I think most people here would agree, drilling kanji and vocab with WK helps with reading. Each review builds an association between seeing the kanji and thinking of the meaning and reading. However, as I’ve been practicing my listening skills, I find that my association between hearing the reading and thinking of the meaning and kanji is weak. This is especially true for words in my passive vocabulary that aren’t super common, but I would never miss while reading, likely due to the kanji helping me out.
In addition to regular listening practice, I was wondering if, analogous to traditional flashcards (kanji to reading and meaning) helping with reading, audio to meaning and kanji flashcards might help with listening comprehension.
While this is easy enough to put into practice with Anki, would it actually work? Has anyone tried this?
I haven’t tried with audio, but I’ve created flashcards with just the kana of a word instead of the kanji. Worked pretty well for me for listening and also for reading children’s books with fewer kanji.
(I also think you can make flashcards in Anki that play the audio, but I’ve never tried that.)
To play the devil’s advocate, more time spent on Anki is possibly at the opportunity cost of additional listening practice. On the other hand, I doubt I would be reading as well if I tried to learn kanji and vocab purely by immersion. The question is, is it an efficient use of time?
For this I personally use Torii. It’s got an option for Japanese audio → meaning, and it’s got the whole 10k on there. I personally think it helps a ton–you can’t really tell how poorly you remember a word from just the audio until you try it. It, surprisingly, also makes reading easier, since I start associating the word as a whole with its meaning, rather than the “kanji addition” method the Wankikani lesson uses for the mnemonics. I’d say give it a try, either on Torii or Anki, and see how much it helps!
I incorporated it into my main Anki deck so I only switch layout if I feel like focusing on audio based reviews. But that’s personal preference whether to have one deck only or multiple decks, I think.
Apart from that, I do something similar to extensive and intensive reading just with audio / video files.
When I do extensive listening I just follow the program and understand what I understand. So I usually choose material close to my level or something I know the story about already… E.g. I recently joint the Escaflowne movie club here on WK. I already knew the story so I knew what to expect. But it was also nice to see what do I actually understand by going Japanese only (no Japanese subs either). A movie gives visual clues as well but you can do the same with audio only. E.g. folk stories I already know I look up on hukumusume.
I don’t think audio only vocab flashcards are a really efficient way to practice your listening skills, since a lot of listening comprehension comes from context, with so many vocabs sounding all the same when the kanji take the onyomi readings. Not saying it will be useless, but there are better ways to practice your listening skills.
Audio only sentences on the other hand could be more helpful, but then again, I always feel like doing SRS on whole sentences just makes you really good at reading/listening that one particular sentence, but maybe it’s just me.
I believe iKnow has done this, along with cloze and context. The audio is both a full sentence and one word. I think that is the best time spent utilizing multiple methods, cloze, context, and building vocabs at the same time. I don’t know if other apps have this kind of feature. I think you can create it in Anki too.
I’m glad to hear this isn’t a novel concept, and other platforms have already been supporting this. @Furiae that is a fair point about homophones and I’m not sure the best way to handle that. I too have experienced the problem of memorizing a context sentence on platforms like Bunpro. Really it’s the same problem @NicoleRauch would run into with kana cards.
Audio flashcards definitely can’t replace listening comprehension. However, I’m finding it acutely frustrating to hear a word, look it up, and find that I already “know” it.
I think you just you have to lower your expectations and accept this will happen a lot with words you “should know” or have forgotten. Eventually it will happen less and less. Just because you’ve SRS’d the word or read it somewhere doesn’t mean you “knew” the word. That takes multiple exposures in both written and listening context . And probably having to use it. That’s why I find ‘knowing’ a word a pretty ambiguous and somewhat redundant idea. Mostly it exists on a scale of 0 to 1.
For the reasons @Furiae mentioned, I haven’t really bothered with audio only flashcards and haven’t really felt the need to.
I was actually thinking of mentioning that in my first reply, but then I thought that it is actually you who wants to introduce audio flashcards and so I thought you’d be well aware of the issue.
