Friday, November 18, 2016

Reading, Writing and the Future


I'd intended to post a wrap up after the August Tadoku round, and here it is two weeks after the end of the October round and still no post.  Which should tell me something, and it does.  But before I get to that, here's what I've read over the past few months.

These are all from つばさ文庫 and they fall into two categories.  こちパー (as it is known) is classified as  小学中級から and I had some difficulty understanding numerous scenes, though I was able to follow the overall story.  The rest are classified as 小学初級から and I'd say I understood these at about the 75% to 80% level.  This is low enough to be occasionally frustrating, but I was able to get at least the gist of every scene.  So broadly I'd say 小学中 is a good fit for intensive reading and 小学初 is a stretch, but doable, as extensive reading.

Since the October round ended I've been wondering if extensive reading is the best use of my time right now.  It seems extensive reading is a good way to synthesize and reinforce what is already known, but not a great way to learn anything new.  I was able to guess at a few words from context, but my understanding is that that's a much more viable approach at 95% comprehension.  So maybe at the moment I should be focusing on acquiring new vocabulary through intensive reading?  On the other hand, having a goal may have helped me build up my overall reading stamina, so there is probably some benefit there.  Well, I don't need to decide whether to participate in the next round for a while yet.

In the meantime I have been pursuing more intensive reading.  Which brings me to Steve Kaufmann (blog, YouTube channel), polyglot and founder of  LingQ is an intensive reading system that I think is a brilliant idea but which entirely fails to work for Japanese.  The broad idea is that one loads texts into the system (or uses their large library of texts provided by native speakers), and marks words as either "known" or "learning."  As "learning" words show up in other texts (or are re-read in the original text) they gradually become "known."  There's also a flashcard component, but even Kaufmann-sensei says he barely uses this feature.

The problem is that the system defines "words" based on spaces between them, which is not an option for Japanese.  They kind of kludge around this and it kind of works, but in my opinion not well enough to be useable.  By the way, I would be remiss if I didn't mention that there's a free open-source implementation of this idea called Learning With Texts, which suffers from same problems.

Implementation aside, the core idea really appeals to me.  It's a systematic version of what I've called "Tae Kim's deluge approach," and indeed Kaufmann-sensei also uses the word "deluge" in at least one place.

As it happens, there's already a program that does all the hard work of splitting Japanese text into words, and that is Rikaisama!  So I put together some utilities that utilize Rikaisama to give me a LingQ-like system that, while still not perfect, works well enough to be useable and useful.

So my process goes something like this:
  1. Buy a KOBO book from Rakuten and download it to the desktop application
  2. Use Calibre and DeDRM to convert the book to HTML
  3. Use a little utility to convert the HTML to plain text (also removing the furigana).
  4. Load the text into a browser and read it with Rikaisama.  Initially all words are colored blue. As I see new words I can press S to mark them as "learning" and turn them yellow, or "X" to mark them as "known" and remove the color.  The upshot is that text goes from this

to this

and, hopefully, eventually, to this

Of course this whole marking process isn't necessary, but I find that hilighting words I'm learning helps me focus on them and, maybe, remember them a bit better.  Another nice feature of this approach is that the process is minimally invasive to the reading, so I actually can focus largely on enjoying the story.  I could also do the same thing with anime transcripts, which would let me both read the material and then listen to it for reinforcement, which is something else Kaufmann-sensei recommends.  I can also use LingQ's library of content, both audio and text can be freely downloaded although it is necessary to set up an account.  Plus, there's a natural path to monolingual when I'm ready for it, I just need to switch to Sanseido mode.

Of course all of this coexists with other Japanese activities; watching unsubbed anime (usually, though not always, after first watching it subbed), playing どう森, flipping through manga (which I am doing extensively), listening to music and podcasts and audio from anime, and so on.  I've also been participating on the KawaJapa forum!  It's seen a surge of activity recently, and there's now a Shoutbox on the site and an affiliated LINE group -- all in Japanese, of course!

Getting here has been a long journey and most of it is still to come, but it finally feels like I'm at least finally on a path and making progress.

