Mind-map learning strategy

phz

秀才
Hey pleco family. This is one of my very few posts, so first of all a few words about myself. Right after that I will describe my idea and form open questions to the community.

I am learning (simplified) Chinese alongside my engineering studies and have started about three years ago. My learning curve was very, very flat at first (enthusiastic mother tongue teacher from northern China but I felt lost since he had no sense for my learning problems). Now I do have a very skilled and patient teacher and I couldn't think of a better one for my needs!! We are focusing on spoken language, are using the swiss book Discover China 1 (in german: China entdecken 1) and write regular vocabulary tests at the end of each unit. I do sort the most important vocabulary (80%) and store these in flashcards. As additional references I have just bought "Modern Mandarin Chinese Grammar" and do use pleco products for about 100 dollars (most notably Tuttle C-E with lots of explanations and examples as well as the new Outlier Mini with great additions for etymology). My vocabulary size is just about 250 words yet.

Right after waking up this friday an idea came flashing to my mind (which was probably wandering in my subconscious for some weeks). And I have learned to not throw away thoughts like this.

I would like to try a mind-map based approach to re-view already learned vocabulary. I recognize some radicals and patterns regularly but the connection is missing to already learned stuff. I definitely am a visual learner and I see great potential for my personal learning curve.

Has anyone tried mind-maps for re-viewing vocabulary yet? I have thought of (1) a story/word based mind map and especially (2) a radical based mind map. I do have some experience with data analysis and typography but really do want to hear your thoughts first about the concept before taking the efforts! Is radical based learning outdated or could it really help me in my case? Do you maybe have some further tips?

Thank you in advance!
 
Last edited:

Shun

状元
Hi phz,

I think this is a good intermediate approach, see below, I would just try it on paper/using a mind mapping app, you will quickly learn how it feels to write characters after having reviewed them this way.

As you surely are aware, almost all radicals also have a meaning element to them. So if you connect the presence of a radical with a meaning element or category of a character, you've already been able to make an important association in the brain, making memorization easier. As I understand it, the radical-based mind map you propose simply groups characters visually, to help you recognize the radicals. What you then could also do is to study most of the 214 or so radicals with their connected meanings first, so you will know about them from seeing any character containing a radical.

For the radical-based mind map idea, I can point you to this website:

http://www.zhongwen.com/

(made from a book, only for Traditional characters, but with Simplified characters in brackets)

There, all characters are grouped by their components, starting with radicals. At your current level, where studying many new characters is particularly challenging, this strategy may well help you as an intermediate step, but once you know the radicals and other components well enough, I think it'll be easy for you just to study the characters by association and repeat them without having to group them at all. But I think as an intermediate stage, to help the brain along, it can surely be great.

By story/word based mind maps, do you mean mind maps based on meaning relations between words? I also think that can help you replicate the semantic structure present in your memory, and extend it more easily by helping the brain put new words in the right place. That's why I called my own mind map a "semantic tree" (now at 400 words or so). I attach the still very much in-progress PDF to this message. (translations are in German) At the end, it should contain many thousands of words. Studying works, of course, by covering up ever-larger parts of the mind map and remembering them. I plan to make finer distinctions between the words clearer in the mind map, based on dictionaries, but especially on the ways I've seen the words used in actual context.

I suggest you try both mind map types out and report your experiences, so we could talk about them in greater detail.

Cheers,

Shun
 

Attachments

  • Semantic Tree.pdf
    140.4 KB · Views: 1,048
  • Semantic Tree 2.pdf
    187.7 KB · Views: 725

feng

榜眼
For characters, this is very helpful. I'm not sure about multi-character vocabulary, but maybe that is my own limitation. Grouping vocabulary is of course useful, but maybe I am just too ignorant of mind mapping to see how it would do anything special for multi-character vocabulary.
 

Shun

状元
Yeah, you almost have to understand the brain's internal structure to know how to put vocabulary in a sensible meaning hierarchy. It was more of a trial. You'd have to take a "brain with perfect language skills", see how it's got its meanings stored, then visualize that in a mind map accordingly. I think that's because the "meaning hierarchy" in the brain is formed while using a language (or thinking about it). But that isn't possible in the near and medium term. For single characters, you could of course use some meaning elements to create a hierarchy. But that has already been done. (Partly by the Outlier Dictionary, and partly by the book on the website www.zhongwen.com) Have you tried those?

