Outlier Kanji Dictionary and etymology for teaching

After the previous thread had a run-in with the police, I wonder if people are still up for a discussion about the general concept of using more ethymology for teaching and the Outlier Kanji Dictionary in particular.

@zenpai @blrandel @reichter @VegasVed @hakusaro

Maybe even @OutlierLinguist is still here for further discussions?


I didn’t manage to participate in the discussion, but, as someone who has taught English in both the United States and currently Japan, I think etymology can be a very useful and important element for teaching language. For me personally, etymology has been the strongest way for me to remember kanji, especially for those that lack the usual phonetic-semantic structure.

Count me in on the discussion! :grin:

EDIT: Just looked up the Outlier Kanji Dictionary since the previous post’s link was removed. That looks really cool!

In line with the WK community tradition to casually make requests that require man-years to fulfill for subjective benefits, I would say what Outlier Kanji definitely needs is their own app. I tested the (aptly named) app Japanese they will piggy-back, and … it’s not that great. Apart from having a distinct “straight-outta-JMDict” feeling and example sentences from the Tanaka corpus that didn’t even make it to Tatoeba (second sentence seen: English definitely wrong), when you want to browse kanji using the reference function you basically get a two-fingernails wide pop-up with the information (for example the Outlier dict). That may be semi-good for the information the app usually has, but terrible for Outlier.

The app already includes an Essential Edition for the kanji of school grade 1, and it is not an advertisement for Outlier (partly because all kanji are boring and don’t display any benefit for learning, and partly because they are super-short). Now with the full edition they want to include vocabulary entries inside dictionary entries as an adaption to Japanese, while the same words are already in the normal dictionary. What’s the point of that?

To be useful the app needs to have strong browsing abilities, like showing all kanji with the same component dynamically. With their current approach is seems like they will end up doing it as hyperlinks inside dictionary entries, while you are stuck with a search in JMdict. (It seems Japanese is still in development (?), but I don’t think it will be awesome anytime soon if ever).

I would say either a decent specialized app, or release an API where people can come up with ways to display the information that doesn’t suck.


Absolutely. Thanks for the new thread. I didn’t want to make it myself, for obvious reasons, so I was hoping someone else would.

As @EiriMatsu said, etymology has been, for me too, fundamental in language learning. Most notably, it has helped me greatly in foreign countries because, at times, I can identify roots and links which allow me to glean the meaning of sentences and statements I otherwise would have no business understanding due to my limited grammar knowledge. It has also helped me in learning words in other germanic and latin languages, because the associations between languages make the memories stronger. Not to mention my interest and fascination with etymology.

1 Like

I think in principle it’s cool and they seem to have lots of expertise and high-quality entries, but I just can’t get excited about it. It looks a bit inaccessible, and the benefit of a new dict over getting the Chinese edition is not that evident for me.

[And the “the learning style everyone has been waiting for” put me a bit off :slight_smile:]


I took Latin in high school for that reason, and ended up taking six years of it. I’ve been gradually inserting etymological anecdotes in my English classes. One JTE resisted it at first, saying it was “too difficult” for them, but I now have several third-years who actually come to me after class to ask about why a word is spelled or pronounced the way it is now. It’s so exciting to see how it works for others.

Yeah, I agree. Thankfully, due to my inherent interest, I don’t mind digging through resources to find information, but it would be nice to have a compiled and sleek reference source. For now, your script (it is yours, right? :thinking:) plus Wiktionary provide a faster enough base resource to help ingrain new kanji.

1 Like

I hope @OutlierLinguist will get a WK subscription of any type of they will participate in the discussion though. There are good reasons why we could be skeptical of anyone who comes along to use a forum without contributing.

@acm2010, the lack of a standalone app is the biggest deficiency that grinds my gears too. It is the main reason I will not be participating in this even though I had in the previous, unsuccessful, campaign. And the app that they wish to associate themselves with is substandard at best, and an obstacle to a smooth learning process at worst. I do look forward to an API though.

The marketing is tacky, but seeing how many companies market like this, I have come to expect no better from anyone in general.


The reasons behind the evolution of words are absolutely arresting to me.

And that’s partly why I am so disappointed in this. I desperately want to like it. But I can’t bring myself to it. I find it quite lazy not to develop your own standalone product. Also, why would you want to depend on others to host your product if it is so good and if creating your own platform is accessible as it is, in this case?

1 Like

When you mean Keisei, yes, basically everything is from me. But I mainly searched on free websites, and very often it was like “there are several theories”, so I just went with gut feeling. If you do it right it will be on a whole different level.

But I just can’t see it in learning, like “when you do a WK lesson you also read whole a dictionary entry of possibly obscure trivia”. For initially learning a kanji it doesn’t matter if “the creation was inspired by emperor FuXian’s second horse Matilda’s taste for lush grass in 600BC”, it’s just too much facts at once.

I think it will attract only “experts” in any case, they will have at least the same success with going “the most comprehensive dictionary, first translation of Chinese expert texts into English, so detailed even we are bored by it, etc.”, instead of going for the hipster who wants to read his sushi menu in Japanese to impress his friends.

That much detail, of course, is definitely too far for standard learning. But more broad facts, such as the morphing of the “meat” radical into “moon,” or the origin of “moon” in 朝 to be “river” can be incredibly useful.

Personally, I like going that deep into the information, but that is for my personal interests, and I actually want to pursue etymology in graduate studies, so my interests are definitely stronger than normal learners.

