From a complete beginners perspective

I did recently let some of these thoughts in chat known to one of the devs and they were super nice, but I think I really failed to demonstrate the point/s I was actually trying to make and perhaps emphasise some things which may not be clear to the devs and more advanced learners who are already much further along in their studies than me, which may prevent those from seeing the wood through the trees :slight_smile: That, and Iā€™d love to know what other peopleā€™s thoughts are about what Iā€™m about to say.

So initially, following registration (and to note, Iā€™ve just done this again to verify my initial feelings), I click Lessons. I got told about lessons, etc then it says the following:

Sound pronunciation

This text shows the pronunciation of the current sound item. It uses pinyin to spell it out. If you arenā€™t familiar with pinyin, donā€™t worry - youā€™ll learn it as you go through the course! Youā€™ll learn more about this bit of pinyin as well as be able to hear it sounded out in the Pronunciation tab.

Fine, Iā€™m a user who knows nothing about Chinese pronunciation or Pinyin but great, weā€™ll learn more as we go. Iā€™m presented with the first ā€œsoundā€ which is shown as ā€œB-, Bart Simpsonā€ I hit the audio so I can hear, and to me, it sounds like buo or boaā€¦ but Iā€™m then told in the pronunciation instructions that

The ā€œb-ā€ sound in Pinyin is pronounced the same as the ā€œb-ā€ sound in English.

Uhm, ok, maybe. I mean, it doesnā€™t at all, but Iā€™ll play along. I hit the right arrow, and Iā€™m told about association and that you want me to associate this sound with ā€œBart Simpsonā€ ā€¦ ok, I mean, it sounds nothing like ba-rt simpson and doesnā€™t have even remotely the same spelling as the letter your showing or my romanized internalisation of the sound (boa in my head). ā€¦Again, I have no prior experience with Pinyin . Now Iā€™m also starting to wonder, ā€œWhat does the hyphen represent next to the b?ā€ ā€¦ anyway, Iā€™ll do as you say and associate ā€œb-ā€ with ā€œBart Simpsonā€ - continuing on.

Next we are on -a, and itā€™s the same as above, but this time Iā€™m wondering to myself why the hyphen is now in front of the ā€œaā€ . hitting the arrow and to your association instructions, and you mention the following:

The ā€œaā€ in the English word ā€œauditoriumā€ sounds similar to this final

Now, Iā€™m thinking to myself, What the hell is a ā€œfinalā€ ?? :slight_smile: Continuing on.

Next, you show me a ā€œ1ā€ and you say the following:

This tone is high and flat. It doesnā€™t necessarily mean it is extremely high, just relatively higher than the other tones.

Iā€™m thinking, ā€œhigh and flatā€ relative to what? There is no audio and no examples or description of whatā€™s actually happening here. Is 1 the highest and 10 the lowest? or does 1 represent something else? Is 10 actually the last number, or is it 6 or 8? But if 1 isnā€™t ā€œextremely high,ā€ does that mean there is a zero (again, Iā€™m playing dumb a bit here to emphasise my point) - Continuing on.

Ok, now we are at something that is more in my wheelhouse, components/radicals - I donā€™t think Iā€™d be a good judge of character here as I come from Wanikani and already know whatā€™s expected and the answer.

Itā€™s worth noting that I could not find a way to get the tooltips (sorry, I donā€™t know your terminology), but the initial highlighting and descriptions describing the lessons and navigation back after the initial viewing, perhaps an oversight?

So on the quiz, your showing me auditorium, and I put in ā€œaā€ and it accepted my answer without the hyphen, so I suppose Iā€™m left to assume the hyphens donā€™t mean anything :wink: I mean wanikani would give me the shake and warning if it did.

Continuing on to the next set of lessons, and am greeted with this

In our modified pinyin chart, there are some syllables that have no final, which we refer to as a ā€œnull finalā€. For ease of typing, we represent it as ā€œ-_ā€ ā€¦

Ok, So again, this is all meaningless to me, but it seems to suggest I need to go find this chart of yours on your site somewhere?.. but I thought we were going to learn as we went along as mentioned earlier (?)

Anyway, so I finish all the available lessons, and Iā€™m still none the wiser as to what it is I actually just learned :man_shrugging: - I canā€™t complete any more lessons to gain more clarity, but I can go over to my user settings and change my daily limit to 20. Okay, I understand why you might want to coddle new users and not allow them to overwhelm themselves, but as mentioned, I havenā€™t currently learned anything, and 20 is kind of leaning on the side of too lowā€¦ I come from an environment where I learn 100 items a day (sometimes I change this depending on the day), so perhaps you should consider increasing the initial item count and also have an ā€œadvanced settingsā€ checkbox or something, and if enabled, allow the user to specify their own number so they can set it to whatever they feel comfortable with.

Anyway, now for the next 10 lessons, which Iā€™ll breeze through. ā€œpouchā€ component is just the same as its main character over on Wanikani, ie: east, so to avoid confusion, Iā€™ll just change the synonym and enter ā€œeastā€ here too, but unfortunately, I canā€™t go ahead and modify or add my own mnemonic to match my association. The dev told me this is a good idea to be able to do this, so :crossed_fingers: this will be added at a later date (perhaps you could even allow other users to view other user-created mnemonics and up/downvote them and add them to their account for the given component/character) Itā€™d also be nice to have the ability to insert an image to reinforce a text-only mnemonics, which Iā€™ve found valuable elsewhere.