So... this blog.  I think keeping it has been useful, both as a way to work through and record my thoughts and for the wonderful advice and encouragement I've received from several people.  But, as the recent gap shows, it feels like it has served its purpose.  Henceforth I'd like to focus on activities in Japanese.  At some point I might even start a blog in Japanese. Just at the moment that seems like more time and effort than it would be worth, but maybe next year some time.  So, until then...


Thursday, July 21, 2016

An entirely expected decision

In a move that should surprise no one, especially not me, I stopped using Anki a few weeks ago.  Yes, again.

There was no deep consideration behind this decision, there was just an evening when I got home, looked at the pile of reviews, thought "I'm sick of this," and stopped.  Although the review counts had dropped off a little after I stopped adding new cards, I soon reached an equilibrium where cards were becoming mature as fast I was forgetting existing mature cards.  The process was almost comically Sisyphean, except instead of one boulder once a day it was 1500 or so pebbles over the course of a few weeks.  Mnemonics helped a little, but not much in the long run.  I'd also churned through a number of alternate formats including audio on the back, pictures on the back, sentences on the front, and other variations, none of which made any significant difference.  Well, almost none.  I had one deck with kanji and furigana on the front and a mnemonic on the back, and that deck went shockingly well, with mature retention rates hovering around 90%!  It feels like this is telling me something important, through I don't quite know what.  I did consider sticking with this approach, but I never really grew to enjoy the process of devising mnemonics, plus this doesn't really solve the problem I want to solve.

I can't say that Anki has been of no value, it's more that the effort to use it always ends up seeming to outweigh the benefits.  I can't shake the feeling that there must be some variation of spaced repetition, or vocabulary drills more generally, where the cost/benefit ratio would be favorable, but just at the moment I'm tired of searching for one.

Instead I'm going to put my energy into immersion.  The thing is, there is broad consensus that massive input is necessary and indeed some (notably Tae Kim) have even found that immersion alone is sufficient.  Conversely, I don't think anyone has claimed that Anki alone is sufficient, and it clearly isn't necessary as people became fluent in second languages long before Anki existed.  So if I'm going to work on establishing habits that are conducive to making progress, it seems that effort is better spent on immersion first and foremost.  I've also struggled with immersion, and in particular with frustration over lack of understanding, but the advice I received from my senpai on my previous post was really helpful.  In particular, Cure Dolly's suggestion to "suppose there is no other language BUT Japanese."  In some ways that's just a small mental shift, but I've found it made a significant difference in the degree to which I'm able to just relax and enjoy Japanese media.

Of particular note, this means I've gone back to a monolingual dictionary!  Whenever a word catches my attention I'll look it up in 例解学習国語辞典.  Sometimes — rarely, but more often than I might have expected — I'll understand the definition straight off, at least to some degree.  If not, well, I don't.  Then every few days I'll go through the words I've looked up (the app helpfully keeps a record) and do an image search for words whose definition I didn't understand.  If the images still aren't helpful then I'll look up, in English, the words in the definition and example sentence that I didn't understand.

Although I'm not putting these into Anki I am compiling a virtual notebook, which is something Rachel suggested a long time back.  The entries look like this:

The graphics are a taken from the KanjiVG project, modified to give each radical a different color.  My experience with the furigana deck suggested that, despite going through RTK (twice!) I still haven't really made friends with the kanji and I thought this might be an easy way of getting to know them better.  One thing I noticed right off, if I look at a random word for a few seconds and then try to write it from memory I almost never succeed.  But if I look at a word with the radicals highlighted I often can!  As noted in the example of the chess masters from this KawaJapa post, it really is all about building a context.

So, with all that said, here are some things I've been immersing with lately!


I enjoyed both seasons of the イカ娘 anime, there's nothing profound to it but it's breezy, good-natured comedy with the occasional moment of genuine heart.  I picked up the first few volumes of the manga and it's pretty much the same (in fact the anime's adaptation was pretty faithful, although the manga contains some additional material).  The chapters are fairly short and there's furigana, making for a pretty quick read.

