Outlier Linguistic Solutions

[pross]

http://www.kicktraq.com/projects/johnrenfroe/outlier-dictionary-of-chinese-characters/#chart-daily

[Kobo-Daishi]

@Pross,

 

Thanks for the link.

 

According to the KickTraq site, they're not affiliated with Kickstarter, Inc., so, I wonder how they come up with their data?

 

Kobo.

[OneEye]

Yep I think this is where I'm naturally sceptical.

 

That claim is based on the work of memory expert Dr. Kenneth Higbee, specifically his principles for effective memorization. His first principle is meaningfulness — in the sense that if your brain is able to attach meaning to something, you will be able to memorize it more effectively. Our idea is an extension of that — if your knowledge of a Chinese character is based on a correct understanding of the writing system, you're able to use your knowledge of the system itself to jog your memory of how to write a character, even if you haven't had to write it in months or even years. That's powerful, and it's very different from the usual "cram it for next week's dictation test and hope it sticks after that" approach.

 

 

 

 Slightly exaggerated but no one would advise beginner students of English to remember what words are from Latin, which Greek, etc (even though personally I really like all that stuff).

 

That's more than slightly exaggerated. We're not saying anyone needs to know ancient forms or pronunciation. We've never said that, yet somehow the idea has gotten out that that's what we're advocating. I'm chalking it up to people not reading and making assumptions based on a few pictures. Maybe that's my fault as the marketer, but we've never said those things.

 

I think another thing that's throwing you off is that you're thinking of learning character-by-character, whereas we're talking about the best way to get to literacy in Chinese — that is, get to the point that you can use the writing system as a tool. Very, very different concepts. If you're thinking in terms of individual characters, you're winning the battle but losing the war, or at least making the war more difficult than it needs to be. It's like building the framework for a database before populating it. If you build it as you go you might get there, but it's likely to be an absolute mess. If you focus in the beginning on structuring the thing correctly, populating it will be easy and the data will be coherent and useful.

 

There also seems to be some misunderstanding that we're against mnemonics. We're certainly not. In fact, I think that the most effective way to learn characters is to combine mnemonic techniques with a correct understanding of how the writing system works. We're not suggesting mnemonics in our dictionary because everyone has their pet approach, but we definitely encourage the use of mnemonics as long as they don't obscure what's actually going on with the character.

 

By "form," I mean exactly that: the physical form of the character. The way it's written. With some characters, like 大, the form depicts something. With some, the form is composed of 2 or more components, each of which plays a role or function in the character. The form of 暮 is 莫 + 日.

 

 

 

Armed with this knowledge, nothing stops you from ultimately using "big sheep/goat" as a mnemonic or learning whatever the "official" radical/陪首 is, but you will now be aware of the sound issues and the possibility that both 羊 and 大 (and similar characters) may play similar roles in other characters.

 

Yes. This exactly. Ash likes to suggest using a glass sheep to emphasize the fact that it's an empty component.

 

 

I look forward to hearing why it is true of 天

 

It's a person with a mark at the top of the head. The mark could be interpreted to mean "what's above the head" (hence "sky" or "heaven"). Or — we still have some phonological research to do before we put this in the dictionary but this is one interpretation — 天 originally meant "top of the head" and was the original form of 顛, borrowed later for its sound to write "sky, heaven." Not sure about that one just yet. Either way, 大 is a form component there. 

[realmayo]

His first principle is meaningfulness — in the sense that if your brain is able to attach meaning to something, you will be able to memorize it more effectively.

 

I agree, hence the usefulness of mnemonics, and/or a system like radicals to provide structure; or indeed your system.

 

Our idea is an extension of that — if your knowledge of a Chinese character is based on a correct understanding of the writing system, you're able to use your knowledge of the system itself to jog your memory

 

It's a nice idea. But it doesn't speak to why the less-correct 'radical' system is less efficient for memory jogging. Here's a question. If an etymologically incorrect system could get learners able to recognise 3000 characters quicker than an etymologically more correct one, which would you recommend?

 

I think another thing that's throwing you off is that you're thinking of learning character-by-character, whereas we're talking about the best way to get to literacy in Chinese — that is, get to the point that you can use the writing system as a tool. Very, very different concepts. 

 

Some people start with textbooks which require you to memorise 20+ unrelated characters each chapter. Others wait until they have some general competency with the spoken language and then binge on characters so they can get reading quickly. I was a binger and remember deciding not to distinguish between 月-as-moon and 月-as-meat when inventing mnemonic stories, because I favoured simplicity over complexity.

