test 1: 7000
test 2: 9300
test 3: 5400

mooou~
Well, didn't expect much anyway.

湯たんぽ!!! ❤

Thora wrote:

benkyou wrote:

ive been studing english for a couple years and now my english vocabulary is about 8000 according to my book's examination. [...]

how many do you think you know english words? its my question which is always in my mind.

According to a paper that someone linked to earlier:

* A few extremely common words make up most of the English language. About 2000 word families cover about 85-90% of the words in general written texts.
*We need to know 8000-9000 word families to read a novel or newspaper with 98% vocabulary coverage.
*"Word family" means all related words. For e.g.  use/useless/user/disuse

Benkyou, I think you're doing really well after 2 years! Please don't hesitate to post.

That is a fascinating and very very enlightening read. It has brought to light many reasons why input = output but what kind of input is needed, how much and why. Really interesting explanations on what text books (and things like KO, smart fm for that matter) do for you're learning (or don't) and production ability. Very illuminating indeed. Thanks for the link.

The results of those studies can vary a lot depending on how you count words. What's included in a "word family", what counts as independent words, etc.

mentat wrote:

Each word Family in English might be bigger, and words can carry more meanings. In Japanese each word seems to have a more specific meaning and the word families are smaller.

I find broad comparisons between languages kind of interesting - the idea that some languages are better suited to certain purposes. But beyond that, and beyond providing a gauge for learners, it's hard to imagine assessing or comparing individual vocab levels or isolating vocab from other language requirements.

Since I mentioned the 2000 English word families earlier, I thought I'd refer Benkyou to to a list. My understanding is that it includes 2000 most common words and their word families  (~8000 words) (by Paul Nation, New Zealand)

1st 1000   2nd 1000

There are several other word lists and other information on that site as well.   Wordlists here  There's also a program that colour codes pasted text by vocab frequency levels (and gives a breakdown). The idea is to better select reading material at the appropriate level.

I wonder if that program exists (or could be modified) for Japanese? I have a cleaned up copy of that Leeds Japanese word frequency list around somewhere (~15,000 words). For self studiers, it might be like having a teacher picking suitable material from native sources (at least in terms of vocab.)

Why not combine them? I think there's a core of general vocab and kanji to be learned before needing to think about area-specific stuff.  But even after that, within a particle area or genre there's still a range of difficulty. A lot of reading at the right level really does build up reading fluency. And it's motivating ("Hey, look at me! I'm reading!") Detailed reading and analyzing what's going on can still be done separately.

I agree with Mentat's observation ages ago that grammar's pretty straightforward, but it takes a long time to build vocab and familiarity with usage and expressions. We need to see familiar words used in many different ways. That, to me, is more effective and fun that SRSing a word in the exact same sentence 8 times.

I thought such a program might be useful to identify writers who tend to use simpler vocab, or to check whether an online article or blog is intended for specialists or general readers, etc.

Then again, perhaps reading native material comfortably requires a certain threshold and after that vocab frequency is too mixed.  If so, graded readers and stuff aimed at junior/high school might still be the best option. I don't know, but I'm curious.