ヤマノススメ is another instance where I liked the anime enough to want to read the source material.  There's no furigana, but the dialog is pretty straightforward and conversational except when they get into detailed discussions of mountain-climbing equipment, and those sections can be skipped without really missing out on anything.  The stories are warm and gentle and are as much about friendship as mountain climbing, and the art is pretty and based on real locations. When I get to Japan someday I really want to visit 高尾山!

I found チョコミミ by searching for "good manga for children."  I'm only a little ways into the first volume so far, but the characters are endearing and the whole thing has this really nice, fun lively feel to it.  And I genuinely laughed out loud at a couple of strips!  There is furigana and conversations tend to be simple... with the exception of Mimi's dog Chiffon who seems to use some archaic or formal language, including referring to herself as 拙者[せっしゃ], a pronoun I don't think I'd ever encountered before (although it is on Jalup's list of 100 ways to say "I").  I gather this series is fairly popular, there are English translations of the first few volumes available as well as a live-action adaptation.  Although I've only glanced at it, I like the way they gave it a manga-like feel!


I forget what lead me to the children's novels at 青い鳥文庫 and つばさ文庫, it might have been a thread over at the koohii forums.  I found this particular title by looking for covers that seemed appealing and cross-checking against titles available as e-books for Kobo.  This turned out to be a good pick!  The core story is compelling; Yuno has just started middle school and her dream is to form a club to revive the magazine "Party" which her late father edited and treasured.  So far it's mostly been about her experiences starting school and encountering the cast of endearingly eccentric characters who, clearly, will eventually become her close friends and club mates.  The book is amusingly genre-savvy.  Yuno's mother writes manga, and at one point she sets all the clocks in the house back an hour in order to make Yuno late, in order to recreate the classic "middle-school girl running out of the house with toast in her mouth" scene for reference!

(Informally, I'm thinking I'd like to finish the first volume by August 15th and then maybe push to read the entirety of the second volume for the Tadoku half-round!  Though maybe half of the next volume would be a more realistic goal.)

One point to note about all the things I'm reading is that I've finally let go of my compulsive need to finish one book before starting another.  This is another suggestion Rachel made a long time ago, and it really is freeing to be able to pick up whatever I happen to be in the mood for at the moment!


I'm not including here the myriad shows I'm still watching subbed.  Honestly, although the goal is to drop the subtitles at some point, I don't think that's going to happen for a while yet.


This has been on my list of series to check out for so long that I don't even remember where I first heard of it.  So far as I can tell it isn't available online anywhere, but there are used DVDs available on Amazon Japan for not too much money (as these things go).  I'm glad I bought them, this is really fun!  The characters are likable and the series is generally easy-going and light-hearted, although with the occasional moment of some genuine poignancy.  The cast is amazing, including Aya Hirano, Mamoru Miyano and Jun Fukuyama.  All of whom have had innumerable notable roles, but I know them best as, respectively, the eponymous lead in "The Melancholy of Haruhi Suzumiya" and Konata Izumi from "Lucky Star," Rintarou Okabe from "Steins; Gate," and  Togashi Yuuta from "Love, Chuunibyou & Other Delusions."  The last is particularly amusing, since the character he plays, Dian:

is pretty much what you'd come up with if asked to draw a Chuunibyou Neko (though he actually is a dark wizard, so I guess it's OK).

My understanding is that fans generally consider the second series in the franchise, ジュエルペット てぃんくる, to be the best.  I've already purchased the box set and will start it as soon as I finish the first series, probably in a couple of weeks.  Neither the first or second series have subtitles available, though some of the latter ones do.


みんな フェアリルって知ってる? この子たちが フェアリル。 みんなが知っている妖精みたいなものかな。 フェアリルは リトルフェアリルというところに暮らしているんだ。 もしキミがフェアリルに会いたいと思ったらフェアリルドアを探すといい。 ドアを開けるときはフェアリルマジックの呪文を唱えるんだ。 リルリルフェアリル!
I don't really know how to describe this.  I can say it's about fairies, and as it progresses it develops a rich mythology and becomes a lovely about growing up, but neither of those statements really capture the essence.  All I can say is that the world of this series is just deeply, amazingly... warm and bright and gentle and whimsical and almost transcendentally sweet.  I mean, just look at some of these screen captures:

Ponies!  And Alicorns at that!
I kind of love this series.  Japanese subtitles for all but the first two episodes are available at kitsunekko and while I haven't been carefully going over every episode with subs, occasionally after watching an episode I'll methodically go through a key scene or two.

