Kickstarter for Outlier Kanji dictionary is now live

Yeah, this is a relaunch. I had backed this at Expert last time and they failed to secure funding. Even created a thread about this last time. People weren’t too keen back then.

Thanks for posting about our project, @zenpai!

I wanted to address some of the concerns.

First, the core value of our dictionary is not in the ancient forms, but in the fact that we explain kanji structure in terms of their functional components. That is, we break each kanji into its components and explain what each component’s function is: is it depicting something (a form component), is it related to meaning (a meaning component), or is it indicating the pronunciation (a sound component)? Sometimes a component does none of the above (empty components either 1. simply serve to distinguish two characters from each other, or 2. are a result of corruption, in which case we explain what the component was originally and what its function was).

As I explain in the video linked above, that allows you to see the real meaning and sound connections between kanji, so that you develop an understanding of the logic underlying the writing system.

We have a few other similar videos on our Facebook page.

The dictionary will be an add-on for the Japanese app, and it will integrate with their SRS and other forthcoming learning tools. So it’s a dictionary/reference, but it will also have SRS functionality.

We’re also planning to build a web version, which will have more robust learning functions, mnemonics, vocabulary, even sentences with audio, etc. The eventual goal is a one stop shop for literacy in Japanese. This will be a stretch goal for our project, which we’re announcing soon.

As for what happens if the Japanese app ceases to exist or support our dictionary in the future for whatever reason, we will make sure that all of our Kickstarter backers have access to it in some form or another, whether a PDF, or building our own app, or something else.

As for “etymology vs. mnemonics,” we’re actually huge fans of mnemonics. But they’re much more powerful when they’re based on an understanding of the kanji’s actual structure, which is what we’re providing.

@acm2010 I’m not sure what you mean by “just pulling in stuff from JMDict,” but that’s not what we’re doing. We’ve hired Ulrich Apel, one of the world’s foremost Japanese lexicographers (he did WaDoku) to create the meaning, pronunciation, and vocabulary data for us. We’re not using any open source data here.

Yes. According to memory expert Kenneth Higbee, the #1 thing that increases retention is that the learner understands the thing to be learned. This explanation tells you that 1. 日 gives the meaning. 2. That 𡗗 does not give a sound or meaning. 3. Why 𡗗 is there to begin with even though it’s not giving a sound or meaning. In other words, this short, concise explanation for 春 tells you exactly why it looks the way it does. And, after reading it, you fully understand the form. Once you understand how kanji really express sound and meaning, the easier it is to remember them, recall them and predict things about ones you haven’t learned yet.

As for how knowing the original meaning helps you to learn, it explains the connection between the kanji’s form and its meaning. One example is 漢: why does it have water 氵? Because it originally referred to the name of a river, then to the people who lived near it, to the dynasty they founded, and then to the ethnicity of the people that dynasty ruled over. You don’t have to memorize all of that, of course, but it does help to know there’s a logical reason for “water” being there.

I’m not sure exactly what you mean. Can you give me an example of a kanji you think our system wouldn’t be able to handle?

Thanks for the discussion!

I was just looking at the JapaneseApp on that page, it would be a wonder if 姿造り = whole fish sashimi arranged in the original shape (from the “Lightning Fast Search”) is not from JMdict. But good to hear that you will generate a good dictionary.

[Not sure if the world is waiting for another closed-source JP-EN dictionary, though.]

For the second part, more info sure is helpful. But let’s say 𡗗 derived from multiple origins. Actively remember which origin it was takes more effort not less. (No guarantees for 𡗗 as an example, but I’m sure there are multiple shapes that started out differently and merged into the same shape).

Edit: the most helpful thing would actually that 屯 is the phonetic component, even if it doesn’t appear anymore and the reading is not とん (anymore?) but しゅん.

I happen to have a copy of Heisig’s Remembering The Kanji: 1 right here. I thought I remembered reading something about this in the introduction, and I was right:

…So fascinating is this story that many recommend studying etymology as a way to remember the kanji. Alas, the student quickly learns the many disadvantages of such an approach. As charming as it is to see the ancient drawing of a woman etched behind its respective kanji, or to discover the rudimentary form of a hand or a tree or a house, when the character itself is removed, the clear visual memory of the familiar object is precious little help for recalling how to write it. Proper etymological studies are most helpful after one has learned the general-use kanji. Before that, they only add to one’s memory problems.