Cheers, Shun
 
Last edited:

rizen suha

状元
Hi phz,

I think this is a good intermediate approach, see below, I would just try it on paper/using a mind mapping app, you will quickly learn how it feels to write characters after having reviewed them this way.

As you surely are aware, almost all radicals also have a meaning element to them. So if you connect the presence of a radical with a meaning element or category of a character, you've already been able to make an important association in the brain, making memorization easier. As I understand it, the radical-based mind map you propose simply groups characters visually, to help you recognize the radicals. What you then could also do is to study most of the 214 or so radicals with their connected meanings first, so you will know about them from seeing any character containing a radical.

For the radical-based mind map idea, I can point you to this website:

http://www.zhongwen.com/

(made from a book, only for Traditional characters, but with Simplified characters in brackets)

There, all characters are grouped by their components, starting with radicals. At your current level, where studying many new characters is particularly challenging, this strategy may well help you as an intermediate step, but once you know the radicals and other components well enough, I think it'll be easy for you just to study the characters by association and repeat them without having to group them at all. But I think as an intermediate stage, to help the brain along, it can surely be great.

By story/word based mind maps, do you mean mind maps based on meaning relations between words? I also think that can help you replicate the semantic structure present in your memory, and extend it more easily by helping the brain put new words in the right place. That's why I called my own mind map a "semantic tree" (now at 400 words or so). I attach the still very much in-progress PDF to this message. (translations are in German) At the end, it should contain many thousands of words. Studying works, of course, by covering up ever-larger parts of the mind map and remembering them. I plan to make finer distinctions between the words clearer in the mind map, based on dictionaries, but especially on the ways I've seen the words used in actual context.

I suggest you try both mind map types out and report your experiences, so we could talk about them in greater detail.

Cheers,

Shun
WOW! keep up the good work = keep updating the map. i will certainly study it by heart (i can read german:) thanks
 

Shun

状元
WOW! keep up the good work = keep updating the map. i will certainly study it by heart (i can read german:) thanks

Hi rizen, thanks! :) The tricky part is just getting the ordering right; it has to mirror the actual structure in the brain, or at least have one that makes sense to it/it can remember. I promise that I will continue, at least for the kinds of words that fit in well.

I had worked on it some more, restructured it, and tried a new system, so here's the updated file.
 

Attachments

  • Semantic Tree 2.pdf
    187.7 KB · Views: 629

rizen suha

状元
Hi rizen, thanks! :) The tricky part is just getting the ordering right; it has to mirror the actual structure in the brain, or at least have one that makes sense to it/it can remember. I promise that I will continue, at least for the kinds of words that fit in well.

I had worked on it some more, restructured it, and tried a new system, so here's the updated file.
yeah, thats the reason (ambuiguity and tim toady) why so many hierarchical ontologies in computer science and machine learning (ai) have failed. but for human learning/understanding, a personal and "fuzzy" taxonomy goes a VERY long way. thanks
 

Shun

状元
Agree, if you explain your personal system. Perhaps you would also like to start an English version of a Chinese vocabulary tree? I think it’s best to just start it with no extra tools or information, just out of your head, adding words you think of and creating a hierarchy with them, finding similarities and differences, and so on. (only if you want) But I will definitely carry on.
 
Last edited:

rizen suha

状元
Agree, if you explain your personal system. Perhaps you would also like to start an English version of a Chinese vocabulary tree? I think it’s best to just start it with no extra tools or information, just out of your head, adding words you think of and creating a hierarchy with them, finding similarities and differences, and so on. (only if you want) But I will definitely carry on.
... soo many things to do, soo little time. grateful to you and fellow minded creative spirits for doing any doing at all. thanks!
 

feng

榜眼
Yeah, you almost have to understand the brain's internal structure to know how to put vocabulary in a sensible meaning hierarchy. It was more of a trial. You'd have to take a "brain with perfect language skills", see how it's got its meanings stored, then visualize that in a mind map accordingly. I think that's because the "meaning hierarchy" in the brain is formed while using a language (or thinking about it). But that isn't possible in the near and medium term. For single characters, you could of course use some meaning elements to create a hierarchy. But that has already been done. (Partly by the Outlier Dictionary, and partly by the book on the website www.zhongwen.com) Have you tried those?

