ABOUT KANJI
Part 3
J. Belote 2001
(last modified September 9, 2002)
Semantic fields are areas of meaning, including connotations and denotations that may be carried by elements of writing. Most kanji, seen as a WHOLE, have a semantic field, or semantic fields (except for those which are used exclusively as phonetics--click here for examples). Elements WITHIN kanji (including radicals and combinations of radicals) may ALSO carry semantic fields.
The semantic fields covered by a single kanji may be varied and broad or quite specific. Some kanji are rarely or never used as single "words" but are combined with other kanji in order to present a more specific meaning. Other kanji may have a semantic field that is specific enough that they can be used independently (i.e. not in any particular combination with other kanji).
As noted above, the elements WITHIN kanji generally carry semantic fields as well. However, within a particular kanji these semantic fields may be only very loosely allied to the overall semantic field of the kanji within which they exist. Or they may not be related at all.
For example, some elements within kanji may be used (or were originally) as phonetics only--that is, they gave a clue as to the pronounciation of a character (remember, most kanji originated in Chinese, not Japanese). In this case, their semantic field/meaning was irrelevant to the semantic field of the whole kanji.
In some cases, phonetics were chosen which not only gave a clue as to pronounciation, but ALSO some semantic clues! Henshall, who provides etoymological data on all 1945 kanji in his book, gives numerous examples of all these cases.
Finally, because most kanji originated many centuries ago, in many cases their meanings/semantic fields have been altered or dramatically changed. Thus the elements within them may no longer be much, if at all, related to the current meaning/semantic field of the kanji. This is illustrated in an analysis of the character/kanji immediately below.
For a deconstruction and reconstruction of a single kanji ("poetry"), especially in terms of semantic fields, click here. For a look at "key" radicals and how they may relate to the semantic field of the kanji of which they form a part click here.
The semantic fields/connotations/meanings covered by a single character/kanji are generally identical or close in the Chinese languages and in Japanese (and Korean).
Thus, for example, the semantic field of 
is "water" in Japanese and Chinese.
And the semantic fields of 
are
primarily related to "difficulty", "danger"
and "disaster" in all these languages. In Chinese, however,
the character also includes an old denotation of "bird with
golden wings" (Harbaugh
103). Don't ask how the the primary semantic field of the character
for "bird with golden plumage" was transformed so drastically--even
the authorities are puzzled by this one.
Note that "bird with golden wings" makes sense in terms of the elements in the character: the right element denotes "bird" and the left element seems to be an altered form of "yellow." This is a perfect illustration of how the semantic fields found in the elements of a character/kanji may not (no longer, at any rate) give any clue as to its meaning!
Anyway, the kanji which combines
with into a single kanji, 
is another matter. Before getting into that "other matter"
I should note that when water, is
used as a radical in another kanji or character, it is most commonly
appears in something like this form: 
(if hand-written); or in this form: 
(if
type-set). As often (but not always obviously) the case, the radical
may provide a hint as to the semantic field covered by the kanji
as a whole. Thus, for example, we have such kanji as the following,
all with the "water" radical: 
(lake);
(sea);
(waterfall); 
(wave); 
(beach);
and many more.
Alright, back to . Not
surprisingly, then, this kanji has something to do with water.
One might guess that the basic semantic field of this kanji is
something like "troubled water." Maybe it is, and maybe
it isn't. I ran across this character in the Kodansha
bilingual atlas of Japan where it was transliterated as nada;
it was always marked over territory that was an ocean or sea.
The kanji did not appear in any of the four Japanese kanji dictionaries
consulted (Halpern,
Henshall, O'Neill,
Rose-Innes )--all of which
were limited to few more than 2000 kanji entries). I first found
it in Chinese dictionaries. In a Cantonese Chinese dictionary
(Meyer and Wempe) it
was under one entry, t'aan, meaning "rapids,"
"shoals" or "sandbanks." Under another entry
(same word, t'aan, but with a different tone) it was listed
as "marsh," "tidelands," and "flat."
In a dictionary of both Mandarin and Cantonese (Choy)
it was given as "beach" or "shore."Then, in
Japan I ran across two sources (including Nelson)
which gave the meaning, "open sea." And, finally, a
Japanese friend, who did not immediately recognize the kanji,
searched out its meaning and came up with "sea where it is
hard to sail because of strong winds." That's the one I like
best, of course! At any rate it is clear from all this that while
it may be true that MOST kanji/characters have similar meanings
in Japanese and in the various Chinese languages, this cannot
be taken for granted. Even between Cantonese and Mandarin there
may be somewhat different connotations associated with the same
character.
