Is there any Japanese vocabulary test like those English ones?

Personally, I'm starting to notice the limits of studying with subs2srs. It's been great and will be for a while longer, but where I was once getting 40-60 cards per episode I'm now lucky to get 20. With children's shows 10 or less. I imagine that at some point the time it takes to make a card will outweigh the benefits and that's when I'll stop using Anki for vocab.

dizmox Wrote:

PotbellyPig Wrote:

dizmox Wrote:I probably have been through around 28k cards myself, started reading a bit back around the 10k mark. I can't remember words very well just by osmosis since I'm not a heavy reader. Native level vocabulary more or less.... always trying to expand active repertoire though, which is what really matters later on.

I'm around 15,000 words now and would like to get to your level. It's amazing how many words you really need to know. Right now reading light novels, I can encounter on average 1-2 unknown words per page over the course of the book. Hopefully when I get to your level it will be less than 1 word per page on average.

Yeah, depending on the book it's usually around a new word every 2-10 pages for me now, so not too bad.

People may say stuff like you only need a few thousand for every day conversation, but the reality is you end up needing "specialized" words and phrases quite often for everything but the most banal of conversations.

Thanks for the info. Assuming I add enough cards to my deck by reading, I should get over the 20k mark of cards studied this year. One new word every 2-10 pages is encouraging. The 28k mark sounds like something to strive for. I was thinking of aiming for 25k-30k words in the first place so this just reinforces that notion.

sunehiro Wrote:I review cards like アルバム to improve my pronunciation of Katakana words, which is often "unnatural" and quite hard. (also considering the fact that I am not a native English speaker)

Couldn't agree more. Apart from the fact that I often have to look up the word, because the Katakana just don't make sense to me. I'm like "Alubamu, what could this be?" and I just don't get it. Could of course be due to the fact, that I too am not a native speaker.

riogray Wrote:

sunehiro Wrote:I review cards like アルバム to improve my pronunciation of Katakana words, which is often "unnatural" and quite hard. (also considering the fact that I am not a native English speaker)

Couldn't agree more. Apart from the fact that I often have to look up the word, because the Katakana just don't make sense to me. I'm like "Alubamu, what could this be?" and I just don't get. Could of course be due to the fact, that I too am not a native speaker.

Even as a native English speaker I feel it's important to put just as much effort into learning katakana-fied English words as any other. The pronunciation is often not what you might expect (ジレンマ comes to mind), and just guessing whether it's 頭高型、中高型、尾高型 or 平板型 will often lead to mistakes. Also, just because you read a katakana word once and know that Japanese people use it doesn't mean you're forever gonna remember that it's a word the Japanese use. I'm sure I've unnecessarily used the Japanese version of words when the katakana English word would have been more appropriate, simply because I forgot or didn't know it was an adopted English word. Then there are 和製英語 like カンニング that mean something different in Japanese than they do in English.

comeauch Wrote:

dtcamero Wrote:if you're able to learn from context, I think the evidence suggests that anki is relatively inefficient at high vocab levels

Which evidence?! Sorry but that's kind of BS lol. I haven't seen any paper specifically about the inefficiency of Anki at higher vocabulary levels for learners "able to learn from context".

Everyone is able to learn from context and memorization is memorization. You probably just mean that there's no point in memorizing a lots of highly specialized words, which is nothing more than your own opinion. If someone wants to memorize those words, there's no reason why Anki would be less useful than before.

But I agree reading is similar to a SRS. The most common words/grammar gets used a lot, so maybe not as "efficient", but so much more reinforcing and enjoyable!

I don't think that Anki is ever completely useless, but it does lose some efficiency as you learn more words; not because Anki itself becomes less effective than constant reading, but that the words you're studying aren't particularly common, so you get less reward for the hours you put into studying it in an SRS.
Sure, if you still have difficulty picking up words without studying them and you want to learn the word, you might as well go ahead and stick it in Anki, but it has less benefit than it did with the first few thousand words.

