Context Sentence scaling


#1

Hullo everybody! First post here!

A problem: I have been consistently glossing over the context sentences as they pretty much always contain kanji I haven’t learnt yet.

A solution: Could the context sentences use only kanji from the current level or lower? Or if there wasn’t enough in one level to make fun/practical examples, perhaps they could be broken up into 5 level chunks (1-5, 6-10 etc). So when you get to level 6 the sentences would contain a majority of unknown kanji, but once you reach 10 you know them all. Then when you reach 11 it resets. It could be a nice wave of ‘I’m doing it! I’m the king!’ then back down to ‘urgh I’m so stupid. Time to work harder!’.


#2

It’s good practice for the real world where you’re going to be seeing first Kanji, and then words, you don’t know, in every sentence you see, constantly, for a long time.


#3

I tried making some at one point and ran into the very problem you foresaw. But that idea about 5-level chunks, especially at the beginning, is a good idea. I’m generally in the same camp as Syphus that the more you expose yourself to native Japanese, the faster you’ll get more comfortable reading it, and any words/kanji you don’t know you can get in the habit of looking up with Rikai, but it might be nice to have a set of simple sentences alongside the more complex ones.

In the meantime, you might be interested in the Tofugu 500 sentences, which tend to be more accessible at a lower level. I rearranged them to be in order of WK level, so that you can find sentences that contain only kanji you’re familiar with. You can get it from my dropbox here. (I actually have a sheet of 5000+ Japanese sentences by WK level but I’m not allowed to share it!)


#4

Thanks BreadstickNinja. I’ll check Rikai out, and those sentences look like exactly what I’ve been looking for. And I agree it’s good to be seeing unfamiliar characters, but all the time can be a little demoralising at first! Still, 頑張ろう!


#5
BreadstickNinja said...In the meantime, you might be interested in the Tofugu 500 sentences, which tend to be more accessible at a lower level. I rearranged them to be in order of WK level, so that you can find sentences that contain only kanji you're familiar with. You can get it from my dropbox here. (I actually have a sheet of 5000+ Japanese sentences by WK level but I'm not allowed to share it!)
 Thanks for sharing!


Syphus said... It's good practice for the real world where you're going to be seeing first Kanji, and then words, you don't know, in every sentence you see, constantly, for a long time.
Maybe, but they could still ease you into it in the first few levels.   I remember getting so discouraged by the random sentences that were WAY over my head, I just ignored them.  It wasn't until I hit level 15 or so that I started to read them again.

Picking out a random word...here is the example for 人口 which is learned in level 1:
庭に人工芝を敷いて見ました

I realize it is hard to write meaningful sentences for someone that only knows 1-10 kanji and zero vocab but I feel like this is too much (or at least it was for me!).

#6
StickyThumb said... Artificial population hits again ;)
 ....wasn't your username ivanpellegrin...?

#7
BreadstickNinja said...In the meantime, you might be interested in the Tofugu 500 sentences, which tend to be more accessible at a lower level. I rearranged them to be in order of WK level, so that you can find sentences that contain only kanji you're familiar with. You can get it from my dropbox here. (I actually have a sheet of 5000+ Japanese sentences by WK level but I'm not allowed to share it!)
Wait, what?  Why not?  Thanks for the ordered 500, though.  Being able to tell roughly which sentences line up with words I ought to know should make that list easier to approach.

As for the context sentences, while it can be a bit frustrating to have most of the vocabulary used in them go completely over my head, the few occasions when I can pick out more words than the one being taught at that moment make for pretty good motivation.  They're a reminder that I've actually learned something.

I'm liking that levelled chunks idea, though, as a balance between "too easy" and "what am I even looking at".  I wonder if something like that could be done with a userscript or some such...  May have to give that a shot if I can figure out how to get an appropriate sentence (with translation) for each item.

#8

@GangsterOfBoats
there’s 4,500 available for sale here: http://store.tofugu.com/shop/4500-japanese-sentences


#9
GangsterOfBoats said...
BreadstickNinja said...In the meantime, you might be interested in the Tofugu 500 sentences, which tend to be more accessible at a lower level. I rearranged them to be in order of WK level, so that you can find sentences that contain only kanji you're familiar with. You can get it from my dropbox here. (I actually have a sheet of 5000+ Japanese sentences by WK level but I'm not allowed to share it!)
Wait, what?  Why not?  Thanks for the ordered 500, though.  Being able to tell roughly which sentences line up with words I ought to know should make that list easier to approach.

As for the context sentences, while it can be a bit frustrating to have most of the vocabulary used in them go completely over my head, the few occasions when I can pick out more words than the one being taught at that moment make for pretty good motivation.  They're a reminder that I've actually learned something.

I'm liking that levelled chunks idea, though, as a balance between "too easy" and "what am I even looking at".  I wonder if something like that could be done with a userscript or some such...  May have to give that a shot if I can figure out how to get an appropriate sentence (with translation) for each item.
 It's not incorporated into WK, but this could be useful for you: /t/Duendecat-Sentence-looper-by-WK-level/4156/1

#10
ShotgunLagoon said...there's 4,500 available for sale here: http://store.tofugu.com/shop/4500-japanese-sentences
I think I'll get started on the 500 first, but thanks.  I guess that's something to keep in mind for the future.

@EskimoJo
Bookmarked it.  Thanks.  That should be good for some reading practice, even if some of the translations sound a bit off now and then.  Reading the many warnings and disclaimers for the Tanaka Corpus (used as as the source for those sentences), though, it's understandable.  What I was/am thinking about is how to show some (theoretically) more "level-appropriate" context sentences alongside the existing ones, with the most obvious challenge being getting a list of sentences together.  I don't know how feasible that really is, but it seems like I've got somewhere to start now, at least, so thanks again.

#11
BreadstickNinja said...In the meantime, you might be interested in the Tofugu 500 sentences, which tend to be more accessible at a lower level. I rearranged them to be in order of WK level, so that you can find sentences that contain only kanji you're familiar with. You can get it from my dropbox here. (I actually have a sheet of 5000+ Japanese sentences by WK level but I'm not allowed to share it!)
 The amount of work you do for this community is somewhat astonishing.

#12

@StickyThumb 
Perhaps that comes off a bit stronger than I intended.  Anyway, I can’t really claim to be an expert either.  I just briefly looked into the corpus after seeing it mentioned in post that @EskimoJo linked.  The warnings I mention, some of which may be out of date, come from the EDRDG wiki on the corpusthis page and the wiki for the current version.  Basically, the main points are these:

  • The original corpus data was entered by students who made mistakes (some of which may remain) and the current project is a public collaboration that can’t necessarily be checked fully, so errors may exist in the material.
  • Unnatural language may be used in places as a result of literal or mechanical translations, or those done by language learners.
  • Translations may use archaic language (which is not to be updated), incomplete sentences, etc.
I guess the TL;DR is that it’s unlikely to be error free and odd language usage in places means the sentences don’t necessarily make for good learning material beyond reading practice.

#13

http://www.manythings.org/corpus/warning.html Here’s a handy warning about the site. 

I would highly suggest using something like ALC or some of the other options on Weblio instead of Tanaka.