I think this may be where we diverge. I confess that I too would be rather unhappy with reading obscure trivia while learning kanji. But the example you provide is the sort of stuff that helps me cement memories. This is because I would be interested in that line, and the wiki the royal dude’s face, just out of curiosity. By the same token as getting slapped during an event would have me remember it longer than if I hadn’t been.
I think I would quite like an API to pull in a blurb of the etymology into WK that I can click on to explore further.

1 Like

This is definitely the most effective way for me.

1 Like

I find the project quite interesting. At least I liked the examples I’ve seen so far quite a lot. Maybe I’m a visual person or I just like things being explained in a way that doesn’t seem random. The “obscure trivia” part is what would make it easier for me to remember. I also backed that thing that shall not be mentioned and hope it will become available.

I would use it as a supplement to WaniKani because while some mnemonics are so great that they stick immediately some are not as much and this different approach of understanding the etymology seems helpful. At least for 開 which I kept mixing up with 間 before I’m now, after the video or demo or wherever it was featured, pretty sure that won’t happen again.

I don’t care too much about the app since what I would do is probably just look up every kanji once, maybe while doing my lessons here on WaniKani. But if I could choose, I would have preferred a version that I can access on my desktop where I do most of my more “serious study time”.

1 Like

I’m not particularly against it, but I already feel that WK is doing the right thing for example by not showing ON and KUN at the same time, it already confuses me at some low level to see more somehow. After I recognized the kanji I can add as much information as available.

Maybe my problem is that they are trying to treat their dictionary as not being a dictionary but a learning resource for kanji. Doing 月=meat would be great for WK as well, but providing a huge wall of text if different, even if it includes that information as well.


Thanks for bringing this topic back, I think this is definitely worth discussing.

I’m cautiously optimistic about this product because they seem to be claiming that they can reduce all kanji to a pictographic interpretation via identification of functional components and corrupted forms. Rather than replacing mnemonics, which I think they accidentally imply they’re trying to do, this allows for the creation of very powerful mnemonics in which the mnemonic is contained within the components of the kanji. It also allows you to learn how different kanji might be related to one another, which winds up creating a reinforcing network for learning. The examples that they have on their blog and kickstarter page are pretty promising . . .

However, it’s worth keeping in mind that those have been cherry-picked, and the effectiveness of the approach may not generalize to all kanji. Kodansha (KKLC) has adopted a similar approach to Outlier with the key difference that KKLC abandons the etymological approach whenever they deem that there are better ways to build a mnemonic. Still, while Outlier may not be the best approach to all kanji, I think the information contained within will be worth having as a reference.

Some people have noted that the learning app that Outlier is being provided with seems subpar, but I personally don’t think that matters. I think Outlier is best used as a reference to build meaningful mnemonics. It won’t replace WK or Anki, but it will supplement our ability to create our own mnemonics (tbh about half the default WK mnemonics make me want to tear my hair out . . .)

Edit: Something new that Outlier brings to the table is etymology for readings. I don’t know how useful that will be because the readings for Japanese kanji are all over the freaking place, but could be interesting and useful.

1 Like

I think that’s because opposing pedagogical factions have generated a certain, “bubbled” idea of a mnemonic. I had a peer in college who would without a doubt argue that an anecdote about the etymology of the components doesn’t constitute a mnemonic, as it doesn’t provide a connection to the readings. The fact that an extra process of memorization of the sound of that phonetic component means that it fails as a mnemonic. :expressionless:

I agree that such connections between components, which illustrate a sort of story on their own, can highly strengthen memory. I used to muddy up the left side of 朝 when writing it because my brain kept thinking of 車. When I learned that it was supposed to depict the sun rising through the trees by a river, I never messed it up again. :rofl:

1 Like

I think it would be kick-ass to have API access to that data. I would rather have a book than a hard-to-use app.

They seem great with linguistics, but you need some innovation in displaying the information as well, and I don’t think they will pull that off with any kind of money.

I asked them via Kickstarter is they also do stuff like Go’On, Kan’On, etc. and they said “not at the moment, but maybe if there is demand”, I wouldn’t expect too much adjustment for the Japanese version.


I’d like to see that friend come up with a mnemonic containing all the readings for 生.

I think the Outlook people struck a few nerves when they called mnemonics ‘cute stories.’ Honestly, if the cute stories work, go for them . . . Perhaps what they were trying to say was that by coming up with novel mnemonics, you lose semantic connections between related kanji. If they’d said it that way, it would’ve been better received, instead they came off as poo-pooing the entire idea of mnemonics (and came off as a bit arrogant too).


I mentioned it just before the last thread was canned, but I think using WK or Heisig to get the characters lodged in the brain and then using etymology for “A-ha!” moments will be the most effective way of using the dictionary. I cannot see myself studying the kanji without some sort of system in place. Not using a mnemonic system is basically still rote learning - or you’re left to make up your own system on the fly. But why reinvent the wheel, when both Heisig and WK have already done that…

At first I was super stoked about the dictionary, but the more I think about the format in which it is presented, the less enthusiastic I become. Burying the dictionary inside someone else’s app just does not make sense to me.


Agreed. I think they mentioned something about a desktop app that might be better presented than their phone apps? Re: readings, that’s not too surprising. Some kanji have so many that you could probably write a dissertation on the various readings for a single one.

Still, even thinking about kanji in terms of their relationships can be helpful. In fact, thanks to these discussions I noticed that 注 reads a lot like 主 which probably isn’t a coincidence; I also found that connection preferable for remembering the reading vs trying to remember whatever WK wanted me to remember about Chewbacca getting wet. Similarly, 仲 and 中 have the same kun’yomi reading which makes it very easy to remember. These patterns don’t hold for all kanji, especially given the multitude of readings, but it’s another tool in my kit which I can’t complain about.