CRIMINALS
OR SINNERS? PRONOUNCING KANJI
In any Chinese language a written character almost always (about 90% of the time) has a single, monosyllabic pronounciation (provided one does not count stress on the last vowel of a diphthong as an indication of a syllable division). In a few cases there are two pronounciations in any one Chinese dialect/language. Kanji, however, more often than not (around 90% of the time) have at least two pronounciations. There are two basic kinds of pronounciations (often called "readings).
ON readings are derived from Chinese
pronounciations taken from the time the character was borrowed
(often more than a thousand years ago) and are now a part of the
Japanese language. There are usually only one or two on
readings for Japanese kanji. Nelson,
however, lists one kanji, meaning "skylark" [#6997]
as "the" example of a kanji which has had many Chinese
readings--8). Although originally monosyllabic, some of the Japanese
versions of the on readings are now polysyllabic: e.g.
RYOKU and RIKI for 
(power, strength).
KUN readings come from Japanese roots,
or Japanese adaptations from non-Chinese languages--and not from
the Chinese pronounciation. They are often multi-syllabic.
Many kanji have more than one on reading and more than
one kun reading. A few have only a single on reading,
and fewer still only a single kun reading. In most dictionaries
the on reading is given in capital letters, the kun
reading in lower-case letters and that is the pattern that will
be followed below.
Note: When a Roman letter representing
a Japanese sound is underlined in this text (e.g. o,
U) the sound of that letter is doubled in length
(in the figures, a macron [line] is placed above the letter).
The difference is phonemically significant. Take, for example,
two locations a few miles to the east-north-east of Mt. Fuji and
only about 16 miles from each other. One is called Oyama
(
- Big Mountain), the other
is Oyama (
- Little Mountain).
So ... watch your vowels!
Examples of kanji readings
Kanji with a single on (Chinese) reading only:
KAN
GOVERNMENT, GOVERNMENT OFFICIAL
KIN PROHIBIT
Kanji with a single kun (Japanese) reading only:
se SHALLOWS,
SHOAL, RAPIDS
waku FRAME,
FRAMEWORK
Kanji with a single on reading and a single kun
reading:
SHO, matsu PINE
HAN, naka HALF,
MIDDLE, UNCOMPLETED
Kanji with multiple on and kun readings:
JU, CHO, e, omo,
kasa HEAVY, DUPLICATE, PILE UP
KA, GE, shita, shimo, moto,
sa, kuda, o DOWN
Ha! You thought that was bad enough! There's more. Not only
do kanji usually have more than one sound used to represent them
in speech, but any particular sound (morpheme) representing one
kanji may be used to represent fifteen or twenty others! Or even
more. KAN, for example, is represented by 49 kanji in general
use! (Fig. 4.c is the kan in kanji
).
(That number came from Halpern.
In Nelson, which includes
thousands of kanji not in general use, there are 133 distinct
kanji listed with the on readings kan. Better yet,
there are 241 entries for the on reading ko!)
Figure
4.
(Note: Kanji 4.g to 4.i and 4j to 4m illustrate the phonetic
nature of some kanji elements--observe the common element found
in each group: 
. The second
["dry"] is itself a radical; the first ["official"]
consists of the radical for roof or building, and a non-radical
element meaning "buttocks" [Henshall,
134]. So much for respect for authority!)
Some (not most) Kanji also have additional
readings used in personal names as indicated in Fig. 4.e,
Fig. 4g, and Fig. 4.h. Here are additional
examples of kanji with "name" readings (Halpern
is the source for all the name readings herein):
KI ERA,
MORALS Name readings: nori, kino, aki,
osamu, toshi, moto, yoshi, sumi, tada
BOKU mutsu FRIENDLY Name
readings: mutsu, mutsumi, chika, yoshi, moku
CHI shi KNOW
Name readings: tomo, satoshi,
kazu, sato, nori, satoru
YU ABUNDANT
Name readings: hiro, hiroshi, yutaka,
yasu, michi
JI, NI BENEVOLENCE
Name readings: hito, hitoshi,
yoshi, makoto, hiroshi, masa, kimi, sato, nori
Note that the sound systems of most of the name readings seem unrelated to either the on or kun readings. Note also that several of the name readings (nori, yoshi, hiroshi)--could be attached to more than one of these kanji. My daughter, Karen, who is in Japan, tells me that she has been told that these variant name readings may be linked to numerological features of kanji (involving stroke numbers) which may be chosen for names (more on this when I know more--or think I do!).
One more thing. Some kanji serve primarily as phonetics--that is, they carry a sound, and may or may not carry a meaning or semantic field. Most were used in the past "especially for the transliteration of Sanscrit Buddhist terms or names" (Halpern, 209).
