January 5, 1995
The existing mnemonic systems of studying kanji always deal with isolated characters and therefore have a number of shortcomings. An attempt to overcome them is made in a new method called "Kanjichain" that is based on the principles of phonetic grouping, plot mnemonics and artificial context. It allows the student to efficiently learn not only kanji writings and meanings but also their Chinese and Japanese readings and compounds. A comprehensive example of using the method is presented, provided with detailed comments. The necessity of proper computer support is grounded. In the appendix, possible organization of the educational computer program "Kanjichain" is outlined.
It is needless to say that the complexity of written Japanese is one of the major obstacles to mastering this language. The simplest way of learning kanji characters, by multiple repetitions, may be good enough for Japanese children. But for a foreigner, this simplicity does not justify the enormous time expenditures that are inevitable for one who is not a superman. This leads to the idea of using the advanced methods of sensible memorization. Mnemonics, as such, is an ancient art, but nowadays it is experiencing a rebirth and is becoming more and more popular among students of Japanese. First of all, this concerns the famous system of James Heisig [1-3], though there are books by other authors as well, for example, see [4-7]. It is an advance over the traditional mechanical approach, but nothing is impeccable. Careful analysis of these systems detects the following weak points they share:
|•||The number of characters to study is limited by the list of joyo-kanji.|
|•||An emphasis is put on meaning; the mnemonics of reading is either poor and irregular or ignored altogether.|
|•||Only one meaning of a character is usually given ignoring all the other meanings as useless.|
|•||A student is often foisted associations; his/her own creativity is limited.|
The goal of this article is to show that all these disadvantages are caused by the approach which dictates the student to apply mnemonic operations on an isolated character. We suggest a novel approach that overcomes the shortcomings of the existing systems.
The essence of the method
The "three whales" of the chain method are phonetic grouping, plot mnemonics and artificial context. When applied to Japanese, the system based on this method could be called "Kanjichain".
In this system, all the characters with the same Chinese reading are gathered in a group (phonetic grouping) and arranged in a chain according to the story invented by the student (plot mnemonics). By doing so, the characters along with their Japanese readings and, possibly, compounds are learned in an artificial context.
As compared to the systems mentioned above, "Kanjichain" provides the following advantages:
|•||Images within a story can be much more vivid and detailed than those for an isolated kanji. The story creates an "associative field" with various possibilities for associations to choose between. Therefore, the associations are more durable.|
|•||Remembering becomes highly reliable due to the plot connections. Every character acquires its own address in memory and its quick retrieval is made possible - in contrast to the host of tiny, static "stories" which are not linked and thus easy to forget.|
|•||Phonetic grouping breaks down the whole set of kanji in a very natural and regular way. If not only joyo kanji are learned, the method can cover all the vocabulary of Japanese, except the loan words written by katakana (covering all the vocabulary means (a) learning all the Japanese readings and (b) learning all the meanings that a particular kanji can impart to its compounds. Compounds can be inserted into chains as illustrations).|
|•||The learning of Chinese readings is amazingly facilitated - to recollect the onyomi of a character, a student has only to remember which story the character is from. This is the prime advantage of the method.|
|•||Due to the easy onyomi remembering, learning compounds (jukugo) is facilitated, too. This is an alternative to the conventional way in which one learns compounds first and then the readings of their components.|
|•||Artificial context allows the student to compactly learn all the meanings of a polysemantic character or word, forming its integral image similar to that existing in the brain of a native speaker.|
|•||The process of studying becomes extremely creative and absorbing - the student is really a writer and kanji are the heroes of his/her own books!|
It would be unfair not to mention any weaknesses of the chain method. The truth is that it does not provide an arrangement of kanji based on whether they are simple or complicated, common or esoteric. When following this method, the student should realize that the desired result is at the end of the path. However, this result is worth enduring some temporary inconveniences. Besides, if a character seems needless, the student can simply ignore it.
A comprehensive example of using the chain method is presented below.
"Kanjichain" in work
The material for the chain
Let us consider the full set of kanji that have the onyomi "ju:". Their shapes, meanings, readings and some compounds are listed below.
Notes on the list:
Several extremely rare characters are ignored. In general, it should be determined by the student according to the goal of his/her studies.
