My last piece on Japanese immersion, was entitled Massive Input vs Anki. A slightly provocative title perhaps.

I didn’t use it because I was trying to start a war. It was because I was and am interested in the interplay between an artificial learning method and the organic aquisition that comes from pure immersion.

I am not suggesting that we all give up Anki. It is a powerful system that helps us acquire and retain Japanese. However I do find that as one uses increasingly massive input one’s relation to Anki changes. I promised to report on my personal experience of intensifying my input.

I spent a month doing an unusually large quantity of input, mainly watching anime with subtitles and reading books. How did it affect my relation to Anki? Let me summarize:

I have two Anki decks. One for vocabulary and a sound-based sentences deck. I found I was actually entering more into the vocabulary deck because I was picking up vocabulary very fast from my extensive reading (strictly with pure extensive reading you don’t look words up, but I was).

Often with a new word I will enter sentences because definitions alone don’t tell you the nuance of a word or how it is used. I did this a lot less. The reason was that with a higher level of Japanese immersion input I found new words recurring much more frequently than I would have expected.

Not only do the same authors/themes tend to use the same words but I was finding words repeating across different books and anime. I felt more confidence in being able to rely on seeing the word in the wild rather than needing example sentences for it. I was popping words (but by no means all my new acquisitions) into Anki to “pin” them. But I felt more confidence in natural encounters with a steeply increased rate of input.

I was also more lenient with myself in Anki. Usually I am pretty strict with myself. “Half-right is wrong”. What I found was that again, with a higher level of Japanese immersion, I had more confidence in my input. So I would think “all right I know pretty much what that word means. I’ll recognize it in context next time I meet it”.

This, of course is how children naturally learn language. They start off knowing roughly what a word means and sometimes use it incorrectly. But with continued exposure their knowledge of the word refines down to exactitude. With confidence in our rate of input we can be less afraid of getting words half-right because we know it is a step along the road and not our “only chance”.

Another thing I noticed is how wonderful the memory really is. Reading and watch/reading quite fast through a lot of material and resisting the temptation to look up things I did really know even if a little ambiguously (tolerance of ambiguity is important here I think) I noticed that I was recognizing words I had learned two years or more ago and hadn’t drilled since. I think we tend to over-drill words for fear of forgetting them when the real drilling they need is actual Japanese immersion: input, input, input.

I might not have recognized those words on an abstract word-list or in Anki. But in context, as living parts of real sentences that were in themselves living parts of a real story, I did remember them. And there was a sense of “Gosh, so that’s what that old glassy-eyed dead word in the attic of my memory looks like when it comes to life!”

Of course, nearly all my vocabulary came from real exposure in the first place, via the Anime Method and its close relations. But some of them had ossified into “Anki words”.

Real Japanese speakers know a huge number of Japanese words. How? By having total Japanese immersion all the time. But even with the degree of input possible to you or me, if it is truly massive it is surprising how quickly words start reinforcing.

Japanese Immersion: The “feel” of the language

Japanese, in both grammar and vocabulary, is a surprisingly “modular” language. As you get used to it you start to realize how beautifully it all fits together.

If you are reading (text or subtitles) more and more you encounter words that you don’t need to look up. You know what the individual kanji mean, and, with the context, it is clear what the two (or three) of them in combination must mean. Japanese has a lot of words, but the good news is that a lot of them come from understandable-in-context combinations of other word-elements one already knows.

One also gets the feel for Japanese sound-symbolism. This is a very unquantifiable subject, so I won’t say much about it. But I do find myself saying “Ah I can tell what that word means by the sound of it”. In some cases I may be subconsciously remembering a word I learned before. In other cases it may just be similar to a word I learned before because similar meanings often have similar sounds. It may also be based on lots of tiny frequently-occurring cues, such as the fact that sharp “s” sounds can tend to imply “doing” while gentle “r” sounds can tend to imply “being”.

I probably won’t ever write anything more detailed on this because there are no exact “rules” and countless exceptions. This isn’t a “method” one can “use” in a conscious way, but as you get used to Japanese you will find it working for you more and more.

Relatedly, as you read more and read faster, especially in books, you will find yourself anticipating. You notice that you read three words at the end of a page and while you are turning the page over you already know what the next one or two words will be. Just as in English, when you read “He had shifty” you know the next word will be “eyes” or “she combed her” will be followed by “hair”. If we know that she is attending to her own appearance (or just that she is alone) we only need “she combed” to know that “her hair” will probably follow.

