Given this is the forum of a SRS based learning tool, I get that all the new SRS apps are what quickly draw eyes towards when posted here; and well… I think SRS are like plants … and you have to water them every day … and most of my plants end up drying and withering during this summer months , so why not read?? having a refreshing drink perhaps
Having that into consideration, I would like to post here a great video featuring Paul Nation talking about extensive reading, how can that do wonders to fix and expand your vocab gradually.
He talks about how Graded Readers work towards easing the steep road towards native material, and how unadapted native material can be counter productive at an early stage… but then further graded material it’s not really necessary… though there’re new ways been developed to fill the gap between actual adult novels and the Graded Readers last levels.
I have one really big beef with it. In the first half of the video, all his calculations about the reading time needed every day to learn a certain number of words in a year are based on one assumption : The reading speed of the learner is about 150 words per minute
That seems crazy freaking high to me. 150 words per minute ? For learners ? In my case, most of the time when reading japanese, my word per minute is dramatically lower than that, and actually drop to zero often, because I stop to check dictionary or some grammar points or some unknown kanji, or just trying to make sense of complex sentences…
More realistically, I think the reading speed is a curve, starting from almost zero for complete beginners to the 150wpm he use, maybe for high intermediate/advance learners ? But I strongly suspect that introducing any kind of variable realistic reading speed would make his results of pure extensive reading for vocabulary much much worse…
In the second part of the video, I really like his points about graded readers. I often wonder why so few graded readers exist for japanese (or any language). The method he describe to create them seems rather straightforward. Take books out of copyright and produce multiple versions adapted to 1000 words vocabulary / 2000 words / 3000 words etc. It should be probably fairly easy nowadays to run some algorithm to highlight words out of the vocabulary scope and replace them by simpler synonym. At least it’s probably much less work compared to the massive task of creating stories from scratch ?
Maybe there aren’t so many compared to other languages, but I don’t think I could afford to buy all the graded reader options that currently exist
But also, I don’t think that approach would work very well, unless you were working on the assumption that your reader has extensive grammatical knowledge but a poor vocabulary. Graded readers are designed to slowly ramp up the grammatical difficulty as well as the breadth and complexity of the vocabulary.
I doubt you’d be able to replace words simply by an algorithm either, because it would probably mis-parse a lot and the synonyms may not always be appropriate. You’d still have to go through it by hand, I imagine, at which point creating a story from scratch, especially if you’re basing it on a folk tale or similar, may start to look a bit more appealing.
Well…I mean, graded readers can be helpful. But the main thing about extensive reading - and he’s right, extensive reading is totally an excellent way forward - is that it has to be fun. Now, I prefer Naruto to most graded readers I’ve looked at, so. That’s what I’m reading. I had to work up to it - started with Yotsubato and didn’t understand all of that, either, but you don’t need to understand 100% to enjoy something
I’m not an AJATTer but that’s something he got right - get books, apply eyes, read however. Read closely the bits you like and skim the bits you don’t, whatever, get words in your eyes. I read Yotsuba quickly and a few chapters of Fairy Tail really slowly and bits of No Game No Life and Harry Potter at varying paces, and they all helped me get to where I can read Naruto at a comfortable pace of a chapter a day and look up, like, maybe one word per page. Sometimes I can’t be bothered so I don’t, and only start again if I start losing track of the story.
I don’t know, Maybe this is the best way, but I still think the best way is the one you’ll actually do. Oh, and that we should all read more probably, and maybe also a bit faster.
If you get the number of vocabulary words low enough, then speed really does get fast quickly.
In my opinion, 1000 words is too many. When I teach Spanish, I start with readers that use only about 50 words—but that use them many, many times to create a (decent but not great) story. Almost all my students can read them, and after only a few hours of class, some of them are actually reading fairly fluently, without translating. Although they don’t learn a lot of words that way, they learn those basic words well, which means that when they get to more advanced readers (with 100 or 150 words!), those easy words just fly by, and they can focus on the more difficult stuff. They also absorb the easier bits of grammar.
Reading this very, very easy material, even if you can read more difficult material, builds speed.
An algorithm couldn’t rewrite the stories, but it could point out the difficult words that need to be replaced. And I can tell you from experience that yes, that can be much easier than writing things by hand—although you’re right, if you’re retelling folk tales, that’s not too hard, either. But writing stories from scratch isn’t for everyone. (It’s currently my goal, and it’s not going very well!)