That’s indeed something I sometimes struggle with when doing the kana cards, especially when it comes to the pesky on’yomi words that all sound the same… It worked very well with my N5 and N4 decks, but starting from N3 (plus all the other words I collect during my readings) the words just get more and more, and it gets more and more difficult to tell them apart.
My usual strategy is to try to come up with as many meanings as I can think of, and if the meaning of the card is among those, then that’s a pass. I think this is helpful on a different level as well.
But sometimes that is not enough, and therefore I try to mitigate this issue in two other ways: Firstly, I show some information about the word type (noun, verb etc.) as well as the tags that I’ve collected for that word (in which books did I see it, in which other contexts did I become aware of it etc.). This already helps quite a bit. On top of that, I’ve started to become pickier with which word I want to learn in which direction, and so I move many of my kana cards to a different deck in order to hide them. This is a bit of an experiment as I’ve only started this fairly recently, and I still need to see whether this gives me good results as well. But at least it lowers the frustration significantly, which leads to me enjoying my studies more, which is a big plus for me as well
In a sense, assuming a one-to-one mapping for most things Japanese seems disingenuous. For instance, clearly most Kanji have multiple readings and often multiple “meanings”, in that they can have distinctly different feelings. Vocabulary words also often have multiple meanings and uses as well. To mitigate the issue with SRS, I feel like you basically have to do what you mentioned, deciding whether recalling one usage is a pass or multiple usages. Wanikani certainly takes the former approach. On the other extreme, you’re basically recalling an entire dictionary entry or several for each card, which at minimum violates best practices for effective SRS and at most is a complete waste of time. Perhaps somewhere in the middle is more appropriate, as you also mentioned.
@morteASD That is entirely valid. In a sense, I’m reluctant to not reach for SRS when it has been effective for me in the past. I feel that SRS is effective in getting you the repetition you need when you cannot get enough input. However, clearly many people have learned without so-called audio flashcards. Perhaps I should at least try it to get a feel for how helpful it is, if I’m going to spill this many bytes over it.
I mean, for WaniKani it’s much easier as there’s usually only one valid reading for a given vocab word.
And it’s usually also easier in my experience to go from the word to its meaning.
For me, the same actually happens during listening (searching for a word whose meaning matches the context), so I don’t see it as such a big waste of time. For me it is also a skill to be able to recall different meanings for a given sound.
But yeah, a good balance is definitely necessary in order to not drown in these words…
True, actually that’s fundamentally where the difficulty comes from, isn’t it? I think flash cards are also helpful because the circumstances under which I can do them are different than listening immersion, and therefore it’s not entirely a zero sum game.
On the one hand, in the ten minutes a day it takes to study one of my Anki decks, for the most part I think I’d be hard pressed to learn more just from native materials.
On the other hand, once you’ve moved past the most common words, what remains tends to get increasingly specialized… A lot of the words that appear in my Anki decks make me wonder how often I’m going to actually come across it… Like I’ve learned how to say mahogany in French, but I’ve never actually ran into it outside of that specific flash card… I mean sure it’s a word you’re likely to know if you speak the language, but as a vocab word it’s not exactly the most bang for your buck…
On the third hand (I’ve grafted a cybernetic extra arm onto my body for the express purposes of continuing this metaphor) along the mahoganys, chainsaws and stablehands (I’ve put a bunch of fantasy books into my Anki deck…) I still come across those words that once you’ve learned them turn up absolutely everywhere
All that said, I’ve stopped using Anki for both French and Japanese at this point, just making sure to spend some time with books or videos each day, but tbh that’s mostly due to being fed up with doing so much Anki every day
Your third, cybernetic arm makes a good point. There are definitely those words that Jisho and all the frequency lists think are super uncommon, but lo and behold everyone and their dog uses them. Absolute madness.
For the past few days, I added my listening immersion with iKnow and let it run for few hours. It might help with cementing the vocabs with context. I found it quite good for listening comprehension. Most of the time, I didn’t need to hold every word in my brain and just knew its meaning. I think you can download the core 6k audio elsewhere and loop it instead, but I like the feature to track the listening duration.