Also, it's not necessarily specialized vocabulary that you don't want to study with a(n?) SRS, since reading specialized papers puts many otherwise uncommon words into regular use. You have a completely different set of common vocabulary while talking about biology or physics, so you would get just as much benefit from learning those words in Anki as you would get from Core 2k, so long as you actually like reading those kinds of things (which I do).

Stansfield123 Wrote:Another point: Core6K, I believe, has a crapload of English loan-words in it, or simple Japanese words that you hear all the time. Stuff like アルバム, スカート, メニュー, etc. So, presumably, people who say that they have the 6000 Core6K words in their Anki deck, have reviewed the word アルバム a good 10 times. Why? Why would someone do that? What possible purpose does that serve, except bragging rights?

That's the beauty of SRS. Words that are very easy will appear less frequently because you marked them as "Easy" when they appeared. And yes, it's possible to sometimes forget even the simplest words. Maybe not これ or 何, but the katakana ones, definitely.

comeauch Wrote:

dtcamero Wrote:if you're able to learn from context, I think the evidence suggests that anki is relatively inefficient at high vocab levels

Which evidence?! Sorry but that's kind of BS lol. I haven't seen any paper specifically about the inefficiency of Anki at higher vocabulary levels for learners "able to learn from context".

well if you read my previous post I mentioned the video and thread, but if necessary I can summarize the linguist and mathematician Anthony Lauder's speech on the seminal work "Learning Vocabulary in Another Language" by I.S.P. Nation. This book is the summary of 600 research studies across 30 years in applied linguistics.

and I did mention the assumption that in order to learn efficiently via reading, one must be able to "learn in context... i.e. noticing, processing and guessing", to use anthony's words.

the point is that anki is a good way to memorize a lot of words, with huge breadth but no depth (contextual meaning... a multifaceted, multivalent concept of a word). Anki is a good starting point for memorization, but only gives a very shallow understanding of any of these words... somewhat similar to what RTK gives you for kanji.

beyond a threshold of efficiency, word frequency is too rare to make intensive study worth the effort. more productive study would be getting exposed to the words naturally. otherwise you learn words in one context only, without vocabulary control... without an understanding of those words usage, particular meaning or context. similarly most japanese students stop anki'ing kanji after a certain threshold of efficiency (2000 or 3000), and instead proceed learning advanced kanji via reading in context.

Obviously reading in a second language, without knowing the meaning of all the words you come across, is difficult... as a result people often shun it and other practical tasks (such as speaking) because they are a painful reminder of things unknown.

But while reading, and EDUCATED GUESSING the meanings, and continuing on, and then hitting those words again later due to a large consumption of material, one ends up at a longer lasting and deeper meaning because you have thought it through.

i.e. repeated encounters in different contexts broaden and deepen the meaning of these words.

your point about anki becoming less efficient at higher vocab levels is a misunderstanding of what I wrote. It isn't that that practice is less efficient at memorizing those words, but rather that as those words pop up less and less often, due to low frequency, the practical use of that practice becomes lower and lower relative to the cumulative language-related benefits from reading.

Aikynaro Wrote:Is there any Japanese vocabulary test like those English ones?

Personally, I'm starting to notice the limits of studying with subs2srs. It's been great and will be for a while longer, but where I was once getting 40-60 cards per episode I'm now lucky to get 20. With children's shows 10 or less. I imagine that at some point the time it takes to make a card will outweigh the benefits and that's when I'll stop using Anki for vocab.

http://www.kecl.ntt.co.jp/icl/lirg/resou...-test.html

Aspiring Wrote:

Aikynaro Wrote:Is there any Japanese vocabulary test like those English ones?

http://www.kecl.ntt.co.jp/icl/lirg/resou...-test.html

Quote:小学生レベル： ５千～２万語
中学生レベル： ２万～４万語
高校生レベル： ４万～４万５千語
大学生レベル： ４万５千～５万語

I got 16,100. JPLT N1 is roughly 10,000. It's daunting to know that the university student level is 4.5 times that.