Anticipation is a huge part of understanding. So is the recognition of words. Many people, reading in their native language, will pronounce a word wrongly when they have never seen it before (or only seen it occasionally) even when the spelling is not ambiguous. Why? Because we have long ago stopped spelling out words we read in our minds. We recognize general shapes and anticipate words. We don’t expect to really have to read an individual word. At most if it is a little unusual we might check if the middle letter is an e or an i. Actually reading a word – looking at each letter – and constructing the sound from that, the way a small child or a foreign learner does, has become an atrophied skill.

That is a good thing. It means that we can concentrate on the meaning of the text as a whole rather than expending mental energy on fully reading words we already know.

Japanese immersion through massive input helps us to develop toward the same level in Japanese. We start to realize that we are anticipating things and sometimes taking a word or phrase super-fast because we only need to glance over it to confirm that it is what we thought it would be.

This isn’t only a reading skill. As we learn what words tend to go together it becomes easier to hear bunches of words in speech.

Returning to the “Anki question” I reiterate that I am not really talking about replacing Anki. I don’t think that is a question one needs to worry about. What is important is to become aware of the importance of truly massive input and its function in organic learning. Considering that one only has so much time, this may impact the balance between Anki and actual input. But it is a delicate and shifting balance and one that each of us can work out for herself over time.

Fortunately Anki is not the all-devouring monster it can occasionally appear to be. Feed it less and it will diminish fairly quickly. My daily Sentences Anki is currently noticeably smaller. I may have occasion to step it up again later and I am still adding to it but not nearly as much as I was.

After an experimental month of much more massive input, my personal finding is that it works if anything better than I expected. I will be continuing with it and I am somewhat shifting the balance from Anki to Japanese immersion through massive input. It is a matter of degree. I always did a lot of input anyway, and I never regarded Anki as my primary learning tool (that was and is native Japanese material).

So really I am only moving a degree or two further in the direction I was always going. One can sometimes, I think, use Anki as a bit of a security blanket. We think it will stop us forgetting words (as you know from some of your old words that pop up looking like strangers, it won’t always do that). We distrust input experience as being too random.

I am beginning to feel that Japanese immersion input lets words and grammar drop only a little more than Anki does, provided it is sufficiently massive. And it keeps them in our mind in a live, organic way, rather than an abstract one, which makes them a lot more useful to us.

NOTE: Naturally this article is relevant primarily to intermediate learners who are advanced enough to be able to step up their rate of input.

Bonus Japanese Immersion tip: SMILE!

One final tip, which you may write off as a bit of doll-craziness. I don’t mind if you do. My royalties on it are very low anyway. In fact, come to think of it, I don’t even get any royalties.

When, in the course of your massive input you “ping” (as I call it) a word or phrase you recently learned, or one you are happy to remember, smile. Smiling makes you feel happy (try it). Or make a little “ding” sound in your mind. Or jump up and down and ワイワイ all over the room. Probably not the latter as you are trying to read quickly.

The point is to tell the pleasure centers in your brain that this is a Good Thing. Even a little fleeting smile will do that. Don’t try to re-memorize it (unless you have a special reason to). Let your Japanese immersion input work on that. Just very quickly “flag” it for your mind.

If it disrupts your reading you don’t have to do it. Just a little trick you may like to try occasionally.

Now read: Japanese Immersion: Massive Input vs Anki

This site’s method is fundamentally Japanese immersion. While we do recommend learning basic grammar as a kind of “cheat” for getting started, the heart of the system is watching anime. It is also vital, after the initial stages, to have a lot of actual communication as well as input.

Which sounds kind of lazy, but if you know anything about the system, you know that watching anime in Japanese with Japanese subtitles is pretty intense, especially at first when you don’t know much.

The idea isn’t that it is easier than other methods (there is no such thing as a “get fluent quick” technique”). It is that you are learning Japanese “in the wild” from early on. Anime becomes your university, exposing you to real Japanese, building a core vocabulary organically and making the language a part of your life. And learning the written and spoken language in tandem.

The strategy of the anime method is to build your capacity for input. At first it takes you hours to get through a 20-minute episode. You have to look up half the words, investigate grammar etc. As you build vocabulary, learn how the language works and generally get used to Japanese, you become faster and faster.

You are approaching the point where true Japanese immersion through massive input becomes possible.

Now from the beginning we have recommended Anki as the means of learning vocabulary and at least some grammar. Anki essentially makes the old method of flash cards efficient and scientific, exposing you to each card at exactly the rate you need to learn it.

In a sense one could say that this is not true Japanese immersion. Some people, notably Tae Kim, whose excellent grammar site we strongly recommend, believe that massive input is the way to learn vocabulary and that Anki is too artificial. So in one sense one can say that massive input and Anki are opposed — in that one could conceivably (at least in some people’s view) replace the other.