 

 

So I'm just unsure whether the extra complexity of your system versus the radical system would help or hinder a student who is looking to memorise characters so he can get started reading Chinese texts.

 

But then again, I learned by using the Wenlin dictionary to break down new characters into the their component parts, checking out other characters sharing a component, and so on. To the extent that your dictionary does the same as Wenlin and makes that process easy, with a super interface, as part of Pleco, I think the 'essentials' will be extremely useful to lots of learners. And the 'expert' looks like great fun, I find that stuff really interesting.

 

My scepticism is simply over the claims that the 'system' itself will be a driver of quicker memorisation.

[Demonic_Duck]

I was a binger and remember deciding not to distinguish between 月-as-moon and 月-as-meat when inventing mnemonic stories, because I favoured simplicity over complexity.

 

How is "‘腹’ ([fù] 'abdomen') = ‘月’ ('moon') + ‘复’ ([fù])" somehow easier to remember than "‘腹’ ([fù] 'abdomen') = ‘肉’ ('flesh') + ‘复’ ([fù])"? The former is confusing, whilst the latter makes perfect sense.

[realmayo]

Ha you're right! I just looked it up in an old deck, it was indeed that way round, almost every 月 was meat rather than moon. I guess I already knew 明天 and 星期. No idea what I did for 朦 though... I did find that with a couple of thousand characters under my belt it was easier to learn new ones without making up stories. Nowadays I go by radical + sound component which is usually enough. In other cases the etymology in Wenlin can be helpful and if I had this new dictionary it'd be helpful too.

[Yang Chuanzhang]

On the topic of learning characters with mnemonics and component analysis: http://sinosphere.blogs.nytimes.com/2015/05/20/early-lessons-in-ethnic-stereotypes-in-china/

[OneEye]

I did find that with a couple of thousand characters under my belt it was easier to learn new ones without making up stories. Nowadays I go by radical + sound component which is usually enough.

 

I submit that this was because after learning a few thousand characters, you had developed an intuition for meaning and sound representation in Chinese characters. Native speakers have that intuition too. Our approach gives it to you much earlier (within 100 or so characters), and develops it much more strongly, because you'll know how they work rather than having a feel for how they work. That's the "mental framework" I've been talking about.

[lechuan]

@realmayo

 

 

 

It's a nice idea. But it doesn't speak to why the less-correct 'radical' system is less efficient for memory jogging. 

 

I don't think it's a matter of radicals vs components. Radicals are correct. BUT, Radicals are an arbitrary subset of components chosen for dictionary lookup. People tend to use the word "radical" interchangeably with character components, but they're these are two separate things.

 

For example, if you want to look up 明 using a radical system, the character will be listed under the main 日 heading, but not under the main 月 heading. Both components are equally as important, but one was arbitrarily chosen as the main radical index for this character (correct me if I'm wrong though, I have rarely used radical dictionary lookup).

[realmayo]

I agree the radicals-system is sometimes arbitrary. But any kind of decent system, arbitrary or not, is useful for memorising characters. What I don't understand is why outlier's overall framework is necessarily better for someone who is memorising Chinese characters. The format, the display, the functionality -- all super useful, nicer than zhongwen.com, cheaper and more portable than Wenlin. And using more modern explanations of etymologies is nice, and probably important for scholars. So all in all, likely a wonderful tool that will help lots of people. But for someone memorising the characters, I don't see the convincing step between 'our system is better than the current radicals-based one' and 'you will memorise the characters faster'. And anyway at the end of the day most characters are meaning + sound which is hardly rocket science.

[Demonic_Duck]

The radical a character is filed under only tells you about one part of the character, which may or may not be related to its meaning. Knowing that the radical for “熊” is “火” is useless information as far as remembering its meaning; meanwhile, knowing that the radical for “囊” is “口” is almost completely useless information as far as remembering how to write it.

 

I think it's very easy to see how a more complete approach, which explains all parts of a character in a way that is systematic, ought to allow for greater recall. Mnemonic systems may be able to help you to remember how to write a character, but they very rarely make the link between form and meaning/sound explicit, or do so in a haphazard way, i.e. only for certain components in certain characters, whilst mixing in other red-herrings which have nothing to do with the true etymology. Does the fact that they're red herrings affect your ability to recall? Well, only insofar as it stops you from having a proper mental picture of the system as a whole. If your mnemonic for “妈” is “my mother is a woman riding a horse”, and your mnemonic for “骂” is “two mouths scolding a horse”, you're missing out on the much more useful information that “马” acts as the phonetic in both characters; of course, that example's rather obvious, but there are other examples where the connection is less clear on the face of it, yet knowing it could certainly be a useful memory aid.