From there follows about five more paragraphs in the intro to RTK that further detail why mnemonics are the linchpin, so to speak, of RTK and WaniKani.


That isn’t to say that your the project is bad, per se – it just doesn’t seem like it’s optimized for learning so much as it is interesting knowledge after the fact.

I know you’re trying to kickstart a product, but unless you have those plans already in development, they’re just empty promises. Japanese one of the most complex and time consuming languages to learn. You need listening, speaking, reading, “writing” (okay, so typing in today’s world), and a whole lot of interwoven grammar and vocabulary. A so called “one stop shop for literacy in Japanese” would actually need to include all of these elements – at once – in order to truly achieve that goal.

I’m only saying this because you’re officially representing the company, and the semantics behind what you said imply a far greater amount than what you’ve presented in your kickstarter campaign.

E: This was intended to be a direct reply to the post I quoted from – not just a reply to the thread.

1 Like

I do agree with the project in the sense that more information is always good. Some things will be remembered, and some things won’t, and that’s OK. The point is to aid in understanding and in retention.

If I got slapped while eating my morning toast, I would remember it for a long time. If I learned something astonishing or interesting about a kanji’s origin, I’d be more likely to remember it. Memory works on associations quite well, since it tends to be one of the foundations of memory.

@OutlierLinguist, I appreciate what you say and that you are addressing concerns pertaining to your kickstarter. But just as a question of good form, it is typically hoped for that if you will promote a product, in any way, you will at Keats be a subscriber of the platform so it doesn’t seem exploitative. Perhaps consider getting a monthly subscription so it’s not too expensive and you look better.

While it is true that mnemonics appear to work, I must confess that RTK touting it does not really inspire confidence in the method. After all, that is what they base their method on. Due to the vested interest, it is expected that they would promote mnemonics…

What I particularly dislike about this project, though, is that it is not standalone. If you’re gonna try and sell me on this type of product, especially at those prices, make sure it’s a product you’ve made in its entirety. I shouldn’t need to depend on a 3rd party platform.

1 Like

Correct me if I’m mistaken (I’m far from an expert), but it’s my understanding that the historical/etymological roots for many kanji are a bit removed from their present meaning. Kodansha (KKLC) uses 東 as one such example, stating that the ancient form of the kanji actually represented a water skin pierced with a stick. While interesting as a bit of trivia to bust out at your next kanji party, referring to the kanji’s etymology doesn’t seem to be especially useful in the learning process. How would your dictionary address this particular item or others like it?

Well of course – but RTK was built by someone who needed to learn the kanji quickly. Mnemonic based learners are really easy to come by, too, and RTK kinda helped pave the way for that. It has standing because of its success, imho.

The biggest thing that irks me is that it’s billed as the end-all be-all solution to learning kanji. There are a lot of great reasons to have a dictionary – but language learning is so hard to get right that I’m naturally skeptical of anything claiming too much.

Also – and this is just to be pedantic – you really need more than one expert/scientist to back up a claim. One person is just one person. The scientific method works because of repeated trials and hypothesis testing, not just the testament of one person.

2 Likes

Completely agree with everything stated here.

1 Like

I had some issues with the Chinese Expert Edition being on pre-order while kickstarting something else that also has an expert edition, and that was six months ago. That edition seems to be still on pre-order?

2 Likes

No one expects you to remember that, only to read it. 𡗗 does have multiple origins. All you’re expected to remember about 𡗗 is that it is not giving a sound or meaning.

The problem is, once your memory stories fall away, you’re only left with sounds and meanings of the words you want to write. When you make up a story that is not inline with how the character actually works, you can no longer use sound and meaning as memory clues to remember the form. We like memory stories. We just think the ones based on how the character actually works are much more effective in both the short term and the long term. When you hide the logic of how kanji work, you lose a powerful tool for both learning and recall.