Character has two onyomi - "chu:" that is more common but irrelevant here and "ju:" that is corresponded only by one group of compounds with a general sense of "wholeness." A couple of examples inserted into our story would work well.
Characters and can also be read as "ju:" in some cases. However, their original onyomi is "shu:". "Ju:" is just the voiced variant for when it follows a voiced sound as in "nanju:" ("difficulty"). is read as "ju:" only when it substitutes in documents - we can treat it with some approximation as a kunyomi. Both the characters are omitted.
At present, the Japanese exhibit a tendency to use kana for many words that were always written by kanji before. One can discuss whether it is reasonable or not. But in terms of our method, it is much more easy and convenient to learn words in connection with particular characters. For this reason, the list below contains the word "hoshiimamani" that is usually written by hiragana.
in compounds also: many, all
e.g. [ju:bun] plenty of; enough
in compounds: utensil
e.g. [ju:ki] utensil
[shiru] soup; juice in compounds: all, whole
[sekaiju:] all over the world
[ichinenju:] all the year round
[karadaju:] all the body
[su.mu] live; reside; inhabit
[su.mai] house; dwelling; residence
in compounds: to fill; to allot
[ju:to:] to allot, to appropriate
[yawa.rakai] soft; mild; tender
(!) [ju:do] judo
in compounds: heaviness, load; hard, severe; serious, important; double; complicated; layer
[omo.i] heavy, massive; serious, important
[omo.mi] weight; importance, dignity
[omo.omo.shii] solemn, grave
[omo.njiru] to esteem, to respect
[kasa.neru] to pile up, to heap up
[kasa.nete] once more; repeatedly
[kasa.naru] to be piled up; to overlap each other
[kasa.ne] pile; layer
[kasane-gasane] repeatedly, over and over again; sincerely; exceedingly
[shitagau] to obey; to follow; to accompany
[shitagatte] therefore; consequently
[shitagaeru] to be accompanied by; to subdue
in compounds: subordinate; secondary; follow, accompany
(!) [itoko] cousin
(!) [ju:rai] before; up to now
[ju:] gun, rifle [tate] length; height
[tate.no] vertical, sheer, perpendicular
[hoshiimama.ni] at will
[kedamono, kemono] beast
Inventing the story
To begin with, let's invent a hero. For this onyomi, a good association suggests itself: "ju:" - Jew. So, all the events in our story will concern a certain Jew. Once the hero is chosen, we must now stretch our imagination and look again through the list above trying to imagine what could happen to this Jew.
An important thing to remember is to put together characters having the same fragment: and ; and ; , and . As shown below, this way simplifies the memorization process. A sequence of such characters within a chain is called subchain and the corresponding fragment is called phonetic as it determines the pronunciation of all the characters in the subchain.
The compounds that are just examples, i.e. they simply show how the meaning of a kanji can be embodied in words, should be inserted in the chain only if they help to connect different parts of the plot or make the story more vivid ("ju:to:", "ju:itsu"), if the meaning they represent is too abstract (like "whole") and if they themselves are very common Japanese words (like "ju:bun").
If a compound is marked with (!), it either has no connection with the basic meaning of the character or has an original reading, not based on the readings of its compounds. Such compounds should be included in the chain, too.
An old Jew had run away from Egyptian captivity and decided to build his own house in the uninhabited Promised Land. He had never built anything before ("ju:rai"). Therefore ("shitagatte"), he called his experienced cousin to help. This smart guy subdued the oldman in the most barefaced way ("shitagaeru"). The poor devil followed his cousin everywhere and obeyed his orders ("shitagaru"). To create a high brickwork, the cousin climbed a trestle and hung a plumb - long thread that hung vertically and perpendicularly to the floor ("tate"). Unfortunately, he did it not according to the rules but arbitrarily, at will ("hoshiimamani"). Meanwhile, the oldman began his serious and important work of carrying heavy bricks ("omoi", "omomi") to the place of the future double brickwork. This process looked solemn and imposing ("omoomoshii") because he respected and esteemed every brick ("omonjiru") - breaking one would be a heinous crime deserving severe punishment (again "omoi"). He heaped them up ("kasaneru") forming a multi-layer pile ("kasane"). During his work, our hero, time and again, sincerely thanked his cousin for his exceeding encouragement ("kasanegasane") while the latter was sitting on the trestle repeatedly twiddling his thumbs ("kasanete"). "How many?" asked he when at last all the bricks were piled up overlapping each other ("kasanaru"). "A great many!" was the answer. "As many as ten!" ("tou") "Oho! That's enough ("ju:bun"). Well, let's now have a bite before they bring the mortar. Come on, take the utensil and ladle out the soup ("shiru"). I'm eager to fill up my paunch, so do allot a rich piece to me ("ju:to:"). Take care not to spill ("ju:itsu"). Oh, by the way: have you got a gun?" "A gun?!" "Yeah, a gun. Against beasts ("kemono"). They've bred throughout all over the world ("sekaiju:") terrorizing people all the year round ("ichinenju:"). Look at me: I've been bitten from head to toe! ("karadaju:") What if they come inside? You're so soft and tender... ("yawarakai") If you don't like guns, you must get a halberd or at worst a wooden bludgeon. As for me, I don't need neither gun nor bludgeon because I know judo!"
