I don’t think spelling is a good analogy here, but I think I understand what you mean! As mentioned in the article, I don’t think that a deeper level of understanding is always better. Most people don’t learn characters just to learn characters, so there’s always an opportunity cost. It would certainly be easier to learn if you have the information readily available, but the filter you mention would be incomprehensible without commentary anyway. It also would directly impede your ability to read modern Chinese, which is after all what most people are after!

First and foremost, thank you for sharing detailed examples of how you memorise characters. I think this is interesting, but I’m sure that many other learners will find it interesting, too! You are pretty close to what I’ve been doing before, even though I don’t focus so much on learning characters these days. I think this approach is suitable if you have some interest in characters beyond merely being able to write them (and maybe if you’re after long-term retention and systematic understanding of the writing system as well). It’s a bit more detailed than I would recommend for the “average” student, but then again, that doesn’t matter. When it comes to memory and especially mnemonics, we’re all different. Your method might not work for somebody else (perhaps someone less interested in characters in themselves), but that’s not a problem for you! I think it sounds like you’ve found a great approach. I think it fits pretty well around level 4 according to the article.

As an avid Outlier Linguistics dictionary user myself (and a bit of a nerd with researching character etymology often since it’s the main time I really enjoy doing!), I feel like I stand between the levels of 4 and 5 here since I just prefer to remember the characters I come across by logic and their real etymology rather than come up with a made up story myself. And because I guess I’m an advanced learner who wants to take learning Chinese characters seriously and know all of the real roles of each component.

Some examples I usually make for myself that I have of this are (and taking some characters from the article too):
– 美: A man (大) wearing sheep horns (羊): beautiful (I know 羊 is an empty component in this character but I personally like to use this mnemonic since it helps me remember the shape more.)

– 藥: Something made out of herbs (艹) that makes a person happy (樂) again: medicine (I know 樂 is sound component here but again, this is to help me remember it’s shape and the sound component this character has)

– 活: Water (氵) in the mouth (𠯑): alive, to live (Here, I’m aware that the sound component 𠯑, which is certainly commonly written into 舌 often, is used here so I use it instead of 舌 (shé) because I know that’s the real sound component)

As for some phonetic loan characters, I would usually try to connect the original meaning and the modern meaning together together in a logical way in something like this:
– 其: picture of a woven basket made out of bamboo
Original meaning: basket made out of bamboo (now written as 箕)
basket > ∆carries object inside > ∆object in a basket shows ownership > his, her, its
(Here I use ∆ in my system to show that it’s not a real meaning but rather to indicate some sort of logical connection between the original meaning and the modern meaning)

And yeah, that’s how I usually memorize characters learning their real etymology and maybe occasionally me coming up with some logical mnemonics that is related to the character itself. Thought I wanted to show you some examples to see if I’m doing this thing correctly at all 🙂

1. “For characters, this means looking at a serious explanation of it (not just a graphical breakdown)”.

By ‘graphical breakdown’ you mean Level 1 depth, or otherwise just trying to remember the components and their orientation within a character without understanding semantic or phonetic meanings, correct?

Yes, I refer to tools that analyse characters only based on what they look like.

2. You mention: “most characters contain components that are there for no other reason than to indicate pronunciation”. Is the SmartHanzi (sourced from EDHCC) etymological breakdown (that both character components have semantics, with one also hinting at phonetic) an accurate depiction (or otherwise inaccurate but helpful)? I see that Smart Hanzi lists the top of 看 as a hand. So, it seems a SmartHanzi-derived mnemonic would fall under Level 2 depth of understanding (except that it leaves room for phonetics)?

I’m struggling to word exactly how, but it seems this question has implications for which mnemonics/what strategy to use for learning characters.

It’s hard for me to verify any single resource, as I’m not an expert in palaeography, but I always recommend using the Outlier Character Dictionary if you want more correct information.

3. “You split the character the correct way, but you are okay with making stuff up if necessary, and when you do so, you do so consistently.”

If I understand then, to be consistent, hypothetically if 看’s corrupted phonetic component did show up elsewhere, I should then treat it like a hand(?). Or because it doesn’t show up elsewhere, this question is irrelevant and I don’t need to worry about consistency in this rare case.

