Sentence miners, when did you start? How do you choose what to sentence mine?

[cheruko]

Basically what the title says!
I’m very interested in starting sentence mining, but I’m not sure if I should wait until I am further along/finished with Wanikani. I’m already N2 and I’ve been doing really good with wanikani, but my current level (28) is the longest one yet (now day 19 ). Mainly because I let vocab lessons pile up to 500+ and finally decided to focus on them before moving on.

I’m afraid that if I started sentence mining now, it would add a lot more time to reviewing

Another question, how do you personally choose what to sentence mine? For words you don’t know, do you do them even if you haven’t learned the kanji for them yet?

Please share any of your experiences with sentence mining! I’d love to hear everyone’s experiences/advice!

[ChristopherFritz]

@cheruko

Mainly because I let vocab lessons pile up to 500+ and finally decided to focus on them before moving on.

I highly recommend focusing on this before getting to sentence mining. Unless you’re really good at remembering kanji readings and meanings, you could be setting yourself up to forget them before you get to the words that use them.

Then again, if you’re currently actively reading, you may be seeing the kanji in the material you read, so that may not be an issue for you. I guess it depends on how much you feel you’re getting out of doing vocabulary through WaniKani, versus using the site only for kanji.

Personally, I do all my vocabulary lessons before I get to any kanji lessons, but I also don’t mind the slower pace that this results in.

@cheruko

how do you personally choose what to sentence mine? For words you don’t know, do you do them even if you haven’t learned the kanji for them yet?

The hardest part for me on starting sentence mining was that it was hard to find sentences with just one thing I didn’t know.

At first, it seems simple: “This sentence has only one word I don’t know. Making a card for a sentence with one unknown thing is considered the best way to learn.” (And indeed I’ve failed at mining attempts with multiple unknown words.)

But then I realize, “This unknown word has two kanji I don’t know. That means there are three unknowns. Unless I count the meaning and reading for the word separate, in which case it’s four unknowns.”

Sure enough, I fail at “one unknown word” sentence if I don’t already know the kanji.

Initially, I started sentence mining only if the unknown word included kanji I already knew, but I was weak at remembering the reading, or if I couldn’t figure out the meaning.

More recently, I’ve been using the Migaku Kanji God add-on for Anki, which auto-generates kanji cards for me. This allows me to create a sentence card that has an unknown word also with unknown kanji. Depending on volume of cards and number of lessons done per day, it’s possible to do the kanji card lessons a few days before I get to the sentence card lesson.

This helps reduce a sentence card from “unknown word reading, meaning, and kanji” to “unknown word reading and meaning”, which is a bit more manageable.

I don’t do enough sentence mining to really say if this is helping me out, though. I’m currently still using WaniKani as my primary source for learning new kanji.

[cheruko]

@ChristopherFritz

I highly recommend focusing on this before getting to sentence mining.

I definitely want to start making the vocabulary lessons more of a priority for me on wanikani. I do enjoy them more than just the kanji, as I can actually see and hear the words I’m learning start to make sense in conversations and my everyday immersion.
Definitely will try doing the vocabulary lessons before the kanji and radical ones so they don’t pile up!

@ChristopherFritz

“This unknown word has two kanji I don’t know. That means there are three unknowns. Unless I count the meaning and reading for the word separate, in which case it’s four unknowns.”

Sure enough, I fail at “one unknown word” sentence if I don’t already know the kanji.

This is what I am afraid of! Okay, seems it would be best to start with words made up from kanji I already am familiar with.

An unfortunate thing is that I am already immersing with lots of Japanese content and looking up words I don’t know, but I really don’t want to overwhelm myself with adding those unknowns into sentence mining.

[ioric]

I think aiming for only “just one thing you don’t know” is counterproductive, especially if you’re going to split hairs and count a word with a kanji you’ve not encountered as 2 or 3 things you don’t know. i+1 at the end of the day isn’t a real mathematical formula, it’s a heuristic that depends on what you consider “+1” to be.

I speak with the experience of most of my learning outside of WK coming from pre-made Anki decks of N5 and N4 sentences. Sometimes, a sentence does introduce more than one unknown. Sometimes a sentence teaches you that 窓 means window and 閉める means to close it. That was completely doable, even though both of those kanji were new to me. I learnt words like 動物園 and 自動販売機 and 健康保険証 before I’d come across those kanji in WK. I won’t claim it was the easiest thing, and it doesn’t mean I learnt individual meanings of each of those kanji, but I learnt that 動物園 means zoo, and I could recognise it in other contexts.

With sentence mining, you already have a massive advantage over studying pre-made sentences because they’re things that you’ve come across in context and are interested in. So I would say don’t fret and limit yourself by trying to find ideal sentences to learn from. Err on the side of adding more sentences and removing them later if they are too difficult or too boring. Limit the workload with how many sentences you learn from the deck each day, rather than how many you add.

[NyappyTiramisu]

I think I started sentence mining when I was roughly around the same WK level as you. At first I decided to only add new words to my deck when I knew all the kanji because I didn’t know where to start and that narrowed it down a bit But after a while I also started adding other words that are either common or that I thought were interesting, and not having studied the kanji separately didn’t feel like too much of a hurdle.

@ioric

So I would say don’t fret and limit yourself by trying to find ideal sentences to learn from. Err on the side of adding more sentences and removing them later if they are too difficult or too boring. Limit the workload with how many sentences you learn from the deck each day, rather than how many you add.

I’d absolutely second this! If you remember the context of a sentence, it’s not that big of a deal if there’s more than one unknown element in it in my experience. You can always replace sentences later on, so I’d say it’s worth experimenting a bit to see what kind of ambiguity works for you in sentence mining.

[alexsandred]

i started around 1000 kanji. what to choose? honestly, the content you enjoy- with yomichan + anki combo installed you can grab any unknown word with the whole context sentence- I usually just watch animes on animelon with Japanese subs.

[ChristopherFritz]

@ioric

I think aiming for only “just one thing you don’t know” is counterproductive

It will definitely vary from person to person. Some people are much much better at it than I am. My own personal experience is spending a few years going through iKnow’s Core 2K, and not picking up any kanji or other words appearing in sentences. (Not everyone will be as bad at this as I am!)

@ioric

Err on the side of adding more sentences and removing them later if they are too difficult or too boring.

I would add to this, don’t worry if things become leeches and auto-suspend (if you’re using Anki). Every review spent on a leech that you aren’t getting is time that could instead be spent on a new card that maybe you will get. Chances are the leech item will be easier in the future after you encounter the word more, and potentially make a new card for it later.

[cheruko]

@ioric

With sentence mining, you already have a massive advantage over studying pre-made sentences because they’re things that you’ve come across in context and are interested in. So I would say don’t fret and limit yourself by trying to find ideal sentences to learn from. Err on the side of adding more sentences and removing them later if they are too difficult or too boring. Limit the workload with how many sentences you learn from the deck each day, rather than how many you add.

Thank you! And very true! There is no correct formula, so I’ll just have to try it out until I find what works for me!

I’ve just set up Anki, and as long as I limit the workload to a very small number a day in the beginning, I think I should be okay with not overloading myself. And since it is all from things I’m interested in, that’ll help keep me motivated!