Cheers, Shun
I had the paper version of zhongwen.com back in 1999 or 2000. Got rid of it after a while as I found it to be more trouble than it was worth in terms of etymology (sometimes fanciful), and the fact that etymology does not seem like a good way to learn (too much time, not enough benefit).

With what I saw of Outlier way back when Mike posted something about it on this forum before they published, it didn't seem like an efficient way to learn (etymology does not seem like a good way to learn), and they just scanned in old characters from scholarly books. But I have not seen their published work. It is a book or just an app?

There are other ways to look at characters, but that is another story for another time . . .
 

Shun

状元
Hi feng,

the zhongwen.com site can perhaps help you with some basic knowledge, e.g. that 三点水 means "water" or verbs describing activities done by hand often have a 提手旁, and so on. Beyond that, I don't believe that knowledge of etymology was all that useful in my learning, either. You just create your own mnemonics. For example, if you have to memorize a number, you just remember it without knowing why the number is the way it is. It has to work similarly with Chinese characters. You simply memorize them without knowing how you did it. It may be in the subconscious. The brain is a big mixture of interconnected memories, so you just have to learn to attach the character knowledge to the other memories as economically as possible. :)

I don't think Outlier exists in book form (yet) because they're working on it right now.
 
Last edited:

feng

榜眼
But Outlier did come out with their app? Back before they came out with it, their website did not have enough information for me to understand what they were really doing. It seemed like it was more of a dictionary than a learning tool, or do I have that wrong?

Anyway, this mind map stuff seems like word groups: colors, parts of the body, prepositions, etc., or am I misunderstanding?
 

Shun

状元
I think Outlier is both a dictionary and a learning tool. A dictionary for paleographers, and a learning tool for Chinese learners. Outlier uncovers an internal logic in the way the characters are composed, offering you a framework to help you remember them.

With Pleco as their partner, Outlier don't need to bring out an app of their own.

The mind map/semantic tree tries to group vocabulary first by rough meaning categories, then by ever finer distinctions of meaning. This should help you to use more new vocabulary in language production, since you should then know exactly when to use which word. But categorizing all ~50,000 core words is a lot of work, and it takes a lot of imagination to really come up with good real-life distinctions.
 
Last edited:

feng

榜眼
Thank you for your replies, Shun.

The reason I am asking, is that the PDFs you put up above seemed to my ignorant eye to be just groupings of characters. Your latest post suggests that mind mapping is also similar to a collocation dictionary. Is that right? One would include all the ways to say "but" in Chinese together, and then have explanations and/or examples of the why different buts are used?

I am still trying to figure out if mind mapping is just a fancy name for word groups and collocations, or if it is really something separate. Can you point me in the direction of examples of finished mind maps for something, preferably language learning? Googling it turns up nothing I can classify as other than bs.
 

Shun

状元
You're welcome!

I'll gladly try to explain the differences between collocations and my semantic tree idea. With collocations, you take a pair of words or expressions and count their frequency of appearing together in a text of, say, 1 million words. You don't know at the beginning which expressions frequently appear together, so the computer will do that for you and create a list of the most frequent collocations. This can help with language learning, as well, since it allows you to learn how certain meanings are usually expressed using combinations of words/expressions.

With a semantic tree, you assume that every word or expression has a meaning (which is probably true), and that the meanings of different words are somehow related to one another. For example, "large" and "small" are both adjectives describing the size of something, so they can be put in a category such as: "Physical world" / "Size" / "degree" (three levels down). You can put other words in that category, such as "medium-sized", and subdivide it further if it gets confusing. Grouping these words of similar meaning together can help you strengthen their grouping in your memory through repetition. This should facilitate the process of extending the range of vocabulary you use, especially when speaking and writing. When reading and listening, you still tend to translate meanings directly, because you associate each word directly with the meaning, possibly accompanied by the memory of an experience.

The same words "large" and "small" can of course also be used more abstractly, as in "a small project" or "a large undertaking". Here, the category "Physical world" wouldn't be appropriate. The solution to this is simply to put the same words in more than one place. Especially with frequently-used words, this is likely to happen a lot, but doesn't hurt at all. You can search for all the occurrences of that word with a computer. So you would have to put it under something like "Abstract world" / "Size" / "degree" (these categories have to evolve while you’re working on the mind map).