NA 
MI, BI, iya, ya 
I 
KA, GA, KYA, togi
In the examples above, "I" is used for the initial sound in "Italy" and the kun readings of the second and fourth kanji do carry meanings. For example, togi (fourth kanji) is translated as "keep company before sleep" and with that meaning this kanji combines with two others to provide the meaning, "fairy tale or "nursery tale" (Halpern, 25).
Chinese does not have a system for representing sounds of foreign words such as names and place names, in ordinary text (it does, for specialist purposes, which will not be discussed herein ... at least not yet) other than using the complete sounds of characters that already exist. Thus in translating a work such as the Bible the characters must be used. Either an approximation to the sound of the name or the meaning of the name must be used. In the examples below, taken from Chinese and Japanese versions of the New Testament, characters are used to approximate the sounds of names (Mary, Simon, Matthew and David) in Chinese. Note that the meanings/semantic fields of the characters used have nothing to do with the original (Hebrew) meanings/semantic fields. With names transliterated into Japanese, however, use is made of the katakana syllabary rather than the phonetic content of existing kanji. This helps avoid the ambiguity of the Chinese system where a reader might mistakenly take the meanings (i.e. "east gate") rather than the sounds (e.g. "sai muhn") of the characters as the significant features. I should note that in the Chinese version of the New Testament consulted, most proper names had a single underline, and most place names had a double underline--probably to help readers avoid the potential ambiguity.
Note: if you squeeze the two characters for Matthew together,
you get one, 
, which
in Japanese is pronounced DA or TA and means "good
for nothing" (Halpern
423). Note also: you can't just squeeze any characters together
to create new terms ... well, you can--but it is not standard
practice to do so.
While a person literate in Japanese might be able to get an idea about what a passage of Chinese writing is about (although there are thousands of Chinese characters not in general use in Japan, educated Japanese can recognize most of the characters commonly used in Chinese writing), it would probably be more difficult for a Chinese reader to get as good an idea reading Japanese since so much written in Japanese is expressed in hiragana or katakana. In the following passage note that only half the graphemes are Chinese derived kanji (the others are all in hiragana, in this case). In fact, I chose this passage because it had such a high proportion of kanji--in most Japanese text, the proportion of kanji is far lower (Check the Japanese sentence given on a previous page--only three of thirteen graphemes are kanji).
Note that only two of kanji/characters used occur in both Japanese and Chinese. The general semantic fields of most of the kanji/characters used in both versions would be recognized by readers of both languages (e.g. the character for devil or ghost in Chinese exists also in Japanese, and the kanji for evil and for spirit also exist as Chinese characters.
While most text in
such media as books or magazines may consist of graphemes which
are primarily non-kanji, a medium such as signs may contain only
kanji. The graphemes on this sign, (along the trail system above
Mitaki temple in Hiroshima), both above and below the figure,
for example, are all kanji. The red kanji from left to right connote:
mountain, fire, thing [together these mean "forest
fire"], prevent and stop. Even without the
grammatical elaboration provided by hiragana, the intent of the
message on the sign seems clear enough. A person who could read
only Chinese characters could probably make out the sense of this
kind of presentation about as well as a Japanese reader (provided
there weren't many kanji combinations with radically different
meanings in the two language systems). Question: does a ducky
baseball umpire take the place of Smoky Bear in Japan?
And speaking of smoke ... one more problem. Katakana are usually used to transcribe foreign words (such as "aloe" and "yogurt" in the photograph in Part 1, and in the case of biblical names) but in a few cases kanji may also be used. Take, for example, the word, "America."
Chinese and Japanese
use the same kanji/characters as phonetics in transliterating
"America" except for the second one, me/mei (see
Meyer and Wempe for
this Cantonese Chinese version and Halpern
for this Japanese version of "America.") That kanji
is "rice" in Japanese. The Chinese character in the
same position means "beautiful." The koku kanji
means the same thing as the kwok character in Chinese:
"country, state, empire." In fact, they are alternate
forms of the same kanji/character. The US, then, is "rice
country" in Japanese, "beautiful country" in Chinese.
Note that Cantonese (unlike Mandarin) has no "r" sound.
In Mandarin the pronounciation of the lei character is
li. Just a guess at this point, but it would seem that
Cantonese speakers came up with this transliteration for "America"since
Mandarin speakers do have at least one character with the ri
pronounciation (the character for "sun"--see Fig.1.d).
Forgetting, for a moment, that any kanji may have several
possible sounds, this is all reasonable enough. But there is more.