[realmayo]

Knowing that the radical for “熊” is “火” is useless information as far as remembering its meaning

 

It's 能 + fire: that's how to remember it. That is, two components: one is the four-dot fire, the other is 能. Easy. The fact that the radical happens to be fire isn't a big deal for the learner who isn't cleaving to any internal logic about radicals: it's just a convenient way to break most characters down to their component parts.

 

Edit: using radicals, you would learn 熊 as 能 + four dots (fire). Presumably, Outlier would break down 熊 as 能 + four-dots (fire) too? So, what's the difference? When I look it up in Wenlin it says "能 depicted a bear; now it is used for the word néng 'able'. Whether the added 灬 four dots depict the bear's feet, or 火 fire cooking bear meat, is uncertain."

 

If your mnemonic for “妈” is “my mother is a woman riding a horse”, and your mnemonic for “骂” is “two mouths scolding a horse”, you're missing out on the much more useful information that “马” acts as the phonetic in both characters; of course, that example's rather obvious, but there are other examples where the connection is less clear on the face of it, yet knowing it could certainly be a useful memory aid.

 

This is the point: where it's obvious, it's obvious. Where it's not obvious, how does it help you remember?

 

I mean, according to some people, 洒 = 灑 = 氵+ 麗 (丽) lì and the 麗 lì is phonetic. If that's true: how does knowing that the 麗 was originally phonetic help you remember the sound sǎ?

 

Or with 达: the full form is 達 but if you're learning simplified characters how does that knowledge help you memorise 达 early on in your studies.

 

 

So pointing out inconsistency or arbitrariness or downright etymological wrongness within the radical-classification system shouldn't be bothering people until they're already reading Chinese quite happily because the system isn't important to most people, they just want to memorise enough characters that they can read Chinese.

[li3wei1]

Just want to make two points here.

 

First, it's too early for Outlier to be making any claims about the effectiveness of their method compared to other methods, because it hasn't been finished yet so it could not have been tested. They can theorise that it will be effective, but that's it.

 

Second, people keep using the word 'systematic'. Any 'system' for looking at Chinese characters can only be as 'systematic' as the set of Chinese characters themselves, and they're just not very systematic. We have a system for classifying animals, but due to the messiness of life, it is full of inconsistencies and exceptions and every branch of the tree looks different from every other branch. It won't help you predict the existence of a species, and to really understand it, you have to study every single branch out to the end. An expert on earthworms won't know how to classify a bird. Outlier's system uses something called 'empty components', which to me look like a fudge factor to bridge these inconsistencies. Every other 'system' will have similar fudge factors, so it's no worse, but I'm not sure it's better, and it certainly isn't the holy grail of the unified theory of Chinese characters that we all dream is right around the corner (right behind the unified theory of English spelling and pronunciation).

[OneEye]

First, it's too early for Outlier to be making any claims about the effectiveness of their method compared to other methods, because it hasn't been finished yet so it could not have been tested. They can theorise that it will be effective, but that's it.

 

That's not accurate. A published dictionary is not the same as a method. We've taught people characters using our methods, Ash has given talks at conferences and ICLP, and we even taught a very well-received course at ICLP. We haven't used the dictionary, but we've certainly used and tested the method with students.

 

Any 'system' for looking at Chinese characters can only be as 'systematic' as the set of Chinese characters themselves, and they're just not very systematic.

 

Could you explain what you mean here, or give an example of a character you think our framework can't accurately explain? I think they're quite systematic, if you understand the processes by which they evolved, the historical phonology, and all that.

 

"Empty components" is our umbrella term for several different phenomena — corrupted components and differentiation marks are two of the most common. Those phenomena aren't especially important for the learner to know about — though we do include that info in the Expert Edition. The important thing for a learner to know is that it isn't expressing a meaning, form, or sound — hence "empty." That's not to say that we don't understand what's going on (that may happen, but we'll always mark when there's uncertainty), it's just that we've simplified it for learners. Chinese characters aren't as complex as life — nobody specializes in a single character.

 

 

I mean, according to some people, 洒 = 灑 = 氵+ 麗 (丽) lì and the 麗 lì is phonetic. If that's true: how does knowing that the 麗 was originally phonetic help you remember the sound sǎ?