Or rather massive input could replace Anki. If you can learn all the vocabulary and grammar you need through massive input then you clearly don’t need Anki. If you can learn them all with Anki you still need massive input, or all you end up with is a massive word list and a lot of abstract grammar.

So theoretically massive input can replace Anki and give a truer Japanese immersion environment. And practically too. One of my most respected (and advanced) senpai has never touched Anki or other artificial learning tools.

I have also made it very clear (well maybe not to everyone as I mostly wrote about it in Japanese) that I do not intend to use Anki forever. In my approach to Japanese immersion, I regard Japanese as my language. In my heart Japanese is much more my mother tongue than English, even though I currently don’t know it nearly so well. One does not use artificial learning tools for one’s own language. Like learning abstract grammar, it is a little cheat: a trick to get you over the (very tough) initial hurdles.

Japanese Immersion: replace Anki with Massive Input?

So, do we need Anki? Can we replace it with massive input? My answer to this breaks into two parts:

1. Yes, we can and should, eventually. Anki is like water wings. Unless we want to regard Japanese as a “foreign language” for our whole lives (and some people, of course, do) we do not want to be using artificial tools forever.

2. It is possible to learn without Anki (or anything similar) from the start. Some people do. But this raises certain questions:

Firstly, how do you learn? Some people can just hear a word and recognize it again, at least after a couple of times. I tend to need in many cases a mnemonic to tie the sound to the meaning. Otherwise it doesn’t stick. Mnemonics are temporary, but they are like the rough stitching that holds the fabric in place while it is being properly sewn. Some people don’t need them. I do, (though increasingly I make my mnemonics in Japanese) and I find Anki the perfect place to use them.

Secondly, and rather more importantly, how do you deal with kanji?

This is not a rhetorical question. There are several ways of dealing with kanji. A lot of people (and probably the majority of Japanese Immersion methods recommend this) begin by working through Heisig-sensei’s Remembering the Kanji which involves learning all the kanji in the abstract without knowing the pronunciations or any words associated with them. At a later stage many people learn abstract lists of the kanji’s various pronunciations.

Other people learn by the traditional method of writing each kanji out hundreds of times.

We don’t recommend either of these methods. They aren’t wrong. They do work for some people. It just isn’t how we do things. We believe in learning kanji along with everything else, organically. Our motto is “learn words, not kanji”.

The problem here is that kanji learning, unlike language acquisition, is not inherently “organic”. Children do not “just pick them up” the way they do vocabulary, grammar, speech and listening comprehension, through natural Japanese immersion

They learn them in class. And it takes them years.

But, they do not learn kanji in the abstract. From their first class they already know the words in which the kanji are used, and how they are pronounced. They never learn lists of on and kun readings or abstract Heisig-style “keywords”. They never (except possibly with some rare ones in high school) learn a kanji of which they don’t already know many real-word uses and (therefore) the main pronunciations.

So learning kanji organically “as you go along” is actually closer to the way they are really learned by Japanese children. Since we don’t have from-birth Japanese immersion we can’t replicate it exactly, but we feel our method comes closest to the real thing and to the way one can naturally learn them with a minimum of abstract study.

But kanji need to be drilled. Grammar and vocabulary may be “drilled” by pure Japanese immersion and massive input (and output), but kanji really do need to be drilled by some essentially artificial method or other. There is the endlessly-writing-them-out method, which I believe is overkill in a digital age. There is Anki (and similar systems). There may be other methods that work for you (do comment below). But one way or another, you are going to have to drill kanji.

For me, the simplest, most painless and most organic way of doing this is to use a simple vocabulary Anki deck with the kanji on the front, and the readings, definitions and any necessary mnemonics (for kanji structure and meaning and reading) on the back.

I won’t go into more detail here as this article isn’t about learning kanji, but by this method one is learning kanji and vocabulary at the same time and learning readings in the most natural way: one at a time, as one encounters them in real words.

But because learning kanji (like learning the alphabet, except that there are a lot more of them) is not natural like learning to speak, there does need to be a somewhat artificial way of drilling them. Pure Japanese immersion alone will not teach one the kanji, any more than it teaches Japanese children the kanji.

But, assuming one does use Anki for this, once one has built a solid organic core vocabulary and made friends with a large number of the more common kanji, is there a sliding scale between Anki and massive input?

Does pure Japanese immersion start to take over? At what point do the water-wings come off? Is it immediate or a gradual process?

As with most things, this will differ with different people, but I have been having some very interesting experiences with sharply increasing my Japanese immersion and massive input and its effects on my relationship to Anki. Read all about them in Japanese Immersion: Why massive input is necessary.