Here our story is over for want of kanji. We can only guess that when the mortar was delivered and the brickwork erected, it collapsed on the spot since the plumb was hung "hoshiimamani", i.e. helter-skelter.
Memorizing the characters
Writing of kanji is memorized according to the following general rule: stable, rigid interpretations for the basic primitives and a flexible approach to less common elements. Seven characters in this chain are formed by a bushu (a primitive used to search the kanji in a dictionary) and a tsukuri (all the rest of the kanji). This type of structure is the most common. Bushu should have only one meaning each, as it is fixed in the list below:
- person - thread - dog - halberd
- water - metal - mouth - tree
A tsukuri can be treated as a single primitive, composition of primitives ("ideogram"), integral picture ("pictogram")  or reference to another chain, depending on the particular case and student's wish. The same concerns characters of a different structure, without an obvious bushu, such as , , , etc.
Below all the characters from the story are analyzed.
For the tsukuri , three cases are possible:
1. The student is familiar with the meaning of ("master") treated as a primitive or as a character. This case is assumed in the story above: the conflict between the Jew ("person") and his Egyptian master - the house was the result of this conflict.
2. The student has already learned the chain "chu:" that contains a subchain with this element as a phonetic (, , , ). Then Jew's new dwelling can be bound up with some image from that chain.
3. The student sees this fragment for the first time. He/she can try to form an association treating it as a pictogram or decomposing it into further elements (e.g. "king" and "drop"). Another way is to postpone this matter until the chain "chu:" or character is learned, remembering mechanically for the present.
A special situation. On the face of , this is a kanji with a complex structure. But on the other hand, it is contained in the second character as a tsukuri. Therefore, it is a phonetic and we should not divide it to its own tsukuri and bushu. Its ideographic interpretation is a hard task due to the whimsical shape of the right element. A pictogram may be more vivid. Let it be the figure of the cousin with symmetrically tousled hair, a flat head and outstretched hand by means of which he gives instructions to the oldman. Let him be standing in a heap of debris. Then the left element will be the future brickwork embryo - three bricks, one upright and two horizontally (a bit askew, but this is exactly what the story is about). The second kanji in the subchain can be very naturally interpreted as a plumb to check the verticality of this queer construction.
The fragment of the story devoted to this kanji is a good example of the strength of an artificial context. Indeed, in the same context we have managed to arrange nine related but different words, some of which were polysemantic. As for the character itself, there are two ways: a pictographic interpretation (something looking like a brickwork, though built a bit incompetently) and decomposition into and or even into and with an image like this: a brick, heavy like a car (), between two () hands of our hero. These two ways can supplement each other for greater reliability. Besides, since this kanji is met with in other characters as an element, it can be learned as a primitive, i.e. an indivisible element with a rigid meaning ("heavy"). This would be the third way.
A subchain formed by the phonetic . If the correspondence between this picture and the concept "ten" is recognized well by the student (that is highly probable), extra associations are needless. If not, the story should be added with some details - for example, the oldman could cross himself before his meal.
The bushu "water" matches well the concept "soup"; the bushu "person" is not so natural and, maybe, a bit abstract for "utensil". However, subchains are learned so easily due to their strict structure supported by the plot that the student very often can manage without too detailed mnemonics.
