's time consuming and not fun at all. (Remember? There are 48 brush stroke kanji!)

And here's why I say WaniKani has a lot of merit. If you know one of the words using this kanji, or one of its readings, you can write that in Roman letters (or Hiragana, doesn't matter) into an input field of most apps supporting Japanese writing and it will give you a list of possible kanji.

That's how kanji are written on Japanese phones - they keep writing their own phonetic language in Hiragana and the phone offers you which kanji fit the sounds. (A lot of sounds have many, many kanji associated with them.) But that's why speaking (or at least being able to "sound in") Japanese is so vital. Because the support for writing Kanji is through Romaji or Hiragana - you need to know the sound / phonetic writing of what you're saying and the computer finds you the matching Kanji.

("wakaranai" - "don't understand")

This is why the Heisig/RTK method helps you so little with actually getting stuff done. Imagine painting all these kanji just to look them up. What you need is a connection between actual words, sounds, concepts, and kanji. Then you can efficiently use modern tools to look them up. This links you back into part of the Japanese mainstream and the tools it offers.

If I write "jitensha" into Google Translate, it will offer me "自転車" to select. I can go from sounds back to kanji. That's immensely helpful and direct, even if I only need one of the three kanji involved.

In the end, I guess, learning Japanese will mean pulling from many different sources. But I think now I'm building a base that I can work with in many regards.

So, that was my journey so far. I still have lots to do! But I don't feel lost. Or so I tell myself.