And it's worse! Sometimes the meaning and sound of a foreign word
may be taken into Japanese and then represented by kanji which
give an idea of the MEANING of the word but NOT OF THE SOUND.
One good example of this is TOBACCO. Tobacco can be, and is, written
in katakana without any problem. But it is also written in kanji,
using kanji which connote "smoke" and "grass": 
. The "smoke" kanji is
normally pronounced EN, kemu, or kemuri; the "grass"kanji,
SO , gusa or kusa. No "tobacco"
is even close to being there. But the "special reading"
tabako is given to the combination, and it means "tobacco"
or "cigarette" (see
Halpern: 255). My questions here are, which part of tabako
goes with smoke and which with grass? Is it ta + bako? Or
is it taba + ko? And does it matter? And is kanji a cognate of
ganja? And why is it that if you say beikoku in Japanese
it could connote either "the USA" or "cooked rice"?
SO ... HOW DO YOU READ THIS STUFF?
. The first kanji
(crime, offense, vice) has the readings ZAI and tsumi;
the second (human being) has the readings JIN, NIN, hito,
bito, ri and to. So what is the reading for this combination?
Zai-jin? Zai-nin? Zai-hito? Zai-bito? Zai-ri? Zai-to? Tsumi-jin?
Tsumi-nin? Tsumi-hito? Tusmi-bito, Tsumi-ri? Tsumi-to? In
this case there are at least two accepted readings, zainin
and tsumibito. The first is translated as "criminal,
offender; sinner" and the second only as "sinner"(Halpern 617). In
Japanese, then, it seems that a distinction can be made between
"criminal" and "sinner" (this distinction
apparently is not made in Chinese--which language, you may remember,
usually only permits one reading for a character). But how does
one know WHICH reading is intended when this combination is encountered?
Well, Japanese writing does make it possible to indicate that.
When very rarely used kanji are written, or in the writing
of names, or for other situations where clarification of the reading
would be helpful, the kanji may be accompanied by small letters
of the hiragana syllabary. Thus in a bilingual comic-book version
of Lady Murasaki's Tale of Genji (possibly the world's
first novel)(Yamato:2)
I noted the following included in a vertical line of text:
The graphics to the right of the two kanji are letters of the hiragana syllabary. From top to bottom they read tsu, mi, bi, and to and thus indicate that this kanji combination should be pronounced tsumibito and thus refers to "sinner" rather than "criminal." (Maybe!)
The phonetic graphemes attached to kanji are called furigana. For more on the use of furigana in Japanese click here.
Had enough? Well ... this never seems to end ...
So here's more.
I did mention earlier that a single sound, such as KAN could be represented by 133 distinct kanji. What I didn't mention is that combining sounds doesn't get rid of all the problems of ambiguity in the spoken (or phonetically written) form of the language. Thus, in Nelson's dictionary there are seven different kanji combinations that can be read as kan-ji (as illustrated above) in addition to three kanji, each of which can be pronounced kanjiki. Or take the sound combination kan-sei. Twenty kanji combinations represent this sound combination; they have meanings such as war cry, dryness, government organization, poverty-striken student, a cold voice, sensitivity, shout of joy, the sound of a valley stream, snoring, etc. Thus sounds, or merely phonetic representations of sounds of the Japanese language can leave room for much ambiguity. Kanji, however (in spite of their varied meanings) may make for less ambiguity in communication. OK ... especially if combined with furigana supplements!
CHINESE PHONETICS
Before leaving (for a while) a couple of more things. First of all, hundreds of years ago Chinese linguists did invent a system specifically for representing sounds so that they could indicate the sounds of obscure characters (for those who did not know them) or the sounds of foreign languages, more accurately than they could by using the whole sounds of characters as phonetics. Using several hundred characters, whose whole sounds would be known to linguists, they divided them into two groups. One group of characters was used to represent only the initial consonant sounds of the characters. The other group was used to represent the "rhyme" sound--that is, the part of the character (if it began with a consonant) that represented only the vowel and, if present, terminal consonant, sound of the characters (which, remember, represented only "monosyllabic" sounds). This system is called "fanqui" ("turn and cut")(Ledyard 36)
At this point, I will just present a brief example of how this could work. Take the English sounds for the Arabic "characters" 2, 5 and 6 to represent only initial consonants, and the characters 4, 8, and 9 to represent only the "rhyme" sound (eight has no initial consonant, so we take the whole sound of that character). We could thus represent the sounds of the English words , "tore", "Tate", "tine", "four", "fate", "fine", "sore", "sate", and "sign" with the following character combinations: 24, 28, 29, 54, 58, 59, 64, 68 and 69. (See Ledyard 36 for some more on this and a similar example.)
There is yet another phonetic system used by the Chinese (other than Roman letters) to represent all the sounds in Mandarin; it was developed in modern times and is called the Chinese National Phonetic Alphabet. It appears related to the older system, having 23 initial consonants and 35 "rhyme sounds" (vowels, or vowels with final consonants such as n, ng and r--which appear to be the only syllable-final consonants in Mandarin Chinese). These symbols do not include means of representing tone in Mandarin. See Choy 24-30 for a presentation of the CNPA.
This is the other--and final--thing for now. The Koreans adopted the use of Chinese characters perhaps some 2000 years ago. The use of Chinese characters to represent Korean sounds was cumbersome and difficult--something for the specialists. In the 1440s, however a new--and alphabetical--system of writing, based on a thorough linguistic analysis of many features of the spoken language and how these could be most effectively represented in writing so that, as was said at the time, even "a stupid man can learn them in the space of ten days" (quoted in Kim-Renaud, 3), was promulgated by the ruler of Korea, King Sejong. One linguist, Sampson, called this "one of the great intellectual achievements of humankind (quoted in Kim-Renaud, ix). This is not the place to present the science and research upon which this innovative new system of writing, now called Hangul was based, nor is it the place to discuss exactly who contributed to it in addition to the king (and exactly what his role [perhaps the principal] in it was). Rather, we will have a very brief look at the major features of this unique writing system.
Hangul consists of an alphabet, with letters for all of the phonemes--consonants and vowels--used in Korean speech. The letters of the alphabet are then arranged in blocks, each of which contains a syllable. This differs from Japanese kana systems in that in hangul each sound of a syllable has its own grapheme. Japanese kana, on the other hand, cannot be broken into their distinctive sound elements (phonemes). Thus also unlike kana, which is limited to representing few more than 100 syllables, the Korean system can easily be used to create thousands of different syllable blocks. This is an essential part of the hangul system because, with a more extensive sound system (including more vowels and consonants and consonant clusters) than Japanese, a simple syllabary in Korean would require thousands of indivisible graphemes to represent the Korean sound system (the same would be true for English). As it is, then, to write Korean one must learn only a handful of letters, and the rules to follow to insert them into syllable blocks.
The figure at the beginning of this section illustrates the two hangul blocks which represent the term hangul (meaning "Korean" [or "great"] "writing") itself. To the right is hangul written unblocked (as is usual for non Korean alphabetical systems.) Note how these alphabetical elements fit into the hangul syllable blocks. The rules for placing the alphabet letters into blocks are clearly presented at http://catcode.com/kintro/first.htm.
Why put them in blocks? Why not write them unblocked as is done with other alphabetical systems? I don't know! When I find out, I'll put it in here. But there is one interesting bit of research which might bear on this question. Students at various levels of learning Korean could read the blocked (syllable) forms on average between two and three times more quickly than they could read the unblocked forms (Kim, 151)!
Finally, I should point out that in South Korea, as in Japan (but probably to a somewhat lesser extent), Chinese characters may still be used in the writing of ordinary Korean text (Chinese characters are not used for writing Korean text in North Korea [Sohn 215]).
A few preliminary comments on Chinese and Japanese sound systems ...
According to my quick analysis Choy's
dictionary, in order to pronounce the sounds of the 3210 characters
listed,Mandarin Chinese uses about 400 monosyllable sounds, while
Cantonese Chinese uses about 800 (compared to 100 or so syllable
sounds used for all the kanji (sometimes, however, using more
than one syllable) in Japanese. Given that Mandarin has 4 tones
which are phonetically distinctive, that gives Mandarin a potential
of 1600 distinctive ways to pronounce monosyllabic characters.
And given that Cantonese has at least 6 tones (more in some technical
senses ...) it potentially has about 4800 distinctive ways to
pronounce monosyllabic characters. I should note, however,
that not all of the tonal possibilities are used in either Cantonese
or Mandarin in order to represent the 3210 characters. In fact,
for those 3210, Cantonese used about 1600 distinctive (when tone
is added to the sound system) sounds, thus having an average of
about two characters per distinctive sound (this is just an overall
average: some sounds are associated with far more than two characters).
The ratio for Mandarin was around 4 characters per distinctive
sound. These calculations are relatively straightforward for Chinese
in that only about 10% of the characters are associated with more
than one monosyllable. In Japanese, on the other hand, not only
is there a much smaller set of monosyllables to go with the kanji,
but also, each kanji (probably more than 90% of them) has two
or three or four or more distinctive sounds associated.