Another subchain. We have to find out an interpretation for the phonetic. One can seek out a sense of "filling" in . However, one can also try to perform mnemonic operations over the second character. It can be interpreted as follows: a stern face with a vertical wrinkle, beetle brows and pendent moustache looking through the backsight of a rifle (the rifle is evidently made of metal and represented by the bushu) - such an interpretation is a hybrid of a pictogram and an ideogram. Apart from everything else, we create here the image of our hero's face - a helpful associative detail and, possibly, an attribute for the whole chain.
As this is a subchain, mnemonics should not be necessarily applied to the first kanji in the subchain. It is enough to remember that this character precedes "gun" and has no bushu ("zero bushu"). Its meaning is recollected by simple moving "up the story."
The bushu "dog" is quite natural here - it sheds light on the mystery of those abstract "beasts." As for the tsukuri, it is so complicated and unique that treating it as a pictogram seems to be the only way. The interpretation: a beast's roar from a cage () in front of which a gaper is standing partitioned by a fence () with his mouth open (). Evidently, this is a deviation from plot mnemonics as there was no cage in our story. However, using isolated associations, when they are reasonable and convenient, is not a sin.
This character is so common that any associations seem needless. However, we should connect it with its place in the chain and with the particular images it begets. Since we easily remember its basic meaning ("inside"), we need not consider the shape. The phrase "What if they come inside?" is sufficient for the connection.
Since both and can perform the function of bushu, this character has no patent bushu and both the primitives are equal in value. This fact is reflected in the story ("halberd or wooden bludgeon").
Mnemonic tricks can also be applied to Japanese readings of kanji characters. For example, the word "kasaneru" can be broken to "kasa" ("umbrella") and "neru" ("to sleep"). However, it is still unclear how to bind up these concepts with the meaning of "kasaneru" ("to pile up", etc.). This problem disappears when we are in the associative field of our story. Let the cousin not simply sit on the trestle and twiddle his thumbs - let him lie down there and sleep, being sheltered by an umbrella to be on the safe side since there is no roof yet. And while he is sleeping, the oldman will pile up his notorious bricks. There is no direct logical connection between the two actions but it is not so necessary. Human memory is not a computer, it needs vivid images rather than strict links.
Remembering Chinese readings is not an end in itself. One should remember them to be able to read compounds. The traditional way is from compounds to the readings of their components; the chain method suggests an alternative.
All the characters in our chain have the same onyomi - "ju:". As this is a chain, i.e. all the characters are bound up by plot connections, just one mnemonic trick is enough to associate all the characters with their onyomi. Whenever you encounter any of these kanji, the story will instantly come to your mind and the nationality of its hero will be an excellent cue to recollect the reading. In the same manner, let Deng Xiaoping be the hero of the chain "den" and Guy July Caesar - of "gai." Let the chain "hen" be about hens and "bi" - about bees. Even you yourself can be the hero of the chain "ai." Long chains with plot transformations may require two or more associations. But in any case inventing a few associations per chain is immeasurably easier than inventing special tricks for every character, which is absolutely impossible. That is why the available mnemonic systems are powerless to appreciably facilitate the remembering of Chinese readings.
Let us assume now that we are not familiar with the chain method. To remember the onyomi of the character , we must learn mechanically some of its compounds, for example, [yaju:] - "wild beast." Then, having met an unknown word, say, , we have to recall this "yaju:", break it to "ya" and "ju:" and establish that the onyomi of is "ju:." The same operation must be applied to and only after that can we conclude that this word is read as "ju:i" ("veterinary"). The redundancy of our efforts is evident. Besides, mechanically learning compounds is more difficult and less reliable since the pronunciation of a jukugo when treated as a sequence of sounds has no internal sense. For most native speakers of European languages, it is especially difficult to remember the length of syllables in Japanese words ("ju:" or "ju"). Many problems can also be caused by the interchange of consonants - for example, if we do not know it beforehand, we cannot be sure that the second syllable in "yaju:" is the original onyomi of but not just the voiced parallel of "shu:", that is quite possible when such a character occupies the second place in a word. Another difficulty is to restore the readings that produce double consonants - they can be of three types, ending by "-ku", "-tsu" and "-chi." For example, the words ("brackets") and ("individual") are both read as "kakko." But the onyomi of their first characters are different: "katsu" and "kaku" respectively. To say nothing of words like [itoko] that are pronounced quite differently from what one may assume on the face of them.