Edited to add:
What I find lacking in most graded readers is volume. In my early reading experiences, I want lots and lots of repetition of the vocabulary words, in a meaningful context. (It doesn’t have to be a fabulously interesting context, but the context has to be there.)
Many graded readers either are full of text but advance quite quickly, or else are full of pictures and have too little text.
I just used this website to test my reading speed in English. I choose a level 8, level 11, and level 13 story. The level 8 and 13 stories (which I got 100% on comprehension) I had a reading speed between 180 and 200 words per minute. The level 11 story I had a reading speed of 230 words per minute, but I only got 50% of the comprehension questions right.
I know I’m a slow reader, but if my reading speed in my native language is barely over 150 words per minute, there’s no way it would be that for Japanese, even with easier texts.
@Arzar33 Maybe word counting would be different for japanese (150 words per minute seems high for me as well, perhaps japanese with the helper and compound verbs make a difference in that regard )
@LaughingLiving The vocab word count in the tadoku series gooes like this.
I thinks the 350 word count and the 400 characters per story it’s an OK entry point. As a reference the only series with a level 5 (that I’m aware) goes up to 2000 words.
You mean like more characters per story??
I find that the real art of making this books it’s exactly put in evidence in those first volumes. The Ask series really surprised me there, as with so little vocab I was totally blown away to get the meaning of an actual story ( I was solely reading my Genki I sample sentences by then).
About the repetition, it’s there, even more, it’s present across stories, for both vocab and grammar, which it’s great since gradually the struggling and sense of accomplishment grows
@Vorvayne One of the points I think it’s missed in the video are the alternatives to Graded Readers for L2 learners / Novels. Lots of material (manga included) goes into a place that could easily overlap graded readers in terms of vocab… for my it’s graded readers but for japanese children (this cover 1st until 6th grade) … the more mature the target audience the broader the spectrum of choices for the readings I can get.
I think the initial levels of graded readers for japanese learners are hard to beat in terms of the design (minimum grammar, controlled vocab count, etc), as they are aproachable while the most common recommendation in native material still its too hard… but as you get into the last levels of these series they become another tool in a wider spectrum of possibilities.
Sorry, I worded badly, yes of course I agree, trying to replace automatically word by algorithm is a recipe for disaster. What I meant is highlighting difficult word by algorithm then replace them by hand. Which is probably not that difficult job for a native speaker, except when the vocabulary allowed is very low, then a rewrite is going to get more and more involved and at some point, probably, creating story from scratch is easier.
@Ncastaneda
But the video is fairly convincing that between end of graded readers and native material there is a hug gap that is not covered. What is your take on his message around 25 minutes mark ? (He is talking about reading the book Captain Blood, around 100000 words)
“If you are a learner with vocabulary of 5000 thousand words and you picked up the book there would be one unknown word every 22 words, one unknown words every 2 lines, and there would almost 2000 unknown words in the book for you. So, reading unsimplified text is a cruel and unusual punishment. For the American (teachers) in the audience, you are immediately prohibited from using unsimplified texts for learners who have a vocabulary size below about 8000 or 9000 words.”
Yep, I really was expecting better advice for filling that gap too (for it’s where I’m now ) , besides his adapted version of copyrights free novels.
My take is using the graded native material, of which japanese it’s full.
I’m tackling the 10分で読める伝記 series now (there’re lots of series aimed for kids in the 1-6th grade). After that point I’m considering some books published by 岩波書店. They have a 少年 series which starts at 小学６年 and goes straight to 中学 aimed novels.
I’m counting on this to fill the gap… for now I know there’s enough material to read with a suitable curve for improvement … besides as I go further I find myself faced with more and better material to read, and the prices are much more affordable than with Graded Readers for japanese learners (which for me it’s the only drawback they have)
I only watched the first 15 minutes, so correct me if I’m wrong, but my hot take on this is that extensive reading isn’t enough as you get out of beginner territory.
While he was laying out his assumptions, he mentioned that some research found that it took an average of seven repetitions in a text before the amount of time spent fixating on a new word evened out with the words you were already comfortable with.