I think the reason why there is no such mind map on Google is simply that 1. so far, it relies on pretty unscientific "methods" to create, and it's difficult to find a scientifically sound method to create them, and 2. it requires a lot of introspection and discipline on the part of the creator of the semantic tree. Nevertheless, I think it's absolutely possible if the creator of the mind map finds generally accepted meanings, perhaps also drawing from data from the field of lexical semantics. Perhaps I should go read some of that before I write anything more.

I'm sorry that I can't present a finished semantic tree to you. It's also important to start the tree with the right words, you can't start at "A" in the dictionary and work your way through to "Z", you need a good strategy to determine the order in which you add words/expressions to the semantic tree.

I hope this makes it a bit clearer!


EDIT: I found this interesting paragraph on Wikipedia on WordNet, which is a kind of semantic network for the English language:

https://en.wikipedia.org/wiki/WordNet#Psycholinguistic_aspects

I think a semantic tree can never be scientifically substantiated fully, but that doesn't have to detract from its potential usefulness.
 
Last edited:

feng

榜眼
Thanks! I think I finally understand you.

I'll gladly try to explain the differences between collocations and my semantic tree idea. With collocations, you take a pair of words or expressions and count their frequency of appearing together in a text of, say, 1 million words. You don't know at the beginning which expressions frequently appear together, so the computer will do that for you and create a list of the most frequent collocations. This can help with language learning, as well, since it allows you to learn how certain meanings are usually expressed using combinations of words/expressions.
My understanding of collocations in linguistics is the juxtaposition of two words in a non-obvious way or in a way that excludes other lexicologically reasonable choices.

Pants and scissors come in pairs; cake comes in pieces or slices rather than chunks or platefuls; we commit a crime (though there is the one-off, slang expression "If you can't do the time, don't do the crime.") .

No computers necessary, and the term was used for language learning and language teaching before computers.

They are almost like idioms in that native speakers know them through simply growing up in the language, but foreign learners find them a bit thorny since the choices involved (X "Pass me a scissor." X) are not inherently obvious. I guess computers might identify some that might not readily come to mind, but I would quote Elon Musk: humans are underrated.

The same words "large" and "small" can of course also be used more abstractly, as in "a small project" or "a large undertaking". Here, the category "Physical world" wouldn't be appropriate. The solution to this is simply to put the same words in more than one place. Especially with frequently-used words, this is likely to happen a lot, but doesn't hurt at all. You can search for all the occurrences of that word with a computer. So you would have to put it under something like "Abstract world" / "Size" / "degree" (these categories have to evolve while you’re working on the mind map).
Here's where collocations and word groups/mind maps may collide -- and this is half of why I have been so puzzled. Metaphorical, and even literal, uses vary among languages.
Out in the hot sun in summer, a Chinese speaker may say 太陽很大. One can't use big in that way in English.
Cute/可愛 in Chinese and English do not overlap fully (I can think of at least three non-shared uses of cute in Chinese and English right now), and this messes up reasonably fluent people.
This is why even highly fluent foreign speakers can say things that make native speakers grin. How does this work within a mind map?

I think the reason why there is no such mind map on Google is simply that 1. so far, it relies on pretty unscientific "methods" to create, and it's difficult to find a scientifically sound method to create them, and 2. it requires a lot of introspection and discipline on the part of the creator of the semantic tree. Nevertheless, I think it's absolutely possible if the creator of the mind map finds generally accepted meanings, perhaps also drawing from data from the field of lexical semantics. Perhaps I should go read some of that before I write anything more.
There's lots of information on mind maps when googled, but for language learning it seems like the same word group idea that has existed as long as the history of foreign language learning.

Interestingly, what I have read says that mind mapping is hierarchical. Foreign language vocabulary does not seem hierarchical to me. Now I understand why all the lines in your PDF were running to a single place, but I don't understand how foreign language vocabulary can be that way. I guess that is what you meant here:
I'm sorry that I can't present a finished semantic tree to you. It's also important to start the tree with the right words, you can't start at "A" in the dictionary and work your way through to "Z", you need a good strategy to determine the order in which you add words/expressions to the semantic tree.
Of course, some words are more common than others.
 

Shun

状元
Thanks! I think I finally understand you.

That's good to hear! :)

My understanding of collocations in linguistics is the juxtaposition of two words in a non-obvious way or in a way that excludes other lexicologically reasonable choices.