All these troubles seem ridiculous when you know the readings of all the kanji by heart. In this case, to memorize a jukugo, you have only to memorize the characters it consists from. Remembering characters, normally, is much easier than remembering sounds - of course, if you are familiar with them. Thus, it is not an exaggeration to say that the easy and full remembering of onyomi is the most important advantage of "Kanjichain."
About the number of characters to study
All the history of written Japanese is a permanent struggle between the reformers who try to cut the number of kanji and conservatives who resist this simplification as far as they can. We will not consider their arguments here since they usually concern the native Japanese, not foreigners. On the other hand, both the sides agree that foreigners are certainly interested in fewer kanji and more kana. But is it really so?
Of course, so complicated characters are rather an obstacle while we treat them just as a way to write the words we already know. In this case, using kana instead looks very tempting. But if we do not know these words yet and have to memorize them, it is better to perform mnemonic operations. The Japanese language has a natural support for this - kanji characters which are both material and tool for efficient and comprehensive mnemonics. Just like Voltaire said about God, we can say: "Even if kanji did not exist, people should have invented them."
This argument works well while we consider common characters and common words. But once you leave the bounds of joyo-kanji, you hear: "Why on earth must I waste my precious time studying these museum pieces? Most likely, I'll never meet with them in my life!" When kanji are learned in a mechanical way or with primitive mnemonics, this position is natural and valid. But it should be revised when the chain method is selected.
The point is that the chain method offers a novel ideology of saving student's time and efforts. Instead of reducing the general number of characters and words to study, it reduces "specific time", i.e. the time spent for one character or word. An interesting correlation exists: the more characters or words are learned the less efforts are needed per one. This is quite explainable. The characters outside the joyo-kanji list usually have very concrete meanings: names of animals, plants, diseases, body organs, household articles, etc. Inserting them into a chain makes the story much more vivid providing great scope for associations. It simplifies the memorization process and reduces the time per character. Note also that these characters themselves are not difficult to remember since usually they belong to subchains, i.e. differ from the adjacent characters only by bushu. Besides, these esoteric kanji often designate very common Japanese words that are normally written by hiragana. Sooner or later, you will have to learn these words. If you learn them now, in a chain, you will kill two birds with one stone, to say nothing about the respect you can win among the Japanese - knowing rare kanji is still treated in Japan as a sign of good education. Moreover, the trend today is to admitting that not all the "esoteric" kanji are useless. Some textbooks on written Japanese begin to offer such characters to foreign students, as . It is pointed out that knowing joyo-kanji only is not enough to read even thrillers, let alone serious literature.
Many characters are rare because they correspond to obsolete words and concepts. They may be of little interest in terms of utility. But even such characters may be helpful when you construct your chain. They can serve as connecting links to join different parts of the plot or as additional details that impart vividness to the story. Only the student can decide whether a particular kanji should be ignored or learned. To make this choice, all the information about the relevant characters should be available in an easy-to-use form.
The example given here convincingly demonstrates the efficiency and flexibility of the chain method. A story, when being invented, is plastic and sticky like an ant-eater's tongue - it readily absorbs any details that may even seem quite irrelevant to the plot at first sight. The secret of this magic property is the associative field of the story that is always wider than we can assume. For successful memorization, the story should not be too logical. Quite the contrary, it should be a bit eccentric - then it appeals to both logical and emotional cerebral hemispheres. Figuratively speaking, an image is nailed by two nails, it is smoothed out and observable, instead of hanging on a single hook like a formless rag, ready to fall at any moment. To nail an "emotional nail" is easy. One has only to pity the poor old Jew for his apparently fruitless attempts to find a roof over his head, or to wax indignant over the brazen behavior of his impudent, lazy and voracious cousin. Note also that the most firm and reliable nail is humor.
Apparently, there are placers of "emotional nails" in "Kanjichain", unlike the methods developed for isolated characters that cannot provide a wide associative field and therefore have to perform cold logical constructions, applied only to kanji meanings and even then to only single ones. In , J. Heisig recommends to decompose to "column", "animal horns" and "mending." Obviously, a student with satisfactory imagination can exert himself and invent a short story that connects these three concepts with the basic meaning of the character ("to accompany"). But he will have to solve a similar puzzle for every new character. In this manner, his efforts will be highly redundant while their result will be very poor - only one meaning and no readings. To study the readings, Heisig suggests a second scan of all the characters in . Thus, there is an apparent gap between studying kanji and studying the language as such. This gap is considerably narrowed by the chain method which embodies a holistic approach - everything is learned in the same set and in the same time. By doing so, the vocabulary of Japanese is being mastered in an active way, instead of passively waiting until you stumble upon an unknown reading or meaning of a known character.