 

The sound formulas which will be in the official release explain variation in the sound series (諧聲系列). That is, they'll give a simple explanation for why 麗 can represent the sound sǎ, or why 立 is the sound component for 泣 qì, etc. If you know the possible variations, 's' <-> 'l' isn't surprising for 麗, nor is 'g' <-> 'l' in the 各 series. The formulas aren't in the demo (there's a place holder on the Demo page) because Ash hasn't published his paper on them yet.

 

 

Or with 达: the full form is 達 but if you're learning simplified characters how does that knowledge help you memorise 达 early on in your studies.

 

You don't need to know the traditional character for 达 if you're only learning simplified. It's enough to know that 辶 is a form component (a road or intersection), and that 大 is both a sound component and a form component (a person crossing said road --> to arrive).

[Basil]

OneEye,

 

Great blog, and v. promising looking product.

 

If you can just send it back in time to when I first started learning Chinese, that'd be dandy.

[OneEye]

We'll make that a stretch goal. ;)

[Demonic_Duck]

Edit: using radicals, you would learn 熊 as 能 + four dots (fire). Presumably, Outlier would break down 熊 as 能 + four-dots (fire) too? So, what's the difference? When I look it up in Wenlin it says "能 depicted a bear; now it is used for the word néng 'able'. Whether the added 灬 four dots depict the bear's feet, or 火 fire cooking bear meat, is uncertain."

 

“能” isn't a radical, so I don't see how the radical system would help you learn it.

 

This is from my memory of how I've heard this character broken down before (not sure to what extent it's correct or authoritative):

 

It can indeed be broken down into “能” + “灬”. The original form for the meaning "bear" was in fact “能”, which then got borrowed for the "ability" meaning (presumably due to being homonymous at the time). In fact, “能” (which happens to be classified under the radical “肉”) was originally a pictographic of a bear: “厶” represented the head, “月” the body (hence “肉” actually isn't too much of a stretch), and the two “匕”s were two legs. So, why the heck did they add “火” to reclaim the original meaning of "bear"? Why didn't they just add “犬” (“犭”) on the left hand side like so many words for animals? The answer is, of course, that it's not “火”; it's a representation of the fact that bears have four legs (cf. “馬”). As for “鳥” and “魚”... those I'm not too sure about.

 

^ For those in the know, I'd be interested to know to what extent that's correct.

[realmayo]

“能” isn't a radical, so I don't see how the radical system would help you learn it.

 

DD I'm not saying that the "radical system" is some super-cool system, it's just a guide to breaking characters into their component parts and given that most characters are meaning + sound, it works fine with most characters. But it's not a system to follow or anything hardcore like that, I'm not suggesting it is 'the way to learn characters'.

 

It can indeed be broken down into “能” + “灬”. The original form for the meaning "bear" was in fact “能”, which then got borrowed for the "ability" meaning (presumably due to being homonymous at the time). In fact, “能” (which happens to be classified under the radical “肉”) was originally a pictographic of a bear: “厶” represented the head, “月” the body (hence “肉” actually isn't too much of a stretch), and the two “匕”s were two legs. So, why the heck did they add “火” to reclaim the original meaning of "bear"? Why didn't they just add “犬” (“犭”) on the left hand side like so many words for animals? The answer is, of course, that it's not “火”; it's a representation of the fact that bears have four legs (cf. “馬”). As for “鳥” and “魚”... those I'm not too sure about.

 

People learn in different ways. When I was learning characters, that would be too much to hold in my head. I'd just remember “能” + “灬”. But others might thrive on that complexity, who knows.

[realmayo]

You don't need to know the traditional character for 达 if you're only learning simplified. It's enough to know that 辶 is a form component (a road or intersection), and that 大 is both a sound component and a form component (a person crossing said road --> to arrive).

 

 

So in, say, 庆, what's the 大?

 

Seems to me that if every time you see a component you have to work out "what 大 is this one, is it a sound component or is it a form component or is it one of those characters which is both, or maybe it's a meaning component, or have I got it wrong and it's one of those meaningless components" seems like a bit of an effort?

 

 

Helping students breaking down components, and showing them which indicate meaning, which indicate sound, and which do neither (but here's the etymological backstory to click on if you're interested) is very important and although it's done elsewhere, quite possibly it's not been done so well as Outlier is proposing to do on Pleco.

 

 

But talk of a wonderful new system leaves me cold because I don't see much that's new about it, and that which is new seems to impose an additional memory burden on the learner. I don't see why this product needs to push the system when it's already offering something that should help people memorise characters quicker.

[imron]

So in, say, 庆, what's the 大?

It's not the da