Ok, now that Iā€™ve maxed out all my lessons for the day, I canā€™t do anything further, given that I donā€™t know what all that audio and pinyin stuff was all about (again, Iā€™m playing this up slightly :wink: ) and given that I knew what all the components were from prior knowledge from over at Wanikani, Iā€™m kind of left feeling a bit lost.

I head over to the ā€œCurriculumā€ tab and select characters. Itā€™s interesting you havenā€™t grouped these per rank or level as shown back on my dashboard, or am I to assume that every group of 50 represents a rank/lvl?

The first thought is, ā€œChrist, I know pretty much all the meanings of every one of these up to a very high level, so Iā€™m going to be stuck learning very little for a long time.ā€ The dev told me they are going to address this, which is great (hopefully not just flat out skip or delete of the item entirely or forever, as I wouldnā€™t mind being reminded of it in a year or some amount of time).

Anyway, I click on the character for mouth, and now Iā€™m starting to get a bit of a better understanding of the audio associations we covered, but this is where I think Iā€™m really going to struggle on this platform, unfortunately, and it breaks my heart to say, as I can see youā€™ve spent a lot of time incorporating the audio/pinyin associations into the mnemonics that youā€™re using for the character/s. The problem, as I see it, is that youā€™ve overcomplicated things. For a single character, youā€™re now requiring me to remember four things to remember one.

[k-] Kermit the Frog is chilling in the [3] basement of the [-ou] bakery, his 口 mouth covered with flour that he has been trying to eat. Of course, he is a puppet, so each handful of flour he sticks in his mouth just hits it and falls right back out.

Why is this a problem? because this is going to hinder and not help, especially long-term, and basically nullifies all the positive aspects of using mnemonics in the first place (again, this is going to be hard to see for those of you who already know the meaning / created the mnemonic), and itā€™s a problem which is only going to condense as time moves onā€¦ Iā€™ll explain as I go on why this is the case.

Now if i was over on Wanikani (and for the sake of argument, just pretend the sound is the same in both Chinese and Japanese), the menmonic would probably be something like this.

The laughing cow had a really big mouth blah blah blah

(cow being the associated sound/reading and mouth being the meaning)

No additional fluff. Sure, Iā€™m not going to learn the tones based on this, but itā€™s actually a useful mnemonic that is actually going to help and not hinder my association with the character, and the tones will be drilled when we get to the quiz and I have to type it out in pinyin and subsequently drilled in reviewā€¦ Better yet, when we actually encounter the character or meaning in media and real context (not some robot audio voice on the platform). Which is obviously going to be far quicker / possible using my mnemonic as opposed to your one.

Other issues with your methodā€¦ Itā€™s great to have familiar scenarios, people/char/s and associations in mnemonics, but they are going to come up several orders of magnitude more compared to the ones over at wanikani just because of the nature of Chinese and because of the way your incorporating the sounds and associations into the mnemonics. This is going to lead to complete confusion and frustration (I already know this and I actually havenā€™t even done a character yet in the lessons). For example, how many stories am I going to have involving ā€œKermit the Frogā€ etc at any one time or in my stacked-up reviews? Wanikani already has a difficult enough time balancing things as it is ie: rearranging similar meaning characters, menmonics with a similar premise, characters of different difficulty levels, characters that look incredibly similar, being too close to one another, etc. I expect that as you get more beginner-level users using your platform, youā€™re going to have a hell of a time trying to restructure things to avoid this.

As an example, letā€™s pretend Wanikani was doing the same thing as you. Youā€™d start, and theyā€™d force you to learn Hiragana, Katakana, and Romaji (take your pick, or just imagine itā€™s all three), regardless of whether you knew them already or not (from my perspective, these would be the analogue of your audio/pinyin lessons), and theyā€™d likely teach them as they would on Tofugu ie: ā€œć›ā€ ā€œseā€ then menemonic ā€œsexy vampire teethā€ (I forget what actual menmonic they use). Theyā€™d then force you to learn pitch accent alongside with no context or explanation (:stuck_out_tongue: I know Iā€™m joking) Then a few radicals, then kanji, butā€¦ Right so now we are learning some kanji, but instead of just associating the entire sound with the character in question, theyā€™d break it down into each individual Hiragana, Katakana, or Romajii (take your pick) and incorporate all the potential sounds based on the mnemonic associated with them and incorporate this mess all into a story so I can learn this single character. (I hope you see why this is an issue.) Now, how many stories are we going to have involving the escapades of our sexy vampire teeth in the coming weeks or months? And when confusion and frustration sets in, what am I to do :smiley: And remember, this is just one out of the 100ā€™s or 1000ā€™s and doesnā€™t even take into consideration confusing pinyin with other pinyin (roman letters with the same roman letter etc) and, in turn, their respective associations.

I mean, I get it. The trouble here is coming down to the tones, but I think youā€™re going to tie most users in knots with your implementation and tie yourselves in knots trying to resolve it as time goes on. Iā€™d like to hear other peopleā€™s thoughts on this, but from my perspective, Iā€™d isolate the audio/pinyin stuff from the main lessons, give it its own insulated section, and strip out all the " Sound Composition" stuff from the characters mnemonic stories.

I also agree with this topic over here.