I get 4.9万 but I'm barely at 3万 really. Don't take it too seriously. I don't think there is a 2万 gap between me and the university students I know.

vileru Wrote:

Aspiring Wrote:

Aikynaro Wrote:Is there any Japanese vocabulary test like those English ones?

http://www.kecl.ntt.co.jp/icl/lirg/resou...-test.html

Quote:小学生レベル: 5千~2万語
中学生レベル: 2万~4万語
高校生レベル: 4万~4万5千語
大学生レベル: 4万5千~5万語

I got 16,100. JPLT N1 is roughly 10,000. It's daunting to know that the university student level is 4.5 times that.

Good post for the discouragement thread.

That test, just like the prior one for English, is whack. I got 8k when I know between 4-5k.

Holy 45900. I don't think I actually know that many =.=

Anyways, imo anki is still just as effective as before at learning new words. However, the question you have to ask is that, is learning new words in anki worth your time, or is that time better spent reading and listening to solidify the words you already know. For "fluency" purposes maybe it's better to know a (nontrivial) subset really well than it is to know a lot of words really shallowly and without much ability to use it. Plus, if you know your kanji's, picking new words that are made up of kanji's you already know is not hard at all even without anki.

I know for me I need to be spending more time in native material and (as a result) less time in anki. That's another reason why I've paused adding new words. The problem, is, of course I spend too much time on the internet and not enough time in novels and anime

The link tests words in different brackets (frequency). 50 words is a small sample tho

OP; every word is a good word -- not as another marker for success, but as a uniquely pleasing symbol,

“For words are magical formulae. They leave finger marks behind on the brain, which in the twinkling of an eye become the footprints of history." –Franz Kafka, one of the most influential writers of the twentieth century

"Realize that words mean nothing on their own. They are simply vehicles for expressing ideas, feelings, and experiences." Sean Clouden

“The last motive in the world for acquiring vocabulary should be to impress. Words should be acquired because we urgently need them—to convey, to reach, to express something within us, and to understand others.” –Vanna Bonta, award-winning novelist

Aspiring Wrote:http://www.kecl.ntt.co.jp/icl/lirg/resou...-test.html

Yeah, but how about one that actually works? Try checking only the last word. Then try adding the first word.

Quote:Aspiring wrote:

http://www.kecl.ntt.co.jp/icl/lirg/resou...-test.html

Yeah, but how about one that actually works? Try checking only the last word. Then try adding the first word.

Hmm. Well, for reference I got the link from here. (Amount of Vocabulary, Koohii)
http://forum.koohii.com/showthread.php?tid=10309&page=3

Try checking a few in the middle, maybe ?

Just for fun I'll have my girlfriend do it later (fairly average college educated adult), I'd be interested in seeing what she gets compared to you guys picking up 40-50k words.

dtcamero Wrote:

comeauch Wrote:

dtcamero Wrote:if you're able to learn from context, I think the evidence suggests that anki is relatively inefficient at high vocab levels

Which evidence?! Sorry but that's kind of BS lol. I haven't seen any paper specifically about the inefficiency of Anki at higher vocabulary levels for learners "able to learn from context".

well if you read my previous post I mentioned the video and thread, but if necessary I can summarize the linguist and mathematician Anthony Lauder's speech on the seminal work "Learning Vocabulary in Another Language" by I.S.P. Nation. This book is the summary of 600 research studies across 30 years in applied linguistics.

and I did mention the assumption that in order to learn efficiently via reading, one must be able to "learn in context... i.e. noticing, processing and guessing", to use anthony's words.

the point is that anki is a good way to memorize a lot of words, with huge breadth but no depth (contextual meaning... a multifaceted, multivalent concept of a word). Anki is a good starting point for memorization, but only gives a very shallow understanding of any of these words... somewhat similar to what RTK gives you for kanji.

beyond a threshold of efficiency, word frequency is too rare to make intensive study worth the effort. more productive study would be getting exposed to the words naturally. otherwise you learn words in one context only, without vocabulary control... without an understanding of those words usage, particular meaning or context. similarly most japanese students stop anki'ing kanji after a certain threshold of efficiency (2000 or 3000), and instead proceed learning advanced kanji via reading in context.