Etymology doesn’t mean “what the ancient form was a picture of”. That is part of it, but only a small part. We have 2 blog posts explaining this which should get posted soon, but here is the gist:

The posts show six aspects of character etymology that are pertinent to learning Kanji:
#1: Identifying the functional components in a given character.
#2: Identifying how the functional components function.
#3: Identifying corrupted components.
#4: Identifying the meaning that a given character was invented to represent and its relationship to the character form.
#5: Restoring the pictorial quality of character components.
#6: The full story.

As far as 東, I don’t know of any paleographers that would describe it as “a water skin pierced with a stick,” so I’m not sure where KKLC is getting that. We describe it as:

東 depicts a bag tied at both ends, which is also its original meaning. The modern meaning “east” is due to sound loan.

It’s important to remember that only the original meaning is tied to a character’s form. We explain what the form is and how it relates to common modern meanings. We also show the logical relationships between modern meanings.

I see a pro and con.

Pro: In combination with WK or Heisig it could be a powerful aid in cementing the meaning of a kanji. The more neurons I can activate in the process of learning something, the more likely I will recall it later. Concepts are also much more easily recalled than specific facts, because they rely on understanding. So if I understand the concept behind a kanji component, I think that will help greatly combined with a memorable mnemonic.

Con: Hiding the dictionary behind an app may help protect its contents, but also does not show trust in the userbase and potentially causes usability to suffer. I think, no matter what, all backers should receive a pdf with the dictionary.

Every time people complain about piracy, they should remember game publisher/store front GOG. All DRM free games, but they still sell well, because people show their appreciation and enjoy being trusted. They also like to feel they actually own a product and that their access to it isn’t nebulous or dependent on outside factors. You can download a game to your harddrive and it will work forever, regardless of whether GOG closes shop or not. The same cannot be said for this dictionary, at this point in time.

I think app access to the dictionary should be a bonus for paying customers. An added incentive. Reputation means a lot online and doing right by the community can have far lasting benefits. The opposite is also true.

I already backed the product for an expert package, but I’d love to have an option for the dictionary to be free of any app ties.

1 Like

I mean that’s nice, but does that help you in any way to learn 春 faster?


Basically I agree that knowing the origins is useful to interconnect the kanji and also get the sound clues, but (after working like this for a while) I don’t like this touting as the best way to learn kanji. Maybe it helps out to iron the last 500 of jouyou kanji. [And you do have to remember the details, right?]

The pet examples of 開 (oh it’s hands and a door!) or 漢 are nice because they have a straight story, but my impression was that a large fraction of the origins are just obscure trivia (like 東) or completely corrupted, which neither help you to figure out the reading or meaning. “Kanji are based upon simple logic” is more wishful thinking, especially with Japanese where they just decided to replace kanji with others quite randomly during reforms.

It’s a lot of work to produce a dictionary and it is useful especially because the best resources are only available in Chinese or Japanese, but once you have access to native sources you wonder what is the big benefit, and it just feels like overselling as a kanji learning resource.

1 Like

Is there not an argument to be made for learning for the sake of interest and not speed? I like knowing details and learning in general.

Having said that, I find you are completely rulight when you say that:

I agree that “speed” is maybe not the right word, maybe effort spent vs. output?

The basic premise is that other methods are inferior because you don’t know the deeper meaning. After you forgot the mnemonics you only have the reading and meaning, so it’s harder to remember the shape.

I was just wondering if [polemic paraphrase] “it was 屯+艸+日 back in the days, 屯 indicates とん, but it actually merged into 𡗗 where it doesn’t mean much anymore, and the reading changed to しゅん anyway, but the sun represents ‘the first of the four seasons’, so it obviously means spring!” will jog your memory more.

As a lookup tool, definitely, maybe the most complete one, cutting edge research, first translations of Chinese only sources, there could be lots of additional sale points. But it reduces the target audience to some die-hard kanji fans like me :slight_smile:

2 Likes

Thanks for taking the time to answer. I backed it yesterday as the project sounds interesting and I’m all for more kanji resources, I wish you guys luck.

By any chance, would it be possible to get more of those videos (the 開 one)? It was nice to have the connections between various kanji explicitly and concisely drawn out in a single place like that.