And Iā€™d say basically just make a more interactive optional section that is like Tofuguā€™s hiragana, katakana, and romaji guide, but of course using Bopomofo and Pinyin. I think there is a real value in first associating a sound with a shape and then remembering the shape and in turn the sound with a menmonic (which the tofugu guide does so well with the image overlays), and then this makes learning other forms easier just by virtue. At least thatā€™s what I found when learning hiragana, and then following learning the sound/pronunciation and subsequently katakana and romaji was easier to learn as a result. Basically, what Iā€™m trying to say is, Iā€™d imagine if you taught me Bopomofo in the same way Tofugu taught me Hiragana, then Iā€™d understand Pinyin much more easily and quickly.

Anyway, I know this is a giant post, and I know you guys are aware of some of these issues already. Overall, the site is super slick and I like it, you guys seem sound and this was purely just constructive criticism and an insight into what a newb will see when using :slight_smile:

Thanks for the feedback!

ā€œb-ā€ in ā€œBart Simpsonā€

You are right that the audio does not match exactly - our descriptions were made before we added audio. Weā€™ll update the descriptions to better match the audio. The mismatch here is explained by the fact that ā€œb-ā€ pronounced by itself in English or Chinese is almost inaudible. Itā€™s just a smacking of the lip. The English unconscious convention is to append an ā€œaā€ vowel when pronouncing it by itself. The Chinese unconscious convention is to append an ā€œoā€ vowel when pronouncing by itself. In reality, the vowel sound is not what we are referring to and not an essential part of ā€œb-ā€.

hyphens

The hyphens indicate whether it is an initial or final. Weā€™ll look into adding an onboarding flow that better explains what each of those are.

tones

I agree that we can have some little audio widgets for introducing tones. We could have the syllable ā€œmaā€ for example pronounced with each of the four tones, allowing the user to press each to better understand the difference. A tone chart like this one would also help.

could not find a way to get the tooltips

Are you referring to viewing the lesson once more before going to the quiz section? You can do that by clicking on the indicators at the bottom of the screen. There should be an icon for each item you have learned. See bottom of picture below. Perhaps they were scrolled off screen? Moving them to the top-right may be better.

null pinyin

Yes this could be better explained for sure.

w/r/t mnemonics

Let me first tell you my own assumptions I had when making HanziHero which still exist today:

  1. The goal of HanziHero is to teach the meaning and pronunciation of each character.
  2. Chinese has a foreign phonological system that includes tones, the majority of which have no English equivalent.
  3. In order to fully know a Chinese character, one must remember its meaning and pronunciation. One without the other is useless for fluency.
  4. Mnemonics are the best way to remember the meaning and pronunciation of a character.

If we mutually agree on the four premises above, then Iā€™m not sure where the issue with having pronunciation information in the mnemonics would be a problem. There are these approaches to teaching pronunciation with a system like HanziHero, after all:

  1. Donā€™t teach it. Not an option, see (3) above.
  2. Say it, but have the user brute-force memorize it. Not ideal, see (4) above.
  3. Have a separate pronunciation mnemonic, like WaniKani. This is the same amount of information required to remember, so does not make a material difference. We cannot use the English-approximation mnemonic approach that WaniKani uses because of tones and foreign phonology, see (2) above.
  4. Have a single integrated mnemonic that teaches both meaning and pronunciation, which is what we do.

Can you let me know what you picture as the best way for learning pronunciation information for each character?

You can learn more about our method in this video. Perhaps it should be integrated into our onboarding flow.

1 Like

For pinyin specifically, I can see us having a ā€œPinyin in 100 secondsā€ video that can be part of our onboarding flow! I fully sympathize with all of your points around pinyin. It is really tricky for beginners and the way we teach it does indeed assume that users already know it. We say we will help users learn as we go through the course, but as you point out the information we teach and the way we teach it is not sufficient, for sure. Iā€™ll work on that as my next video. :bulb:

1 Like

I will chime in about the mnemonics for pronunciation topic, since I was already thinking about it as well.

I love what the site provides to me and I see myself continuing to learn all the characters here, but personally Iā€™m barely using any of your mnemonics and I doubt I will in the future: I prefer to create my own.

This is because of two points that @jiminyd1234 touches on too:

  1. The main one is having the meaning and the pronunciation in the same mnemonic. For my brain this is too much information to recall in a single story. Also, the story in the mnemonic starts out with the character & location for the pronunciation, but I need to get to the meaning first to be able to recall that. Because the starting out input is the character Iā€™m seeing on the screen, the logical order for a mnemonic should be: Iā€™m seeing components that I recognize and lead me to the mnemonic for a meaning. From that meaning, I can recall a story involving a character and a location that leads me to the pronunciation. The two parts can be related or not, whatever seems simple in my brain when I think about it.

  2. The second point is about using the system of character & location for initials, finals and tones instead of how WK creates a story with a similar sound in English (kou->cow example). Personally, I do like this system and I think it fits better to Chinese than the WK approach. WK kind of gets away with it because Japanese has very simple sounds, but I always found it too random. Having too many stories about a specific character I doubt will become too much of a problem, since once you start gaining familiarity with the character you stop using the mnemonic anyway. Having said that, I think this method is truly powerful using my own set of characters and locations, since then I can use some that I know really well and make the mnemonics more memorable to me, which is why Iā€™m also not using the ones provided here.