Obviously reading in a second language, without knowing the meaning of all the words you come across, is difficult... as a result people often shun it and other practical tasks (such as speaking) because they are a painful reminder of things unknown.

But while reading, and EDUCATED GUESSING the meanings, and continuing on, and then hitting those words again later due to a large consumption of material, one ends up at a longer lasting and deeper meaning because you have thought it through.

i.e. repeated encounters in different contexts broaden and deepen the meaning of these words.

your point about anki becoming less efficient at higher vocab levels is a misunderstanding of what I wrote. It isn't that that practice is less efficient at memorizing those words, but rather that as those words pop up less and less often, due to low frequency, the practical use of that practice becomes lower and lower relative to the cumulative language-related benefits from reading.

I really like Anthony Lauder, but it's not a scientific study... And even if it were, I'm pretty sure I DO understand your and his point and I agree with most of it: if you're not going to use specialized vocabulary, then learning it as you go is the most efficient way to go (if you're "going"! aka reading, communicating with native speakers, etc.) It's the way you said "if you're able to learn from context, I think the evidence suggests that anki is relatively inefficient at high vocab levels" that made me jump. Who is NOT able to learn from context? Some are too scared to try when it comes to language learning, but we're all experts at guessing and making presumptions!

In the end Anki is a tool and it does its purpose equally well for any "level" of vocabulary. It's just probably more useful to become an expert of say ~10-20k words (and grammar! and style!) than memorizing 100k words in 50 years. But if you're a chemist working in Japan, there's no reason why repeating chemistry-related flashcards would not work as well as it did with basic vocab. So it's not less-efficient, it's just that there always need to be a head behind the tool.

So to get back to the original topic, I guess the ideal vocab goal - both in range and depth - would be the point where it becomes more time-efficient (and enjoyable) for you to use that language than srs it. I feel like I'm not very far from that turning point at just below 10k, but I consider myself quite fluent in German and it came way before that (due to German being so much more transparent to Latin languages speakers than Japanese)

afterglowefx Wrote:Just for fun I'll have my girlfriend do it later (fairly average college educated adult), I'd be interested in seeing what she gets compared to you guys picking up 40-50k words.

My wife got 3万2百. However, she did one year of high school and four years of college in the U.S.

Edit: although it's implied, I should specifically mention that my wife is Japanese and the rest of her schooling took place in Japan (although she participated in a special English program at her high school. The program involved triple the amount of English classes compared to the regular course).

Aspiring Wrote:http://www.kecl.ntt.co.jp/icl/lirg/resou...-test.html

Cool, apparently I started Japanese knowing 2700 words. That's how many the first test gives you for knowing what champion, line, and parallax mean).

I guess that's plausible. But I still doubt I learned another 10.000, which is how many it gave me.

riogray Wrote:

sunehiro Wrote:I review cards like アルバム to improve my pronunciation of Katakana words, which is often "unnatural" and quite hard. (also considering the fact that I am not a native English speaker)

Couldn't agree more. Apart from the fact that I often have to look up the word, because the Katakana just don't make sense to me. I'm like "Alubamu, what could this be?" and I just don't get it. Could of course be due to the fact, that I too am not a native speaker.

I understand that just because you speak English, doesn't mean you're automatically perfect with everything to do with English loanwords.

My point is that you don't need Anki to learn them. Why? Because it's easy enough to learn them through immersion alone: when you come across most such words in either spoken or written form, you can figure out what they mean without having to pause and consult a dictionary or other resource. It's more like learning a strange dialect than another language.