There were a few consecutive weeks when #リルリルフェアリル was trending on twitter after the episode aired, so it seems it's picked up an adult fanbase, I'm hoping that leads to Blue Ray box sets, as it did for アイカツ.  For the time being the first episode is available from the series' site at TV Tokyo.


In advance of the premier of ラブライブ!サンシャイン!! (which has so far been everything I  hoped it would be!) a number of singles were released, with accompanying drama tracks.  Although I certainly don't understand them completely, or even mostly, I do understand enough to enjoy them.  My favorite is the drama on 意になりたいAQUARIUM, in which the members of Aqours split into three groups and each visits a different marine park in an effort to find a theme for their next PV.  The first year students go to the Numazu Deep Sea Aquarium (a real place!) and become a bit obsessed with Coelacanths.  An obsession which was apparently shared by the seiyuu!

Hm, I can't seem to embed the tweet.  Here's the link!

Err... okay, sorry, this went significantly longer than I'd planned.  And I haven't even mentioned とびだせ どうぶつの森, which I've been playing for about 15-20 minutes a night,  or the effort I've made to be more active on the KawaJapa forums, or various other things I'm watching!

So I'll summarize with this: I think I've finally reached a point where I can enjoy living on the other side of the wormhole, and there is so very much to enjoy there!  My comprehension is not 100%, or even especially close, but it can only increase from continued exposure.  Maybe at some point I'll find a way to accelerate that process, but I'm not going to stress about it much.

Thursday, May 26, 2016

An anniversary

Flipping though my offline diary this week I noticed that I just passed an anniversary.  I first started learning Japanese, with these cards, on May 5th... 2008.  Eight years ago.  It doesn't feel like I have a lot to show for that time.

Although as noted on AJATT it's not the years that matter, it's the seconds, and I can't in all honesty say I've really put in much time, or more significantly effort, during those eight years.  I've jokingly described myself as lazy, but I don't really think I am.  Rather, I think I've been making unwarranted assumptions about what the process "should" feel like based on others' experience learning Japanese and my own experiences learning other subjects.

For example, Cure Dolly's comments here:
A very important piece of advice that one can't really give is: "find something in Japanese that you adore so much that pushing your way through it at a snail's pace (at first) is a labor of pure love".
(I would also recommend reading her comments, and indeed the whole discussion, on the "Making Learning Japanese Fun" thread over at koohii).  This is reminiscent of something Khatzumoto said here:
Because you work so hard on the material, you don’t get to sit back and relax with it as much. This does take out some of the enjoyment. Not a lot, just a tiny deduction because you’re exerting yourself so much. [...]

This is not at all to say that work in the beginning sucks; it’s fun; it is fun.
These parallel my experience learning how to program, for example.  The process wasn't easy, but it was fun.  I've tried to replicate that experience with Japanese, buying manga and light novels that were adapted into series I loved, applying Cure Dolly's method to アイカツ and so on.  In every instance the grind of the process and frustration over how much I didn't understand quickly overwhelmed any enjoyment of the material.

Which is where my decision to return to Anki came in.  I found intensive activities so difficult and unpleasant that I figured I wouldn't be able to get enough input to remember new words.

Speaking of Anki, here's what the latest month of my 1000-word vocabulary-only deck looks at the moment.

Still pretty dismal.  However, compare it to the previous report.  The retention rates are still low, but they've gone up.  Likewise the number of daily reviews is very very slowly decreasing, and the fraction of the deck that is mature is very very slowly climbing.  These results are interesting because this is the first time I've kept a "failed" deck.  Previously when my retention rates were this low I deleted the deck or gave up on Anki entirely.  However, this tells me that had I just kept chipping away at those decks the effort would eventually have started to pay off.  Looking back now, no doubt a major reason I haven't made much progress over the last eight years is that I kept restarting more or less from the beginning.