Another aspect is the reliability of remembering. "My head is like a sieve!" people often complain. This is a problem, indeed. To resolve it, there are at least two ways: to reduce the size of the sieve holes and to magnify the size of the sieve contents. Let us compare a grain of rice and a noodle. Which one has better chances of falling through a hole? To all appearances, the first. A single kanji is like a rice grain, even when it is suspended by a mnemonic thread. If once this thread breaks, the kanji inevitably falls into the coarse-holed sieve of mechanical memory; its further fate is sad. A chain of kanji is quite a different matter. Even if one of the chain tips gets into a hole, we can just pull it out by the other end (of course, the plot links must be sufficiently tough). Furthermore, when we invent our stories, we can let our heroes roam from plot to plot and let old details turn up in new situations. This provides links between chains, which can be supported by identical graphems or homonyms. We operate not in isolated associative fields but in a multi-dimensional associative space. We invent not merely a set of stories but an entire epos. The best illustration for that is a plate of spaghetti. No one part of it can fall through a sieve, they can do it only as a whole - but such a hole would correspond to full loss of memory.
It is generally assumed that words of a foreign language should not be learned in isolation but contextually. However, only a natural context is usually intended (texts, conversational patterns, etc.), though the context can be artificial as well, i.e. created specially and solely for better memorization. Of course, it cannot be so idiomatic as a natural one, and therefore most words learned in this way should be "naturalized" during further language studying. But in any other aspect the artificial context is no worse off. In some sense, it is even preferred due to its flexibility. The suggested approach introduces a novel ideology of language studying and, probably, we do not completely understand all its advantages yet. But it is already clear that the field of its application is not limited by Japanese. A similar approach can be applied to any language where we can naturally perform phonetic grouping.
Evidently, the story given here is not the only possible one - most likely, it's even far from being the best variant. However, the criterion is only one, how it helps a particular student. People can differ in type of memory, sense of humor, imaginativeness, outlook, etc. Some images have repercussion in our imagination and some have not. The way to success is simple: the student invents a chain by himself and for himself. He is free to experiment with his own tricks, free to grope for his own golden mean between logicality and "schizoidism." Our only task is to allow him to fluently orient himself with the host of characters, their elements, readings, meanings and compounds. This begets a number of problems to be solved. The first one is the systematization of kanji elements. They all should be divided in several groups: 1) basic primitives to study beforehand; 2) less common primitives to study gradually, when the student meets them in new characters; 3) elements that are kanji themselves; 4) simple and complex elements without any interpretations suggested (in particular, phonetics). These groups will overlap. Therefore, we must solve one more problem: to define the most rational order on the set of onyomi and recommend it for student's successive studying. This order must minimize the number of confusing situations like that with (c). Besides, we can take into account the level of use so that more useful kanji and words will have priority.
These and other problems should be solved within the comprehensive project of creating a powerful computer program to support studying written Japanese using the chain method. Without such a program, the extensive use of this method is hardly possible yet. Forming a chain "by hand" requires long search through dictionaries, much paper work and even some linguistic intuition not peculiar to everyone. Preliminary estimations show that proper computer support could reduce the time needed to form a chain by 4-5 times. Besides, it could provide a lot of extra facilities, that are now inaccessible. There are about 300 different onyomi in Japanese, so about 300 different chains of various length (from one character up to 50-100 or even more; too long chains can be split up) are to be constructed to cover all the stock of kanji. If the program is developed, this task will be quite feasible within admissible time, and students of Japanese will obtain an efficient, powerful and flexible tool for their studies.
James W. Heisig. REMEMBERING THE KANJI I. A Complete Course on How Not to Forget the Meaning and Writing of Japanese Characters. Japan Publication Trading Company, 1977.
James W. Heisig. REMEMBERING THE KANJI II. A Systematic Guide to Reading Japanese Characters. Japan Publication Trading Company, 1987.