  3. Also into the equation of coming up with my own mnemonics is the fact that I already know some of the components from their WK definition and Iā€™d rather keep using those than learning a new one when it differs, but thatā€™s just me.

In the end, I still think itā€™s great that the website gives us a framework to work with and suggested mnemonics if people want to use them, I just prefer to create mine.

1 Like

@kevin Thanks for the reply, confirmations, and clarifications. +1: Hopefully I didnā€™t come across as too harsh. I know these things take time, and things are evolving. Just trying to give you my raw perspective and first impression and also to possibly show if Iā€™m struggling (and thatā€™s probably with a 50% advantage due to my knowledge of Japanese), then Iā€™d imagine people with no leverage advantage will probably double struggle, and Iā€™d imagine them just giving up.

@damia Thanks for providing your perspective. Itā€™s good to see other peopleā€™s processes; I know we all learn in our own way :slight_smile:

I think we have to separate out the ā€œreadingā€ of a character vs the ā€œpronunciationā€ of a characterā€¦ Wanikani does not actively teach us the pronunciation (or pitch accent); they rely on us inferring the pronunciation based on us looking at the hiragana (or whatever), typing in the hiragana (or whatever), the audio, and the pitch accent markers (if thatā€™s your thing, I generally donā€™t even bother looking).

I largely agree on the four points you made, but Iā€™d say if we are using Wanikani as a point of comparison, then I know people who use it purely for input and/or use it primarily to gain input and learn and apply output (usually at a later date) through other means (for example, maybe somebody goes to Wanikani because they simply want to understand kanji to help them better read manga, for example). Regardless, Wanikani does not force teach pronunciation at all, nor does it force teach pitch accent. Sure, it might provide the information on the ā€œcardā€ but neither is mandatory and requires no association anywhere to get it. - I guess you could argue "they do because hiragana (or whatever) is in the ā€œcardā€ , but again this is the reading and not the pronunciation (pronunciation, again, is inferred from the written form), and knowing hiragana and/or romajii is a mandatory pre-requisite in order to use Wanikani and is completely isolated from the main ā€œcourseā€ .

As I touched on, I think you should isolate the Pinyin (and Bopomofo if you want to implement that also) and strip all references to it from your main course and hanzi mnemonics. So users can get to grips with the system, tones etc and itā€™s written form independently (just like you would on Tofugu and then on Wanikani).

In my mnemonic example, this would get my brain 90% of the way there as is, and as I mentioned above, Iā€™ll get myself the other 10% of the way there based on the other factors mentioned. I donā€™t think you need to forcibly or actively go out of your way to do it.

Letā€™s go back to the mouth example (which is arguably very easy as it already aligns with a previously learned component). Hereā€™s yours:

So your Hanzi for 口 mouth is

then you have the ā€œsound componentsā€ as follows:

[ k- ] Kermit the Frog +[-ou ] bakery + [ 3 ] basement

Then you have your mnemonic.

[ k-] Kermit the Frog is chilling in the [3] basement of the [-ou] bakery, his 口 mouth covered with flour that he has been trying to eat. Of course, he is a puppet, so each handful of flour he sticks in his mouth just hits it and falls right back out.

You have presented the pinyin components and the ascosiations youā€™ve created, and even though you might call them ā€œassociationsā€, they are in and of themselves individual mnemonics (albeit not great ones, sorry).

So, now, in order to recall the reading and pronunciation, using your method for the character for mouth 口 I not only have to recall all three of the individual sound components (and the associated spelling of the pinyin) either by recalling the pinyin or by recalling the respective association/mnemonic to get me back to the pinyin. either/or

So, Iā€™m basically forced not only to go through this mnemonic inception :wink: but I also have to remember the absolute order of said association/mnemonics in the context of the primary mnemonic (which is completely separate). In total, Iā€™m required to recall at least (emphasis on the at least part) four pieces of information (there could be more). (And thatā€™s before considering other characters with far more components.)

In contrast, my example would be more along the lines of what I mentioned prior.

Hanzi for 口 mouth,
[You could keep the sound components here if you want; I donā€™t really mind.]
then the mnemonic

The laughing cow (kĒ’u ā€¢ kou3) had a really big mouth blah blah blah

In this case, one word association In one mnemonic, pinyin is there and inferred. There is no mnemonic inception at play here, no concern about ordering, and the information getā€™s my brain 90% of the way there, and you donā€™t need to needlesly lock yourself into some strange association/mnemonic ecosystem (which is what youā€™ve done).

In regards to overuse of association/mnemonic premises, which youā€™ll be fixed into, but wanikani (and my example above wonā€™t)ā€”for example, Wanikani will routinely associate 恗悅恆 with shoe, but they recognise that constantly involving a shoe in every mnemonic involving this reading / hiragana of the Kanji that use it, which are close together (in the LVL), would be confusing. They have the freedom to alter said association and instead say ā€œshoelaceā€ or ā€œshootā€ or ā€œchuteā€ or whatever, you see my point. - not that this issue is going to be as prevalent on Wanikani anyway (but will be in your case using your implementation).

As mentioned I completely understand we all learn differently but it does seem in your effort to simplify youā€™ve inadvertantly over complicated :thinking: but maybe itā€™s just my learning style which doesnā€™t fit which is why I love reading other peoples impressions :slight_smile:

I think that component breakdowns are great, the way that 藝 contains 埶 which contains äøø, 圄 and 土

Crazy pills, pills, mushroom (now Iā€™m in doubt actually :laughing:) and soil.