This was fine because he was talking about learning the first 2000 most common words through extensive reading with lower-level books. Due to their limited vocabulary, those books might actually have the required amount of repetition.
What I’ve found parsing books on Floflo though is that intermediate (and even some beginner books) are never gonna give you even remotely close to seven repetitions on any slightly uncommon word. In fact, around 50% of words only appear one time in a book. The amount of words appearing more than three times is minuscule, relatively.
So although I think reading is definitely useful, I don’t think that it’s sufficient for getting into the higher levels of reading fluency. You’d really have to blast through novel after novel to get the number of repetitions that he suggests, and even then those repetitions might be stretched out too much to be sufficient. It feels to me like SRS-supplementation is literally a requirement.
Which is why you constantly look for more challenging books in different genres. Yeah you might see words like 端正 only once in a book aimed at younger audiences, but you might see it a few times not too far apart in a book aimed at adults. Similarly, you might see 成仏 once in a metaphorical sense in that same adult book, but see it a ton in the younger book if it is about ghosts and whatnot. The point of each book is to familiarize yourself with words common in that book, not every word that might appear in it.
Edit: Not saying that SRS can’t be beneficial even as a more advanced student. Just arguing that it isn’t a requirement.
To my understanding, it seems like you’re more into reading for enjoyment while I’m more interested in translation and writing. I feel like the latter requires a higher standard of accuracy and because of that I feel like learning just the more common words isn’t enough. That’s what I meant
I think to remember that he initially said 12 repetitions where believed to be the sweet spot (over the course of a year), but then actual studies revealed that 7-8 repetitions achieved the goal, as measured by eye fixation on words while reading…
He later mention, that the problem could arise if you get that repetition in the first books only… and then don’t see the word again
So I guess 7-8 repetitions on a 1 year interval sounds right … that amount of repetitions gets harder to get once you are pass the 4-6K word count… so you’ll need to read more … Finally he tells that extensive reading could be well suited to be used with some other resources (as after some extra thousands words… you would have to read like a lot to keep extensive reading your only resource for vocab improvement)…
so all in all I guess I’m slowly trying to find a way out of SRS systems (supported by state of the art research … of course…)… I mean… I used to have a life… …
I made it through the entire video! (watching videos at x1.5 or x2 in places can make all the difference!)
My notes and thoughts:
*150 words a minute for a beginner?!? He really did blow off the question about slow readers. His suggestion is a speed reading course. (and that a speed reading course should get an ESL student up to 200-250 wpm!) I have no idea where you would find one for Japanese. I suppose you could do a lot of timed JLPT reading section tests? Of course with Japanese, you would measure characters per minute. I have no idea what would be considered a reasonable range.
*He is for looking up words (deliberate learning) when you don’t understand them. I completely agree with this and will look up words in graded readers if I don’t understand them.
*He says that you do need to use other methods of study, including rote memorization.
*I’m curious how many series of graded readers exist for Japanese. I’m aware of two. The most commonly suggested Japanese Graded Readers, and also the green series of readers. There are also the digital graded readers from White Rabbit Press that are being rereleased. If there are more, I’d like to know and compile a list.
*Japanese graded reader books don’t go nearly as high as the top 9,000 for English, but I know some websites are doing great calculating word counts for books. The only common lists I’m aware of for Japanese are Core 2/6/10k, and JLPT vocab levels. I would love to see vocabulary frequency data on popular manga and books. Unfortunately, I don’t have the skills to gather or compile the data.
I completely agree. I believe that they are mostly statics used to say how large someone’s vocabulary is as a way to easily classify the reader’s level. I know the Japanese Graded readers level 0-1 is 350 vocabulary words, but they do not use most of those words in each story. If I remember correctly, they do start off focusing on about 50 words with 400 characters a story.
*I remember reading some articles and examples about comprehension. Knowing 90-95% of the vocabulary makes such a difference when reading. With that level of comprehension, it’s much easier to decipher the meanings of unknown words. I know when my vocabulary knowledge is lower, I’m glued to a dictionary.
)
… the more mature the target audience the broader the spectrum of choices for the readings I can get.
… besides as I go further I find myself faced with more and better material to read, and the prices are much more affordable than with Graded Readers for japanese learners (which for me it’s the only drawback they have) 
… of course…)… I mean… I used to have a life… 