Yes

Maybe this is where I’m confused, and where the distinction between SmartHanzi’s etymological character component breakdown and the conventional (1 semantic, 1 phonetic) breakdown might clear things up(?).
Indeed, if we used a different character whose ostensibly phonetic component is used in other characters, SmartHanzi’s phono-*semantic* definition of that phonetic component would theoretically be worth remembering(?). But, perhaps this then just becomes a similar strategy to what you advocate – applying a meaning (be it visual or solely ‘logical’) to both char. components (e.g., eye + hand 看), even if one mnemonic is factually incorrect (the hand).

Yeah, this seems to be inline with what I’m saying.

Albeit, thereafter it seems we return to the question of which mnemonic is most effective. This brings up 2 quick last questions. In the above case, we are applying made-up meanings to phonetic components when that is more helpful for remembering the character than trying to link the phonetic component to the character’s exact pronunciation ([1.] which would almost always be the case?). And, 2., is this Level 3 depth?).

I’m not entirely sure what you’re asking here, but in my opinion, remembering pronunciation is much easier than remembering how to write the character. I rarely, if ever, rely on mnemonics to remember how a character is pronounced. That knowledge will also come more or less automatically as you learn more characters. So, using your words, I think that the phonetic component is (usually) already linked to the character’s pronunciation; you don’t need a mnemonic for that. If the link is extremely weak, as in 他 for example, I would just ignore the phonetic information and memorise it as if it were a meaning + meaning compound. Level 3 would mean that you do this, but you’re aware that this is for memorisation, not actual etymology.

Apologies, seems I’m unable to comment within comment threads (maybe it’s Safari).

1. “For characters, this means looking at a serious explanation of it (not just a graphical breakdown)”.

By ‘graphical breakdown’ you mean Level 1 depth, or otherwise just trying to remember the components and their orientation within a character without understanding semantic or phonetic meanings, correct?

2. You mention: “most characters contain components that are there for no other reason than to indicate pronunciation”. Is the SmartHanzi (sourced from EDHCC) etymological breakdown (that both character components have semantics, with one also hinting at phonetic) an accurate depiction (or otherwise inaccurate but helpful)? I see that Smart Hanzi lists the top of 看 as a hand. So, it seems a SmartHanzi-derived mnemonic would fall under Level 2 depth of understanding (except that it leaves room for phonetics)?

I’m struggling to word exactly how, but it seems this question has implications for which mnemonics/what strategy to use for learning characters.

3. “You split the character the correct way, but you are okay with making stuff up if necessary, and when you do so, you do so consistently.”

If I understand then, to be consistent, hypothetically if 看’s corrupted phonetic component did show up elsewhere, I should then treat it like a hand(?). Or because it doesn’t show up elsewhere, this question is irrelevant and I don’t need to worry about consistency in this rare case.

Maybe this is where I’m confused, and where the distinction between SmartHanzi’s etymological character component breakdown and the conventional (1 semantic, 1 phonetic) breakdown might clear things up(?).

Indeed, if we used a different character whose ostensibly phonetic component is used in other characters, SmartHanzi’s phono-*semantic* definition of that phonetic component would theoretically be worth remembering(?). But, perhaps this then just becomes a similar strategy to what you advocate – applying a meaning (be it visual or solely ‘logical’) to both char. components (e.g., eye + hand 看), even if one mnemonic is factually incorrect (the hand).

Albeit, thereafter it seems we return to the question of which mnemonic is most effective. This brings up 2 quick last questions. In the above case, we are applying made-up meanings to phonetic components when that is more helpful for remembering the character than trying to link the phonetic component to the character’s exact pronunciation ([1.] which would almost always be the case?). And, 2., is this Level 3 depth?).

Hi! You bring up some good questions here. In general, understanding is the key to remembering anything, so the first thing you should always do is to at least try to understand what you’re trying to memorise. For characters, this means looking at a serious explanation of it (not just a graphical breakdown). For words, it means understanding how word formation works, which will solve almost all casse for you. If you haven’t already, I suggest that you check out my series about the building blocks of Chinese, especially the part 5 and part 6, which are both about words. In most cases, no mnemonic is needed. Most people don’t need a fancy story to remember that “fire” + “vehicle” = “train”. At least for me, making sense of the construction is enough.

This is true for single characters as well, but it requires a thorough understand of how they work. The issue with characters, though, and this is the topic of this article, is that oftentimes, the true etymological explanation of the character is not helpful. I don’t think what you describe is on level 5, that would imply that you cut no corners and never invented anything. I like using the example of 看 to explain this. Most people, including native speakers and teachers, think that this is a hand shading an eye. It’s not. The top part is actually a corrupted phonetic component that is not used anywhere else as far as I know. Now, it would, in my opinion, be a mistake to ignore the hand-shading-an-eye visualisation, because that’s so easy to remember and makes so much sense, even if it’s wrong. This is the level 3 I advocate in this article: You split the character the correct way, but you are okay with making stuff up if necessary, and when you do so, you do so consistently. Level 3 also implies that you don’t make a point of remembering that the top of 看 is actually not a hand and not even a meaning component, because the goal is to remember how to write the character, not learn the origins of all characters.

I don’t think this fully answers your question, and if not, feel free to ask a follow-up question based on what I wrote! I’ll consider addressing this in an upcoming podcast episode, but it’ll need to be in a somewhat simplified form.

My struggle is likely partly b/c I don’t fully understand how best to create a mnemonic (I’ve read your great articles).

My Current Strategy: if I fail to remember a word in Anki, I check SmartHanzi’s component breakdowns (drawn from an Etymological Dictionary that breaks characters into one phono-semantic and one semantic component). I guess this is ‘Level 5 depth’. DongChinese is my backup (it’s breakdowns differ; matching your teaching: characters have one semantic and one phonetic component).

My (untested) impression is that creating a “convoluted” story with many details (that don’t *logically* follow each other), might be difficult for me; I’m certainly overwhelmed by the prospect of having hundreds(+) of these stories, and the prospect of initially setting them up poorly vis-a-vis each other.

Whereas, if this dictionary’s authors are correct: that 94% of characters have one phono-semantic and one semantic component, then I can use logic (something we intuitively understand) to deduce character meanings, without creating anything additional (story or otherwise).

Yet, I assume a story often is more memorable, and that the meaning of characters is not (indeed often isn’t) straight-forwardly deduced by a learner who knows component meanings, (this is evident in the dictionary’s entries).

So perhaps the upfront work is more intimidating to me (inventing vivid stories, and doing so correctly vis-a-vis each other), but long-term retention is better. Whereas, I guess what I (and maybe Fearchar) are currently doing is relying on our existing understanding of logic/rationality to piece together the semantic and phono-semantic (actual/etymological) meanings of components.

E.g.,

1. ‘上’ often means ‘on/above’, and

2. ‘坡’ is made from ‘土’ ‘earth’ (Semantic) + ‘皮’ ‘align’ (Phonosem.);

together depicting: “aligned hillsides, slopes or embankments”.

So, granted I memorize the components (which is possible I think right(?); i.e., there aren’t too many),

my mind doesn’t need an additional unrelated story to link these two components/to realize the meaning of: ‘上坡’ (uphill; upslope; to move upwards, etc.). i.e., logic does that linking for me.

However, this linkage/realization is probably very difficult, because it’s not always crystal clear how the two components connect to create the word meaning (a comparatively small number of components cannot cover the hundreds of thousands of words/meanings), and (I think?) each component can mean multiple things.

And to your question, I guess the only way a visual story differs from this process is that there is less that is being created/invented (b/c 1. component meanings are memorized/known, 2. logic is intuitive to humans).

As you can see, I’m still unsure what strategy is better, but probably it’s your story/mnemonic method.

Appreciate any insight and advice you have, this has been bothering me a while.

1. The use of stupid, helps to…,in what way
2. …It’s far less problematic to invent the story about Noah and his ark to remember to write 船 than it is to try to use that story to prove that the ancient Chinese were actually Christians. (Ancient Chinese were actually Christians), The use and making of this character was in what year, does it pre-date Christianity?

Yes, what I meant was “written exactly like 羊 is written when it’s the top component in other characters and it actually is 羊”!