But when it comes to mnemonics, Iā€™m most of the time skipping them because theyā€™re overly complicated.

Hereā€™s one of my favorite ones that stuck with me from WK: äø€ę—„ is pronuounced ā€œtsuitachiā€ because on the first date you only get two touches. Thatā€™s all. Itā€™s a fuzzy mapping, but itā€™s memorable and good enough to get the point across.

Now hereā€™s one for ē“…:
[hu-] The Hulk is [2] inside the [-eng] English manor reception hall, which is lined with ē“… red ē³ø silk carpet. The Hulk wears some ē“… red ē³ø silk robes and carries a å·„ cross-beam (also ē“… red) on his shoulder.

$person is $preposition and $place, to me, is mostly just fluff that is impossibly hard to remember. The one upper hand Chinese has over Japanese is that the characters actually make way more sense. I remember it by "Itā€™s red silk, and itā€™s pronounced hong2 because the right side is gong1.

Sure, theyā€™re not a 1:1, but once look at the character, you can most definitely make a qualified guess.

é“» is hong2 as well, and you can easily take the first wild guess by spotting å·„, but the mnemonic is:

[hu-] The Hulk is [2] inside the [-eng] English manor where he is chasing a giant 鳄 bird. He holds a å·„ cross-beam in one hand as he runs after the é“» swan the velvet carpet in the large vaulted halls. The é“» swan keeps on ę°µ spraying poop back at him, sometimes temporarily blinding him, as it continues to fly out of The Hulkā€™s reach.

Woah! I wouldā€™ve said ā€œitā€™s a bird with a å·„ so it must be a hong2. And it splashes, so itā€™s a swan, or a goose, or whateverā€.

1 Like

Ouch, this is making my head spin a little just reading thisā€¦ Can you tell me whats going on in your head when you encounter another character which your learning or which comes up in your reviews simultaneously which also include ā€œHulkā€ in the mnemonic?

For me Iā€™d start crossing mnemonics, I mean I do it on wanikani occasionaly.

Thanks for the tips :slight_smile:

I want to clarify my usage of ā€œpronunciationā€, which can refer to two things:

  1. The literal pinyin (or any phonetic transcription) associated with a character.
  2. How to sound out that pinyin with your mouth.

We do require knowing (1), as does e.g., WaniKani.
We do not teach (2), nor plan to. We merely give a broad description of each pinyin sound for reference. Knowing those descriptions or how to form the sound with your mouth is not required.

The laughing cow (kĒ’u ā€¢ kou3) had a really big mouth blah blah blah

This does not work in practice, because ā€œkouā€ could be kou1, kou2, kou3, kou4, or kou5. You would need some information to capture that tone as well. It should be noted, and emphasized, that these are completely different pronunciations/readings with completely different meanings in Chinese. A common mistake for Chinese learners is to believe tones are ā€œsprinkled on topā€, but that is not how they are perceived by Chinese themselves. If I say kou4 when I mean kou3, I would not be understood, for example.

mnemonic inception

Made me laugh :laughing: . This is indeed how I have used this method to learn 3,700 characters. Effectively.

  1. Look at components, mentally verbalize them.
  2. Try to ā€œfindā€ the story that contains those components in my head.
  3. Use that story to get meaning and reading.

I think part of it comes from having a single mnemonic. Iā€™ll think a bit more about how a split mnemonic works.

Looking at component pronunciation clues is a great way to guess the pronunciation. One issue with it, though, is that the mapping is ā€œlossyā€. It merely narrows down the possible choices, but will not point to any definitive pronunciation answer.

In your example, ē“… is pronounced hong2 and å·„ is pronounced gong1. So to get from one to the other, both the initial and tone changes. If we assume that every character with å·„ in it will share the same final but have differing tone and initial, then all we have done is narrowed down the 1200 possible pinyin syllables (with tone) to 1 final * 55 initials * 4 tones = 220 options. However, in reality this is not true, as many characters with this same component have completely different pronunciations that do not relate to å·„ā€™s at all.

These pronunciation hints from components are great - sometimes a certain group of components will nearly always have nearly identical pronunciation across multiple characters! But it is hard to make a system out of because these clues are imprecise and full of exceptions.

my own set of characters and locations

I can see how that would be very useful. After all, this system worked well for me because I was the one who came up with all of these mnemonics using locations/etc that fit in my brain the best. I can see us eventually doing it. The main thing that needs to be figured out is the level of integration that is desired.

The simplest form of this is the ability to add a note/association to each sound. More advanced would be able to see that note/association whenever you hover over it in a mnemonic, or in the little ā€œcardā€ that shows up in the item info on the quiz.

This, combined with the ability to make oneā€™s own mnemonic (again, likely as a ā€œnoteā€ field) I think would help here. What do you think?

From that meaning, I can recall a story involving a character and a location that leads me to the pronunciation.

I see what you mean! For example, for 口 kou3, we could split it like so:

  1. The meaning of this is the same as its component: mouth (component + meaning only).
  2. Kermit is in the basement of the bakery making a cake that is the shape of a mouth (pronunciation + meaning only).