Pants and scissors come in pairs; cake comes in pieces or slices rather than chunks or platefuls; we commit a crime (though there is the one-off, slang expression "If you can't do the time, don't do the crime.") .

No computers necessary, and the term was used for language learning and language teaching before computers.

That may well be an earlier meaning of the word "collocation". Where I got my understanding of the word "collocation" from is corpus lingustics, for example from using the Sketch Engine website. There, the computer looks through a corpus and just displays a statistic of all expressions occurring together frequently.

They are almost like idioms in that native speakers know them through simply growing up in the language, but foreign learners find them a bit thorny since the choices involved (X "Pass me a scissor." X) are not inherently obvious. I guess computers might identify some that might not readily come to mind, but I would quote Elon Musk: humans are underrated.

That is certainly true for creating a semantic tree. I think computers should be looked at as "a bicycle for the mind" (to quote Steve Jobs). In the ideal case, they free humans of the fetters of long and dull meticulous work, but the human stays in control and retains all freedom of movement in what they wish to do.

Here's where collocations and word groups/mind maps may collide -- and this is half of why I have been so puzzled. Metaphorical, and even literal, uses vary among languages.
Out in the hot sun in summer, a Chinese speaker may say 太陽很大. One can't use big in that way in English.
Cute/可愛 in Chinese and English do not overlap fully (I can think of at least three non-shared uses of cute in Chinese and English right now), and this messes up reasonably fluent people.
This is why even highly fluent foreign speakers can say things that make native speakers grin. How does this work within a mind map?

The mind map/semantic tree just has to be structured properly for that language, so you'd have one 大 meaning "intense" like in your example (for wind, rain, sunshine, i.e. natural weather phenomena) and another for size, and others for things like "不大高兴".

So you just have to put all the different 大s or 可爱s in the right places using the right structure. It's the job of the creator of the mind map to understand all of these uses and express them through the naming of the subcategories. That's why computers won't be able to do such a thing for quite a while.

There's lots of information on mind maps when googled, but for language learning it seems like the same word group idea that has existed as long as the history of foreign language learning.

Surely the wish to categorize and group words together has existed for a long time. I just observe that full-fledged "semantic trees" for different languages aren't found in the wild yet.

Interestingly, what I have read says that mind mapping is hierarchical. Foreign language vocabulary does not seem hierarchical to me. Now I understand why all the lines in your PDF were running to a single place, but I don't understand how foreign language vocabulary can be that way.

It's a good idea to question that, as well. This hierarchical arrangement may help you find a word you're looking for by moving through the hierarchy. Instead of having to qualify the meaning of a word using a set of qualifying words, the word is more elegantly defined by its place in the hierarchy. The semantic tree would simply be a reference, a way of recording fine meaning distinctions so you will remember to use the words properly. It wouldn't make much sense to learn the semantic tree by heart, of course.

Of course, some words are more common than others.

Yes, but that's also just one part. The mind isn't very comfortable deconstructing meanings like that. It prefers to work with complete meanings instead of deconstructing them, and it needs a lot of input/thoughts and real experiences so as not to feel very bored while composing a semantic tree. So maybe we'll have to wait for computers to catch up and do this 20-30 years from now, after all.
 
That may well be an earlier meaning of the word "collocation". Where I got my understanding of the word "collocation" from is corpus lingustics, for example from using the Sketch Engine website. There, the computer looks through a corpus and just displays a statistic of all expressions occurring together frequently.
Perhaps we should distinguish between "collocations" and "n-grams" (https://en.wikipedia.org/wiki/N-gram). E.g., even though 我 and 也 occurre together frequently, as far as I know 我也 is not considered a collocation, but a bigram.
 
Last edited:

Shun

状元
Oh yes, thanks! Also in Wikipedia, it says that "There are about six main types of collocations: adjective+noun, noun+noun (such as collective nouns), verb+noun, adverb+adjective, verbs+prepositional phrase (phrasal verbs), and verb+adverb." So it has to be two completely independent words or expressions—which can easily be substituted for another expression—occurring together. You'd probably need a tagged corpus to do that reliably. (tagged for parts of speech and sentence constituents)

If I remember it right, n-gram searches are also frequently used to find all the personal names in a text, because n-grams are just repeated chains of characters of length n. (Google Ngram Viewer is also an interesting (diachronic) tool.)
 
Last edited:
Top