James W. Heisig and Tanya Sienko. REMEMBERING THE KANJI III. Writing and Reading Japanese Characters for Upper-Level Proficiency. Japan Publication Trading Company, 1994.
Michael Pye. THE STUDY OF KANJI. A Handbook of Japanese Characters. The Hokuseido Press, 1984.
Kenneth G.Henshall. A Guide to Remembering Japanese Characters. Charles E. Tuttle Company, 1994.
Andreas Foerster and Naoko Tamura. KANJI ABC. A Systematic Approach to Japanese Characters. Charles E. Tuttle Company, 1994.
Michael Rowley. Kanji Pictographix. Stone Bridge Press, Berkeley, California, 1992.
Possible organization of "Kanjichain" courseware
The kernel of the program should be a database containing all the kanji having JIS X 0208 code (5446 characters) provided with necessary accompanying information. The kernel should be surrounded by various utilities to realize a powerful user interface that supports the following basic modes:
Chain forming. In this mode, the user looks through all the kanji of a particular onyomi, selects the needed characters, invents a story and arranges the characters accordingly. Different readings of characters as well as different meanings of polysemantic words are arranged in the required order, too. Then the formed chain is saved in a special subfile and can be modified if needed. The story itself can be put into words and saved, too, or else it can stay in student's head only, depending on his wish. This mode should also be provided with a number of options to facilitate the memorization process: decomposing a kanji; informing about its elements; indicating the level of use (five levels may be established - from kyoiku kanji down to obsolete forms); searching for its compounds with the characters already learned, synonyms, homonyms, kanji with similar writings, etc.
Chain view. This is a mode to repeat the stories. Characters, readings and meanings successively appear on the screen accordingly to the order in which they were saved.
Drill. The characters from the saved chains (or their readings, meanings, compounds, etc.) randomly appear on the screen to allow the user to check his/her memory.
Beginner's training. A mode to acquaint a beginner with the basic primitive elements and let him/her memorize them distinctly. The main tool for that is a drill procedure that decomposes to their elements randomly drawn characters which are provided only with their keywords.
Dictionary. In this mode the user can obtain any information about a particular kanji or word.
Answers to possible objections
Story-making requires outstanding literary abilities. Most students are ordinary people incapable of creating masterpieces.
A story does not necessarily has to be artistic. Those two stories in the article and newspaper are just examples specially selected and written down to illustrate the method most vividly. As for me, I never put my stories into words and, apart from these two, they all are quite clumsy. Nevertheless, they do work. When a student invents a plot by himself, artistic merits do not matter. As for the story-making technique itself, it is by no means hard. Beginners can master it very rapidly passing from short chains to longer ones.
A student may be frightened by the huge amount of stories he has to invent.
The number 300 reflects the real situation in Japanese. 300 large-scale stories are no more than 2,000 tiny ones when you invent them for every character. In terms of student's efforts, long stories are more efficient: a snowman of three segments requires much less time than a large number of snowballs made from the same amount of snow. Besides, inventing a story is a really joyful and absorbing process.
Studying kanji so profoundly will distract students from studying grammar, conversation, etc.
Full-time students of Japanese will be able to combine mastering different areas of the language very successfully. If they can only use their spare time, I would recommend them to divide the path into two phases as I did: first accumulating kanji and then activating them during ordinary language studies. With good computer support, the first phase can be no more than one year. The obtained result will greatly facilitate mastering the language in the second phase: terrible Japanese characters will turn from enemies into allies.
Evidently, the intensive accumulation of kanji can begin only after some initial knowledge has been received. So, the first phase should be preceded by a "zero phase" in which the student acquaints himself with the fundamentals of Japanese. The program "Kanjichain" will be able to help him in this phase, too (in the "Beginner's training" mode).
Most students will be reluctant to learn much more kanji than is usually recommended and analyse them so profoundly as it is done in the numerous "notes."
The students do not have to learn as many kanji as I did. The computer program is to satisfy their needs whatever level they have selected. All the characters will have the utility index, and setting the program options properly will allow the students to ignore all the unwanted kanji in advance. The program will also automatically carry out such functions as the selection of compounds, distribution of readings and meanings when a character has more than one onyomi, gathering together characters with the same phonetic, etc. Thus, the initial analysis of kanji will be transparent for the users.