This also changes the way I think about Anki, from "maybe it just doesn't work for me" to "it works but not well, how can I make it work better?"  One answer to which is possibly mnemonics.  I've tried mnemonics before but not in conjunction with Anki, and I was inspired by this post from Cure Dolly to try the combination.  It's a little early to draw definitive conclusions, but here's what my trial deck with 100 words looks like:

So... worse.  Though really I should be comparing this to the deck without mnemonics when it was the same age as the new deck.  My impression is that mnemonics help early on and they help words become mature faster, but by the time the intervals stretch into a few weeks I no longer even remember the mnemonics.  Though the more I think about it "mature" probably shouldn't be taken to mean that I'll remember a card for a month, it should be taken to mean that I'll remember it long enough to see it in context.

If I were getting enough immersion.

Which I'm not.

Which brings this post full circle.

Just as I ultimately had to admit that my memory is poor and I need to force myself to use Anki even if I don't enjoy it, I think I now need to admit that I won't enjoy immersion very much until I reach a much higher level of proficiency, and I won't reach a higher level of proficiency without a lot immersion, and the only way to break the catch-22 is to force myself to immerse whether I enjoy it or not.  Here I mean not only intensive activities, but extensive as well.  The answer to "ugh, do I really have to watch another episode of anime tonight?" should, in fact, have been "yes, yes you do."  Or, well, rather it should have been "no, but only if you switch to a drama or travel show or something, or some manga, or a game, most of which are likely to be even more effort."

In parallel, I'm frankly weary of making such slow progress on new vocabulary.  Five new words per day has been my standard, I'm bumping that to ten and might reach for more in the near future.  I'll put on the brakes if (or, more likely, when) I start hitting two hours of reviews per day, until that point I can just focus on the fact that this seems to be what I need to do if I am to ever get to where I want to be.

Maybe this state of affairs won't last very long.  Maybe I just need to get past the plateau I've been stuck on and I'll find myself in a positive feedback loop.  Maybe that could even happen in as little as a few months.

Sunday, February 21, 2016

Counting grains of sand, ignoring the beach

I'll start with my recent Anki statistics, not because they're particularly interesting but because they illustrate a point.  As a recap, the plan this time around was to add 15 cards per day, review them very quickly and without much effort, and rely on repeated exposure to make them eventually stick.  Here's what the deck looks like:

Nothing particularly noteworthy, although that average ease is a bit low.  Which leads to the retention rates:

As expected, pretty dismal.  However, compare these to the results from the last time I tried using Anki, around this time last year.  Neither set could be called "good," but last year I worked much harder and had a lot less to show for it, so in that sense something has improved.

Which is not to say this latest attempt at Anki went well.  It didn't.  Once I passed 200 reviews per day things fell apart.  It's like a conveyor belt carrying boxes through a gate, there's no problem if the gate is wide or if there are only a few boxes.  However, trying to send through too many boxes at once will cause a jam at the gate, and after that the situation can only keep getting worse as more and more boxes pile up.  The gate of my memory is quite narrow, and I hit a point where no new words were making it though at all.  At least this time, rather than giving up and deleting the deck, I just stopped adding new words.  Since then the backlog has been draining, although slowly.

Around the time I stopped adding new words my thoughts turned to passive immersion, and the realization that I'm really not getting enough.  I decided to aim for a total of four hours per day, including both listening and watching.  Here's how that went:

There were a few days I missed the target (for good reasons) but mostly I managed it.  Which is not to say that this went well.  It didn't.  Typically by the end of the night I'd have a headache.  Even worse I often found myself watching something I'd enjoyed previously and thinking "is this almost over yet?"  Or even as remarkable as it may sound "ugh, do I really have to watch another episode of anime tonight?"  When I stepped back and realized that that's what I was thinking I put the brakes on immediately.