Anki comes in when the going gets tough: with complex native words, as a beginner you can listen to spoken Japanese until you go deaf, and you probably won't be able to make any significant headway. Anki breaks down that intimidating volume and complexity into tiny tasks, of learning one word at a time.

Also, if you use Anki with example sentences (as everyone should, frankly), loanwords will still show up, in the example sentences for the native words. So you'll still be hearing and reading loanwords in a slow paced, didactic setting.

Just to sum it up, loanwords are not a hurdle for Japanese learners who already speak English (even non-native English). Yes, they're not automatic, but they're the least difficult aspect of learning Japanese. There's no reason to waste precious study time on this, when there's so many far more difficult things that actually require study.

comeauch Wrote:I really like Anthony Lauder, but it's not a scientific study...

dtcamero Wrote:Anthony Lauder's speech on the seminal work "Learning Vocabulary in Another Language" by I.S.P. Nation. This book is the summary of 600 research studies across 30 years in applied linguistics.

comeauch Wrote:It's the way you said "if you're able to learn from context, I think the evidence suggests that anki is relatively inefficient at high vocab levels" that made me jump. Who is NOT able to learn from context? Some are too scared to try when it comes to language learning, but we're all experts at guessing and making presumptions!

dizmox Wrote:I probably have been through around 28k cards myself, started reading a bit back around the 10k mark. I can't remember words very well just by osmosis since I'm not a heavy reader. Native level vocabulary more or less.... always trying to expand active repertoire though, which is what really matters later on.

comeauch Wrote:So to get back to the original topic, I guess the ideal vocab goal - both in range and depth - would be the point where it becomes more time-efficient (and enjoyable) for you to use that language than srs it.

dtcamero Wrote:so my advice would be to keep anki'ing vocab until reading becomes easy enough to stop actively studying vocab.

dtcamero Wrote:

comeauch Wrote:I really like Anthony Lauder, but it's not a scientific study...

dtcamero Wrote:Anthony Lauder's speech on the seminal work "Learning Vocabulary in Another Language" by I.S.P. Nation. This book is the summary of 600 research studies across 30 years in applied linguistics.

comeauch Wrote:It's the way you said "if you're able to learn from context, I think the evidence suggests that anki is relatively inefficient at high vocab levels" that made me jump. Who is NOT able to learn from context? Some are too scared to try when it comes to language learning, but we're all experts at guessing and making presumptions!

dizmox Wrote:I probably have been through around 28k cards myself, started reading a bit back around the 10k mark. I can't remember words very well just by osmosis since I'm not a heavy reader. Native level vocabulary more or less.... always trying to expand active repertoire though, which is what really matters later on.

comeauch Wrote:So to get back to the original topic, I guess the ideal vocab goal - both in range and depth - would be the point where it becomes more time-efficient (and enjoyable) for you to use that language than srs it.

dtcamero Wrote:so my advice would be to keep anki'ing vocab until reading becomes easy enough to stop actively studying vocab.

It's a speech on a summary of studies. I'm still waiting for ONE study talking about Anki, higher vocabulary and how it doesn't work for people good at guessing. Specifically. This is the guy's opinion based on what he read, not A study. It's okay, he can make correlations, make his opinion and present this, but it doesn't make it true. And summarizing often means not talking about standard error, method of measurement, etc. It's a long shot from "a scientific evidence".

I don't know about you not remembering through osmosis... It's not really about remembering as it is about having a better idea of the language in general (including, but not limited to the meaning of some words). Anyway, I doubt it's because you can't guess. And if you really can't, then, doesn't that mean it would be a good idea to SRS those words? I'm confused.

And yes. We both have pretty much the same opinion, I know. Sorry for picking up on that single sentence lol.