The one difficulty I foresee here is the scenario of very common pronunciations. Take shi4 for example. Here are five characters with that pronunciation:

  1. ę˜Æ is
  2. äŗ‹ matter
  3. ē¤ŗ show
  4. 士 solider
  5. äø– generation

If we focus just on this meaning-pronuncuation story, using the same characters/locations approach:

  1. ę˜Æ Shrek writes ā€œisā€ in goo-y swamp slime on the tile of the bathroom within the house.
  2. äŗ‹ Shrek is depressed in the bathroom of the house. He lies in the bathtub, thinking.- ā€œnothing mattersā€.
  3. ē¤ŗ Someone is showing Shrek the bathroom of the house that he is thinking of buying.
  4. 士 Shrek is dressed as a solider in the bathroom of the house.
  5. äø– Shrek is looking at a book of all the previous generations of ogres as he sits on the toilet in the bathroom of the house.

In reality, we have 23 characters in total with the pronunciation ā€œshi4ā€ within our curriculum! :face_holding_back_tears:

So we would have 17 more scenarios with Shrek in the bathroom of the house.

Here is what the meaning mnemonics could look like:

Then, expanding to the component-meaning mnemonic:

  1. ę˜Æ The sun uses a pointer to indicate what is.
  2. äŗ‹ Everyone looks at you wondering what is the matter, because you havenā€™t said anything for a while. The reason? You have a pen and a cross stuck in your mouth!
  3. ē¤ŗ You have a cool altar that you want to show everyone. People are sick of you always showing them your altar.
  4. 士 Same meaning as component: solider.
  5. äø– The generations on a family tree are represented by branches and leaves.

Iā€™m curious your thoughts on this. Is this more memorable to you than having it all in one? I can see the benefits and disadvantages of this approach. Likewise with @jiminyd1234. Does an approach like this address your concerns?

1 Like

Forgive me, because I donā€™t know pinyin I just put ā€œblah blah blahā€ so you could fill in the gaps so I donā€™t embarass myself :slight_smile: (not literally the cow saying blah blah blah :smiley: ) but from what youā€™ve described, if you want that in the reading mnemonic Iā€™d just do this.

Reading Mnemonic

The laughing cow (kĒ’u ā€¢ kou3) has 3 gold rings in itā€™s ear and has a really big mouth and wonā€™t stop blabbering on about cheese

The key points for me in terms of recollection is the cow and the mouth, my brain will take care of the rest.

Now for @Youyujuan example

Iā€™d probably change this to

Reading Mnemonic
Iā€™m going to Hong Kong (hĆ³ng ā€¢ hong2) on the 2nd of july to see the swans at the docks

I hear ya :slight_smile: but sometimes we can bias favour our own mnemonicsā€¦ I mean I made a whole spanish deck utilizing the wanikani style and there was a good chunk where I got the mnemonics I created fine but when my friend tried the deck he found a few of them overly complicated. In retrospect yeah i think a few of my mnemonics had too much thought put into it and I overcomplicated it, but because I came up with it and I had my tunnel vision engaged, it was inherently easier for me.

Agree, splitting it into a reading and meaning mnemonic will also simplify too (like on wanikani)

you read my mind :open_mouth: I was just typing that very thing above.

3 Likes

Gotcha, thanks for your feedback. Has given me a bunch to think about, and Iā€™ll probably reply with more thoughts tomorrow!

Not sure if you got a notification because I edited you in after, but have a look at the system I outline in the post above. How does that look to you?

Edit: since it is a bit long, I am specifically referring to splitting the meaning/pronunciation mnemonics while retaining the character/setting/location system.

E.g., for ę˜Æ (meaning: IS, pronunciation: SHI4) which contains the components ā€œsunā€ (top) and ā€œpointerā€ (bottom):

MEANING: The sun uses a pointer to indicate what is.
PRONUNCIATION: Shrek writes ā€œisā€ in goo-y swamp slime on the tile of the bathroom within the house.

This seems a step in your direction, which is why I am eager to hear your thoughts.

3 Likes

And for completions sake, here is what a split mnemonic for 儽 good/hao3 would look like, since that is the example I used in my videos and articles, previously.

MEANING: It is good for a woman to be with her child.
PRONUNCIATION: Harry Potter is in the cellar of the dark barn looking for some good milk to drink. (He uses his magic wand to light the way so he doesnā€™t trip over any farm tools.)

1 Like

Iā€™ll have a look over them tomorrow (getting a bit late) so I can can cross reference a few things :+1:

1 Like

If you read The Chaos you will start wondering if English ever gave more than pronunciation clues :wink:

Iā€™m well aware that this method is ā€œlossyā€, and Iā€™m also guilty of picking some characters that gets my point across. For example, 鳓 has nothing to do with 口 and is pronounced ming2 :man_shrugging:

But these lossy phonetic components can help jump-start my brain from nothing, whereas ā€œHulkā€ is quite arbitrary. éµ° is ā€œDiddyā€. éµ” is ā€œWalter Whiteā€. All my brain sees is ā€œbirds something somethingā€, and itā€™s not apparent why I should make the leap from ā€œbirdā€ to either of the three personas, if not at least taking a peak at the phonetic component.

I love it! To nail the tone, one could even add ā€œwhy 2nd, and not 1st? Because the 1st is a public holiday, so the swans are off from work on that dayā€.

