MEANING: A woman is holding a child while saying: “You’re a good boy, you’re a good boy…”
PRONUNCIATION: The good boy is Harry Potter, who’s just been abandoned in the cellar of a dark barn after Voldemort killed his parents.
This example really ties together the points you were making. I understand now what you mean about going from component triggers → meaning → pronunciation!
I realized yesterday after typing up my post(s) that the pronunciation will indeed need to be related to the mnemonic. Not only for the ease of recall via association, as you point out, but also because there are many characters with the same meaning (e.g., “see” or “look at”). If we have an in-isolation pronunciation mnemonic that has no relation to the meaning/components, then we would have multiple stories with characters that are all about “seeing” and have no relation.
I really quite like this method. In reality, it is more or less what we do today already - in terms of total information - but we just have the ordering out of whack.
I’ll give this one some deeper thought, its advantages are immediately apparent to me.
That’s a good summary. The holy grail is an ad-hoc system that carries complete information. I’ve searched and thought about it, but haven’t been able to come up with one. Just the toneless phonetic pronunciation of some syllables in pinyin simply do not have any English equivalent for us to lean on. For example, I can’t think of any words that could differentiate between the following effectively:
jiao
xiao
qiao
shao
chao
zhao
Even if we can encode the “tone” into the mnemonic with random numbers, or even a system like the one we have today, this fundamental inability to have a unique mapping of these sounds to English words is a big roadblock. This is why HanziHero does the approach where sometimes our sound associations are based on the spelling or simply a bit arbitrary.
One other question I have about this mnemonic-reordering insight you have.
MEANING: A woman is holding a child while saying: “You’re a good boy, you’re a good boy…”
PRONUNCIATION: The good boy is Harry Potter, who’s just been abandoned in the cellar of a dark barn after Voldemort killed his parents.
It seems like the main point is having a better ordering of elements within the story, as you say:
components → meaning → pronunciation
What do you think about having the ordering like you have here, but still as a single mnemonic:
A woman is holding a child while saying: “You’re a good boy, you’re a good boy…” The good boy is Harry Potter, who’s just been abandoned in the cellar of a dark barn after Voldemort killed his parents.
In other words, what do you view as the main advantage of splitting the mnemonic into two separate fields/sections, when they are still related? Or do you think that the main thing is just the ordering, and whether it is split up in our UI into two or kept as one is immaterial? I can see the advantage of either approach (one vs two) but wanted to hear your thoughts.
We have the same problem in Japanese too, and in these cases, if you can’t find a word that sounds the same, you either break it up into multiple pieces until it does, or you base the association on spelling instead, like you say. In all cases, they only have to be good enough, to bring you back to the actual pronunciation.
For your examples (remember, I don’t truly know the reading or pronunciation), just off the top of my head
I’m competing in a jousting (jiao) tournament this weekend, and I hope my horse is in good health.
x’s and z’s are tricky, but this is how xiao is broken down on a pronunciation dictionary: - Phonetic Spelling:[ sh ih ow ]
The new yorker said “sure, I’m goin to tha show (xiao)”
Or you could break it up and say
she yelled ow after hitting her thumb with a hammer
close enough in my book:
The Italian waiter said to my wife, ciao (qiao) bella and I got insanely jealous.
I think most of us know this Italian across all languages, right?
The shaolin (shao) monk prayed in his temple all winter.
I’m starving, let’s get some chicken chao mein or I’m hungry let’s get some chow
I told him a really terrible joke (zhao) about the guy with the big, sagging jowl
That last one I heard was pronounced two different ways (both without tones) (?)
Anyway, they’re all good enough in my book, and they’d stick in some mnemonics for me. I’m also very sure better ones could be created with a bit more thought
It could still work as a single mnemonic, but I feel splitting it into two makes it more digestible to the brain. Splitting would also give you a bit more flexibility if there is a case in which linking the two stories doesn’t make sense (which can happen in some occasions), and in all the characters that are one single component.
Just the reordering in a single one would make a world of difference though!
Surely! These examples are better, but I still find it very ineffective to have to introduce a number in every single mnemonic and a good reason why it’s that number and not one of the other four.
I’m still using my own system of character+location for a fair number of pronunciations though, which is why I’m arguing the, even though it requires a bit more buy-in in the beginning, for me it works better than the ad-hoc system from WK.
But of course it’s very logical that we all have different ways that work best for our brains!
If you mean having a separate meaning mnemonic and reading mnemonic, then from my view it’s less information all at once, though, this assumes that your showing them in their own separate cards.
Meaning “card” only relates to component association with the character to get to the meaning
the reading “card” associates the reading with the meaning.
It just makes everything easier and kind of fits real life, I’m constantly told by people who are fluent in other languages that they strain to translate from the fluent to the native on the fly as the two occupy “different parts of the brain” I completely understand this point of view now from my own learning where I didn’t quite get that before, kind of like your looking at it from the natives perspective. You don’t even hunt for the english meaning after a while and just instinctively fill out the reading without acting like google translate
I think you can get away with a single mnemonic when the alphabet largely remains the same, anything else, I feel needs split
Good point. But I think this is an issue with 1a and 1b equally depending on the specific mnemonic. Remembering if some contrived story took place in a bathroom or in a basement is pretty much as hard as remembering the tone. But if the story necessitates a certain location for it to make sense, then it works. For ad-hoc mnemonics the same thing could apply. Shoehorning 2nd of July is not useful. May I suggest
I went to Hong Kong (hóng • hong2) and saw swan loveboats filled with couples in the harbor
So there are a few things conflated here
the pronunciation of “the letter b” (which in English is “bee”) [1]
the utterance people choose when expressing the sound a “b” makes (“buh” in English, “boa” in Chinese) [2]
the intrinsic sound/set of sounds a “b” contributes to a syllable [3]
Users need to learn [3]. The issue is you physically cannot produce [3] without including some vowel, voiced fricative, etc. which then makes it just an alternate like [2]. Maybe one thing they could add the entire pinyin table row (“bi”, “bin”, “bo”, etc.) so that learners can intuit [3].
As an aside, you should look at the phonology/pinyin page on wikipedia. “qi” is like the “ch” in English, but the tongue is further forward in your mouth and “palatized” (the middle is raised toward your hard palate like you’re about to say “eeee”). This btw is true about the equivalent sounds in Japanese. An additional note is that in Chinese it only is followed by the vowel “i”, and that’s why in the lesson you’re being taught “qi”.
Pinyin “ch” on the other hand is like English “ch”, but the tongue is further back in your mouth and curled up with the tip close to the roof of your mouth.
Yeah I feel that. I think
is a decent strategy. Use orthographically similar words in the mnemonic.
Another thing you could leverage is the fact that the retroflex series (“zh”, “ch” “sh”) and the palatal series (“j”, “q”, “x”) are in a completely contrastive distribution. The palatal series is only followed by the medials “i” or “ü”, and the retroflex series is never followed by the medials “i” or “ü”.
So words like “jeans”, “cheese”, “sheet” uniquely signal the palatal series. Unfortunately there is no good way to approximate palatal + “ü”, without confusing it with retroflex + “u”
Got it. Yeah, I can see the split being good for the case where the two stories are only marginally related and having them in the same “block of text” may do more harm than good.
I discussed this split-approach (along with the re-ordered mnemonic elements) with my brother. We are still digesting it and may try it out internally for a bit. Of course, with over 4k+ mnemonics by now, replacing them will take a while if we go down this path, but if we find it is definitely better, it’ll be effort well-spent.
One other advantage is that it is easier to teach, I think. Even in the videos I made about our approach, I found it cumbersome to need to cover all bases before covering the mnemonic. This way the user can focus just on the meaning part and just on the pronunciation part separately, since either “stands on their own” in a way.
By the way: we are working on adding a simple note field to the quiz item info and item pages which you should see in the next week or so. We’ll have an announcement when it is launched with more details.
Thanks for your thoughts! Yeah, I can see the benefit of the split approach.
I can see the advantages of the phonetic-approximation approach, but it also has its own trade offs. I can see us working towards a split mnemonic and changing/simplifying the mnemonics elements in the story first before to see how much of an improvement that is for everyone before going down that path.
We are also adding a note field within the next week or so to allow people to easily write any personal notes or mnemonics for when ours don’t fit their learning style. I think regardless of what mnemonic style we use, it will never be able to work for everyone, so this is an important feature that we are excited to get out soon.
Yep, for sure. I agree that “x in y of z” by itself without conscious integration of other elements relating to x y and z is only marginally better than brute force memorization of e.g., tone numbers.
What I’ve been trying to do for later mnemonics is to incorporate unique aspects of those settings/locations in the story to make them feel less arbitrary. E.g., the bathroom of the barn will have manure, and the cellar would have milk jugs or farm tools and so on. Or how the “inside” of the “highschool” is really a hallway with lockers. Mnemonics I made earlier may not have this improvement, however.
That’s awesome!
Yes, definitely it’s going to take some time if you decide to do it in the end, I assume it will take some careful consideration.
And it’s great that we’ll get that note adding feature meanwhile, can’t wait!
I posted a new thread which has a proposal for splitting our mnemonics into meaning and pronunciation mnemonics for characters.
I’d love it if those who participated in this thread can give their thoughts over there!
Of course, our new approach doesn’t address all of the mnemonic feedback in this thread, but we think it is a step in a right direction for most folks and want to confirm that others feel the same before proceeding.