CaLeDee Wrote:

riogray Wrote:

sunehiro Wrote:I review cards like アルバム to improve my pronunciation of Katakana words, which is often "unnatural" and quite hard. (also considering the fact that I am not a native English speaker)

Couldn't agree more. Apart from the fact that I often have to look up the word, because the Katakana just don't make sense to me. I'm like "Alubamu, what could this be?" and I just don't get. Could of course be due to the fact, that I too am not a native speaker.

Even as a native English speaker I feel it's important to put just as much effort into learning katakana-fied English words as any other. The pronunciation is often not what you might expect (ジレンマ comes to mind), and just guessing whether it's 頭高型、中高型、尾高型 or 平板型 will often lead to mistakes. Also, just because you read a katakana word once and know that Japanese people use it doesn't mean you're forever gonna remember that it's a word the Japanese use. I'm sure I've unnecessarily used the Japanese version of words when the katakana English word would have been more appropriate, simply because I forgot or didn't know it was an adopted English word. Then there are 和製英語 like カンニング that mean something different in Japanese than they do in English.

I'm a native english speaker and I agree also. Seems the more advanced you get the more important good katakana word knowledge and pronunciation is to reach a very high level. Ever read a japanese magazine lately ? Almost 30% katakana words.

So think we should all review them to get an easy bonus in our ability.

I'm going to assume we have collective forum amnesia or people just didn't see the thread, but not so long ago there was a similar thread on here and this image popped up. It shows the probability that you will know a random word (y axis) versus number of words you know already (x axis), assuming you learn the words in the order of highest frequency->lowest.



So as you can see, it has a number of curves. The top two are word lists which contain 15,000 and 60,000 with no proper nouns, and the "cb's innocent novel frequency list" contains 300,000 words with proper nouns included.

90% coverage is a pretty good benchmark for I'd say the minimum knowledge required to actually enjoy reading. For the smaller lists, it occurs at just under 5,000 words, and for the big list, it occurs somewhere near 12,000 words.

The novel list will obviously be biased towards more rare and poetic words, and the smaller lists might be more restrictive, so you can take it with a grain of salt, but it gives you an idea for how many words you need.


However, I'd stress the minimum part when we talk about the numbers above. You probably want more than 90% coverage. That tends to mean at least 1 unknown word per sentence, which is a bit of a bumpy ride.


Personally, I did core6k + all the (P) words in edict - all the katakana words after core6k. I also deleted all words that I considered obvious (it happens-- after you've done 6k, you get a sense for things). That landed me at 13k approximately.

I agree that studying katakana words for recognition is pointless 90% of the time. Its pretty easy to recognize the english word they are going for. What you really need is production-style katakana cards so you can actually remember which words you can import from your existing vocab, and which you can't.

But anyways, I thought at first that this would give me like 99% coverage or something, but now I realize that I'm still not there yet. Core6k and the (P) tag are both based on primarily newspapers/academic sources, so they tend to leave some gaps lying around.

Thus, I'm still adding to anki. I don't think I'll ever stop in the near future. I have no problem with the whole attitude of "add everything you don't know". If it came up once, maybe it will come up again, maybe not. A single card in anki takes up so little time to add with rikaisama, and then over its lifetime, you get like 20 reps at 5 seconds each, its like 1.5 minutes of your life in total. You could easily make that up later, when you spend less time puzzling over trying to glean the meaning from context. Also, I've talked before about what I see as a bit of a snowball effect when you start getting an intuitive sense for the kanji.. one vocab allows you to make connections to multiple others, making it easy to pick up more and guess at meanings, etc. Its a hidden benefit of adding lots of vocab that makes the deal just a little sweeter..

I do see what dtcamero is saying though, because after you've been through so many cards, you become so naturalized to the process that at some point you gotta make an effort to break the routine and move on to more practical study. Still, I don't think we need to shame people over using anki a lot. The average learner doesn't come into the language learning scene super hyped up to learn 10,000 vocab, but that's pretty much what it takes... so why not just tell em "shoot for the stars"...