1 Like

It doesnā€™t? I guess the confusion comes from the fact that in English our ā€œconsonant in isolationā€ sound is this: Mid central vowel - Wikipedia

While in this sample they are using a dipthong with this: Close back rounded vowel - Wikipedia
and this: Mid back rounded vowel - Wikipedia

This is one reason I donā€™t use the mnemonics. I think in the beginning it works pretty well, even if it is large overhead. I remember that story about Yuigo being in the basement protecting his cards from the rain => é›Ø. What Iā€™m worried about is what happens when Iā€™m 2000 characters in? Now I have ~50 characters in the 4 main locations like 10 times each on average. Iā€™m afraid at that point all the stories are going to just going to blend together. Personally I avoid this problem by ignoring the mnemonic 90% of the time. For some reason, most characters are just easy to remember, while others are hard. For the ones I mess up on reviews with, I go back and read the mnemonic for extra help.

However hereā€™s the other side:

  1. The funny thing is, you actually can memorize 3000 silly stories in your head like Harry potter being in the basement saying ā€œVoldemort took everything good from meā€. Your brain is really really good at memorizing when there are associations and meaning involved. The issue would just be if the stories start to drop in quality (not saying they do).
  2. You donā€™t actually have to memorize every mnemonic or even every part of every mnemonic. Itā€™s probably good enough to memorize parts of it especially the parts you struggle with. The point of the mnemonic is not to be able to generate a word that you donā€™t know. The point is to be able to recall something thatā€™s actually still in your memory.

I was with you on the previous point, but have to disagree here. It seems like youā€™re saying

A) there is a tradeoff between mnemonic cost to memorize and memory payoff
B) you prefer that the mnemonic be simpler and contribute less information, specifically, to not include tones
C) you think the parts the mnemonic doesnā€™t teach you can be drilled in reviews

But what if Iā€™m a user who has no issue remembering that 口 means mouth, but struggle to remember that itā€™s pronounced ā€œkou3ā€? In fact Iā€™d say all English speakers are going to struggle with inherently remembering the tones, and therefore that should be incorporated into the mnemonics.

This is why the system forces you to go through the sound + component items. Itā€™s teaching you an encoding of the sounds. E.g. you encode ā€œqiā€ to Kirby. The idea being that its a lot easier to make other mnemonics out of characters like Kirby than it is to do so from ā€œqiā€. Of course, if you skip through the sound and component lessons/reviews, youā€™re going to have poor decode speed (or maybe not be able to decode at all).

To give an analogy, for piano you learn to read sheet music. It would be way easier for a complete novice at sheet music to see ā€œC - E - Gā€ and play C major. But you can read much more efficiently if you internalize the bass + treble clefs.

TL;DR

This thread is about two things

  1. ā€œThe mnemonics should be systematic and built on top of the sound/tone/component associationsā€ vs ā€œThe mnemonics should be ad-hoc and requiring little to no prepā€

  2. ā€œThe mnemonics should carry complete information about the characterā€, ā€œThe mnemonics should be carry partial information"

The site currently chooses 1a and 2a. I agree that 1b would be easier, arguably more effective, albeit harder to come up with. But I couldnā€™t disagree more with 2b, especially if the tone is dropped from the mnemonic - itā€™s setting English speakers up for failure, because we donā€™t yet have an tones a phonemes (deriving meaning from tone differences). I like the following example of 1b + 2a:

You encoded the meaning and reading in a single medium length sentence.

3 Likes

I think that would be a really good solution! The advanced solution of being able to see your own association in the card would be fantastic, but even without this it would be great.

This is the format I was referring too, but:

  1. I think itā€™s important for the meaning mnemonic to always start from the component. Those are the trigger, I donā€™t think it works if you start the mnemonic with the word youā€™re trying to remember.

  2. Maybe itā€™s a misunderstanding because I said to me itā€™s too much stuff in a single story, but the two mnemonics can (and when possible should) still be related. The key is that the pronunciation mnemonic starts from the meaning. Following the previous point, the recall trigger for the pronunciation is either the meaning itself or its whole meaning mnemonic.

With this, to me the example would look more like:
MEANING: A woman is holding a child while saying: ā€œYouā€™re a good boy, youā€™re a good boyā€¦ā€
PRONUNCIATION: The good boy is Harry Potter, whoā€™s just been abandoned in the cellar of a dark barn after Voldemort killed his parents.

2 Likes

I would definitely still go for 1a and one big reason is remembering the tones. Every word has a tone, and encoding it to the mnemonic with its number would be incredibly ineffective to me. Why would I recall that I visited the swans on the 2nd of July, or that the cow has 3 rings in the mouth instead of 2? Thatā€™s basically the same as remembering the tone without a mnemonic.

The systematic character+location approach works for me exactly because of what you laid out: once youā€™ve encoded the sounds, the decode speed is pretty fast. And similar to you, I donā€™t use the mnemonic for all characters. I use it when Iā€™m having trouble remembering a specific aspect of one, so I donā€™t see the number of stories becoming an issue once Iā€™ve learned 2000.

1 Like

Great, this is much better than mine and further reinforces the tone.