There's a lot in common between these recent experiences, and I think they point squarely to some common problems I've had all along.  The first is a rather obvious error in retrospect.  I can walk a mile in 15 minutes but that certainly doesn't mean I can walk eight miles in two hours!  After the first mile I'd be tired, after the third I'd be exhausted.  Likewise, just because I can handle 15 minutes of Anki reviews without difficulty that doesn't mean I can handle over an hour, and just because I can enjoy an hour of unsubbed anime (at least to some degree) that doesn't mean I'll enjoy three consecutive hours.

More fundamentally, I think even setting specific daily goals was misguided.  I tend to get obsessed with the minutia and lose sight of the bigger picture, and end up doing things like, well, turning watching anime into a chore, or neglecting the stories I'm reading in favor of just trawling through the text for enough words to meet the day's quota.  As the writers on Kawaii Japanese might put it, I haven't actually been using Japanese I've just been reducing Japanese media to a collection of study exercises, as dry and painful as any textbook.

However, my immersion experiment did yield one positive result, it kept me from wasting a lot of time on things like flopping around the internet, activities that are easy to fall into but not really enjoyable and certainly not satisfying.  Although I didn't much enjoy the time, it was noteworthy how much of it there was available for Japanese once I forced the distractions out.  So it seems the sensible thing to do is flip things around; cut out the distractions to make space for Japanese and allow, rather than force, that space to fill with things that I enjoy.

So the plan going forward is simple, outside of work I give myself complete freedom to do anything I want, so long as it isn't in English.  Read what and when I want, watch what and when I want, look up words and intensively analyze sentences not out of a sense of obligation but rather to help me enjoy what I'm doing even more, and add to Anki (or perhaps Skritter) those words that, for whatever reason, I particularly want to hold onto.  And if I need some quiet time to catch my breath and want to just stare at the wall for a while?  That's fine too.

And, really, this isn't so much a plan as it is just living the way I'd eventually intended to.  When I chose the name of this blog one image I had in mind was that my apartment would actually, physically be in Akihabara and my front door would be a wormhole linking that space to the US, where I work and where my friends and family live.  In practice, although I've done a good job of removing all English material from my shelves and walls, too much of my time at home has been spent connecting back to the English-speaking world through the internet.  Time to start actually living in my Japanese apartment!

Wednesday, January 6, 2016

Envisioning the future


While it is common at this time to look back on the previous year, I'd instead like to look forward, to an evening in 2016 (or, perhaps, 2017):
I return home from work, as always a little tired.  I put solive24 on in the background while having dinner.  Even though I'm not paying any particular attention, it does seem that recently I've been I've been picking out words a bit more often.  After dinner I do a burst of Anki reviews.  I got through a good chunk during the day, but there's probably still an hour's worth to do this evening.  After 20 minutes I step away from the computer, sit down by the coffee table, pick up the Kobo eReader and return to the novel I'm reading.  There are still a lot of words I don't know, but I have a firm grasp of what's going on in each scene.  This volume is well-past what was covered in the anime adaptation, and the fact that I can spend more time with these characters is still something of a thrill.  I won't look up unknown words as I encounter them, that breaks up the flow too much, but I will highlight them.  This weekend I'll sync the desktop reader, go through all the highlighted words and add them to Anki.  Speaking of which, after reaching the end of the chapter I return to today's reviews and push through most of them.  Then time to relax with an episode of anime!  I'm still watching new series with subtitles, but weekdays are for rewatching favorites raw.  While watching I hear a few words that I added to Anki recently, that's always a nice feeling.  After that I finish up today's reviews, then start getting ready for bed.  Before going to sleep I casually read through a chapter of manga.  This is a purely extensive activity, when the text gets too dense to follow I'll just look at the pictures for a while. Though that also seems to be happening less often recently...
So, there's a look into the future.  Or a possible one, anyway.  Some might quote Yoda's "always in motion is the future," but I think I'll go with "未来はいつもミステリー/それは Sweetest Sweetest ラズベリー"

Saturday, December 12, 2015

An unexpected decision

I thought I was done, moved on never to return, but... shortly before Thanksgiving I created a new vocabulary deck in Anki.

The impetus for this was finishing the iKnow Core2K:

So, you know, グッジョブ 俺.
On reaching this landmark I realized a few things.   I started iKnow as a safety net, a way to ensure that I was making some kind of progress while freeing myself to to pursue long-term strategies like intensive reading.  However what I found is that, at least right now, drilling with an SRS is a lot easier than trying to engage with native material in any intensive way.  Intensive activities require a level of energy and focus that I just don't have after a day of work.  Or maybe I'm just lazy, but if that's the case I should be honest with myself about being lazy and plan accordingly!

That seems to work out OK for Anzu
I also realized that the vocabulary in iKnow really isn't great.  A large portion is comprised of words I already know, English loan words, and simple variations of other words (once I know 高い, 長い  and 高さ I don't really need 長さ as a separate vocabulary item).  Genuinely new words weren't really sticking any faster or easier than they did with Anki.  For a while it seemed like they were and I attributed that to the multiple different kinds of "cards," but on reflection I think iKnow just seemed easier because so many of the words were, for various reasons, rather easy.

This all culminated in my next realization, which is significant enough that I'm going to separate it out and bold it
I have a poor memory for vocabulary.

I should have just admitted this a long time ago.  If I had just acknowledged my own limitations instead of blaming Anki for my frustrations I could have saved myself a lot of time and effort looking for something better.

So if no method is going to work particularly well, and trying to learn directly from native material is more effort than I can (or will?) exert, and SRSs with premade vocabulary lists are unsatisfying... then what's left?  Right.  Anki.  Again.

However, the one important thing I did get from iKnow is a sense of what using a spaced repetition system should feel like.  I'm having a difficult time putting this into words, but iKnow somehow makes it easier to adopt a relaxed, laissez faire attitude.  When using Anki I would obsess over forgotten words which lead to both mental strain and a lot of time spent repeatedly looping over each day's forgotten words.  iKnow somehow makes it easier to just shrug things off and allow words not to stick today because, well, maybe they will tomorrow.

So this time around when I see a word I don't remember I'll fail it and when it comes around again I'll read it and mark it "good", regardless of whether I remembered it, and that will be the last I do with for the day.  This makes Anki far more pleasant and much faster.  Come to think of it, this isn't really that different from the attitude Tae Kim espouses for reading.  See a word, look it up, it either sticks this time or it doesn't and there will be another chance later.  Except "looking it up" just means flipping the card.

So far this has been going fairly well.  Adding new cards is easy, I just turn on Rikaisama and drag the pointer over text I'm interested in, hitting S every time a word pops up that I don't know.  I'm not making any particular attempt to read for content, I'm just trawling for words, though of course if I do happen to understand a sentence or clause that's a nice bonus.  I've done this with my twitter feed, the Kobo edition of 魔女の宅急便, and anime transcripts.  The latter works nicely immediately after watching the corresponding episode.

The daily review counts are increasing pretty rapidly, but the reviews go quickly so it isn't taking too long in total.  Also, Anki reviews are the one Japanese-related activity I can actually do during the day.  Ten minutes or so on the bus to and from work, a couple of minutes waiting on line for lunch, a word or two while waiting for a program to finish running.  The unit of work in Anki is so small that there turn out to be a lot of little gaps in the day where I can do some, and these add up.

But... I need to be very clear with myself about where this path leads.  It is the nature of SRSs that so long as new words are added the daily number of reviews will increase.  It will do so slowly for someone with a high retention rate... but my retention rates have been, are, and will continue to be quite low.  It's a matter of when, not if, the daily reviews take an hour per day to complete, and then two hours, and then... If it ever becomes more than I can handle I can stop adding words for a while, but of course there's a cost to that too.

I also need to be very clear about where this path doesn't lead.  This is not learning Japanese.  This is mapping a lot of Japanese words to rough English approximations.  When I return to learning from native material and it is certainly my intent to do so I will still be faced with many sentences that I won't understand despite knowing all the component parts.  However, I think that right now native material is so much effort for me because I have to deal with both the meanings of individual words and the meaning of the whole.  My hope is that after building up a decent vocabulary it will be easier to grapple with meaning at the level of sentences.  In a sense my plan is to do half as much work for twice as long.

I don't plan on focusing on Anki forever.  I might not even need to focus on it very long in the grand scheme of things.  I've been stuck on this plateau for almost three years now, I think maybe all I need in order to start making progress up the mountain again is to, say, double my current vocabulary from about 2500 words to 5000.  At 15 new words a day, if I can maintain that pace, that's only about 6 months.  After that maybe I can try learning from intensive reading again and/or try going to monolingual dictionaries again.

It's not like I need to cut off all connection with native material until then!  In fact I can concentrate on more fun, extensive activities.  I'm still rewatching series raw that I've previously watched subbed and I'm still listening to audio dramas and music.  I've also started approaching manga in a new way.  I'll look over the words but without any real attempt to read anything, instead I'll mostly just look at the pictures.  It's kind of odd that despite manga being a visual medium I've been so wrapped up in trying to understand the text that I tended to ignore the art!

Monday, November 2, 2015

Checking in


There are two reasons for the recent lack of entries.  First, my job basically ate my life.  Long days, and frenetic ones, so that even on those occasions when I got home at a reasonable time I didn't have the energy to do much, if anything, in Japanese.  Or in English, come to that.  I think things are finally starting to calm down a bit, or at least I'd like to believe that.

The second reason for the lack of updates is that there hasn't actually been much to say.  I've been keeping up with iKnow, each day spending about a half hour on reviews and adding 15 new cards.  Here's where I am:

I also started using Skritter a few weeks back, more as a way to practice writing than as a way to drill vocabulary.

The original impetus was to facilitate dictionary lookups.  There are many amazing programs and tools that recognize handwritten kanji, but they all frequently required me to make several attempts before finding the character I wanted.  I thought that getting a better feel for stroke order would help and I think it has, but in the process I also discovered that I just like writing kanji.

While this level of studying is going fairly well I'm still struggling to figure out a strategy for immersion that I'm happy with.  There are a few activities I genuinely enjoy and that might even be helping me to learn.  Music and audio dramas certainly, although recently work has been such that I haven't been able to have headphones on at all during the day.  I enjoy rewatching anime series raw that I've previously watched with subtitles, and I've found a small handful of dramas that I've enjoyed watching raw.  Though to be honest I have my doubts about how useful these activities are, it seems like when watching TV I just tune out the words entirely.  Perhaps because I can rely on the visuals to convey the story?  There is probably more I could be doing there, either extracting the audio and listening to it repeatedly and/or going back to Dolly's method.

In any case, where I'm still having real difficulty is reading.  Around the time of my previous entry I had been trying to read 喫茶部へようこそ intensively and I found the process of looking up nearly every word in every panel and then frequently not understanding the sentences anyway! to be so dreary and frustrating that it outweighed any enjoyment I might have gotten from the story.  I tried switching to easier material and I tried streamlining the lookup process by using material with furigana or reading online manga with the help of KanjiTomo (which, while not 100% accurate, is a pretty amazing piece of software).  At the moment I've settled on both simpler text and easier lookups, I'm reading 魔女の宅急便 via Kobo and Rikaisama... but somehow it still isn't quite working.  I'm following Tae Kim's advice:
[...] read as much as possible and look up each word you don’t know. You are essentially simulating how your mind would have worked if you had known all the words with a couple seconds lag for each word lookup. Finally, read the sentence again with all the words in your short-term memory to reinforce the process you want to attain.
but too often, most of the time in fact, I'll really have no idea what a sentence or clause means even after looking up all of its component parts.  And then the frustration and irritation sets in.  Generally speaking a few minutes per night of this is about all I can stand, although admittedly being tired and stressed from work wasn't helping.

At some level I know this is part of the process and that there's no way past this stage except to keep going.  Presumably every once in while I'll come across a sentence that "clicks" and helps me make new connections, indeed this has actually happened at least once already.  The strange thing is that I read a whole book extensively and understood basically none of it, and I didn't find that frustrating.  Somehow by making an effort to understand I'm putting artificial pressure on myself and creating the very frustration I'm trying to avoid!  Probably all I need to do is, you know, stop doing that to myself.  Come to think of it, didn't someone once advise me to "worry less, read more"?  Oh, right, that was me!