It might just be me; if Iā€™m saying the English word / letter ā€œbā€ Iā€™d be enunciating it as either ā€œbeeā€ or ā€œbuhā€ . To me, the audio sounded more like ā€œboaā€ - I think a lot of this is me not knowing whatā€™s expected of me more than anything :slight_smile:

I have the same feeling on the Japanese front, I think Iā€™m currently at about 8000 items, and now Iā€™m using the mnemonics less and less, albeit not for the reasons you mentioned, ie: confusion, but rather because I have a familiarity with the characters now and can more easily associate characters/items with other things previously learned.

I wholeheartedly agree, but what Iā€™d say is, and Iā€™d think weā€™d all agree, that the purpose of the mnemonic is to actively and quickly get the ā€œthingā€ into our short-term memory, and when ultimately the said thing moves into our mid- to long-term memory, the mnemonic slowly drifts away, leaving us just with the meaning (or whatever).

But if the mnemonic is still ā€œactiveā€ and we have other ā€œthingsā€ (being learned) with said elements of that mnemonic also at play (more active items), how long is it going to be before the stories conflate? I mean, this is not as much of an issue if youā€™re only learning 5, 10, maybe even 20 items a day (where we decrease conflation), but if youā€™re learning 50+ and of those 50, letā€™s say 30 of them have shared elements in their mnemonics. Well, thatā€™s where I tap out and just learn it rote.

Letā€™s have an example. I tell you five unique and separate stories involving a piano. I need you to tell me these stories the next day. Youā€™ll probably get one of the stories right or conflate the rest because theyā€™re all active. Now, if I told you one of the unique stories of a piano once a week over the course of five weeks and gave you a chance to absorb the stories, youā€™d likely give me the correct story every week. So weā€™ve lessened the problem, but weā€™ve created a new one in the form of item limit absorbtion and maximum allowable items. But it doesnā€™t have to be this way :smiley: letā€™s just change one of the five stores to be a pig, the second to be a house, the third to be an apple, and so on and so on :smiley:

Agree :slight_smile:

Apologies; as I touched on, I was tentative to include the pinyin aspect in my example due to my lack of familiarity, but Iā€™m 100% with you; the tone has to be in the mnemonic.

Iā€™m not saying Iā€™d skip it. My point is more that Iā€™d learn it independently (as Iā€™d learn hiragana on tofugu). (Again, sorry, Iā€™m not familiar, but if ā€œqiā€ sounds like ā€œkeyā€ (?) then thatā€™s what Iā€™d associate it with, but I digress.)
(edit: so afaiu itā€™s pronounced ā€œcheeā€ in which case Iā€™d associate with ā€œcheeseā€ or maybe with buddhism in some way, digress again,ā€¦sorry)

What Iā€™m driving at is that I donā€™t like the fact that the mnemonics have an overreliance on these associations, which puts us in what I mentioned earlier, this fixed ā€œmnemonic ecosystem,ā€ which I think hinders more than it does help for the reasons I mentioned above and previously.

Iā€™m basing this on the assumption that the meaning is learned prior (like on Wanikani): the ā€œcardā€ has been given enough time to be shifted into short to mid-term memory, and then afterwards the reading ā€œcardā€ will be shown, the ā€œcardā€ comes up (maybe thatā€™s + 10 ā€œcardsā€ later or based on the results of the reviews, either/or), and the mnemonic ties it back to the meaning. Itā€™s like double reinforcement, which lessens confusion yet again.

The issue here, I think, is more down to my mnemonic not being very good; , @Youyujuan example is better. I just came up with mine off the cuff while typing. There is plenty of ways to better encapsulate any element of a mnemonic in a story.

So maybe instead of this

The laughing cow (kĒ’u ā€¢ kou3) had a really big mouth blah blah blah

You could make the element your looking for more memorable, like

The cow (kĒ’u ā€¢ kou3) was heavily malnourished and, as a result, was only born with three udders. To compensate for this, he developed a really big mouth

You can make them as crazy or not so crazy as you want; some people find that helpful.

The important thing for me is the ā€œcowā€ association, as I mentioned; this gets me 90% of the way there; everything else falls into place all on its own. :slight_smile: (Also, it might have gone over a few of your heads depending on where you live in the world, but the ā€œlaughing cowā€ is actually a thing, offering further reinforcement :smiley: )

Letā€™s also change the other one too.

Iā€™m going to Hong Kong (hĆ³ng ā€¢ hong2) to see the rare black swan, There are only two in existence, and they are the last of their kind. They are both males, so this is our last chance!

This has a bit more of a play on words, ie a ā€œblack swan eventā€ etc and ā€œrareā€ but it does a better job of tying everything together compared to my original example.

What Iā€™m seeing, though, and this is the overarching theme, is that you guys are either not using the menmonics at all or are rarely using them. I mean, this is the crux of the problem. Not using them is a problem (which avoids my complaints completely), and partially using them lessens or significantly reduces a lot of the issues Iā€™ve raised by virtue and what I mentioned earlier. :man_shrugging:

@Kevin Iā€™ve not forgotton you, I will take a look at you example in the coming days :smiley: :+1:

1 Like

As an aside, I do find seeing how youā€™ve seemingly filled in the gaps here, super interesting :slight_smile:

afaict, Youā€™ve internalised the 3 rings, which I initially envisioned as earrings, and (again, this is seemingly and nothing more than a guess on my part) turned my bad mnemonic into a better one. I assume youā€™ve associated the three rings with lip rings instead because of the mouth reference.

Mnemonics at their finest :smiley: