American Association of Teachers of Japanese is collaborating with JSTOR to digitize, preserve and extend access to The
Journal-Newsletter of the Association of Teachers of Japanese.
http://www.jstor.org
Some Observations on kokuji
Author(s): James H. Buck
Source: The Journal-Newsletter of the Association of Teachers of Japanese, Vol. 6, No. 2 (Oct.
15, 1969), pp. 45-49
Published by: American Association of Teachers of Japanese
Stable URL: http://www.jstor.org/stable/488823
Accessed: 30-09-2015 12:01 UTC
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/
info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact support@jstor.org.
This content downloaded from 147.251.102.99 on Wed, 30 Sep 2015 12:01:33 UTC
All use subject to JSTOR Terms and Conditions
45
SOMEOBSERVATIONSON kokuji
James H. Buck (University of Georgia)
The Japanese writing system includes various syllabaries and Chinese
characters (kanii) supplemented by conscious creations of the Japanese
usually termed kokuii, or national characters. These inventions are also
called waji, wazokuji, or most aptly, perhaps, waseiji 'made-in-Japan
characters'.
Specific reasons for these inventions -- or the occasions for such
activity -- are by no means entirely clear, although Professor Tsukishima
has suggested that kokuji are relatively recent creations, probably because
there was no need for them during the early periods of massive borrowing of
kanji from China and adequate infusions of additional kanji were possible
during the periods of subsequent lesser borrowing. He also has written that
kokuii may have developed from a need for the Japanese to represent graphically
certain gainen 'ideas' which existed in Japan, but which did not exist
in China.1
Certainly many kokuji are of relatively recent origin, yet Douglas E.
Mills has pointed out their existence in the Shinsen Jikyoo, a dictionary
completed in the years 898-900.2 Analysis of kokuii found in current
Japanese dictionaries leads one away from Professor Tsukishima's second
point: namely, to a conclusion that kokuii were most often created to represent
words for things, not ideas. Nevertheless, many of the things so
represented by no stretch of the imagination pertain exclusively to Japan.
Not only is there some difficulty in ascertaining the "when" and the
"why" of kokuji, a problem is also presented by the "what". There is no
argument about defining kokuji as a generic term, but there is not universal
agreement concerning which individual characters do, in fact, constitute
kokuii. The most extensive list of kokuii (250) I have found is contained
in the appendix to the Lehmann and Faust Grammar.3 The Ueda Daijiten list
213 of the characters on this list, but "recognizes" only 183 of them to be
kokuji. To further complicate matters, one may find other kokuji in such
dictionaries as that by Nelson4 and in the Shin Kanwachuu Jiten5 which are
not listed either by Ueda or Lehmann and Faust. To finesse the obvious
difficulties presented by this lack of consensus, most of my observations are
limited to the kokuii recognized by Ueda.
As a percentage of total listings, the kokuii in Ueda form 1.2% (183 of
14924), a percentage nearly identical to that in Nelson (77 of 5446). This
is perhaps statistically insignicant, but kokuji present interesting comparisons
with kanji in such matters as principles of formation, grammatical
function, semantic classification, compound formation, stroke count, and
pronunciation.
Means of Formation
It is usual to classify kanji according to the criteria for character
formation known as roku-sho, the six categories traditionally assigned by
This content downloaded from 147.251.102.99 on Wed, 30 Sep 2015 12:01:33 UTC
All use subject to JSTOR Terms and Conditions
46
Chinese dictionaries. Such classifications are assigned to 74 of 183 kokuji
in Ueda. The vast majority (66) are noted as kai-i, "suggestive" or "logical'
characters in which the combination of two or more elements suggest a logical
meaning. An example of this category is hanashi (1536) 'speech, a story,
talk'. The meaning is "suggested" by combining kuchi-hen 'mouth' with shin
or atarashii 'new'(or should one interpret it as 'news'?). As a general
rule, none of the elements controls the pronounciation, although there are
exceptions, e.g. tsuma (10735) 'skirts of a garment' is written by combining
koromo-hen 'clothing' with the character for 'wife' (tsuma).
Two kokuii are listed as kei-sei 'form-sound' characters and six are
considered a combination of kai-i and kei-sei. The paucity of characters in
the kei-sei category forms a genuinely remarkable contrast with kanji as a
whole. It is estimated that 90% of Chinese characters fall into this class-
ification.
Within the scope of what are termed kokuji there is considerable latitude
for creativity. The best-known characters of this type, however, show
little imagination, principally the three dozen or so characters invented to
denote Japanese pronounciations for foreign words designating measures of
weight, capacity or distance. For instance, the Chinese character for rice,
pronounced mei, is used phonetically to represent meetoru 'meter'. For
kirumeetoru 'kilometer' the Japanese simply combined the character for mei
with sen '1000'. Similarly, other characters were combined with mei to show
the seven basic units of the metric system, from millimeter to kilometer.
This process was also followed for units derived from the gram and the liter.
Although such characters are generally well-known to Japanese, the Ueda
Daijiten excludes three of them from any listing at all. Such oversight
makes one wonder whether the Lehmann and Faust listing may not be more
reliable.
Non-metric measures also had to be dealt with, e.g. yard, mile and
pound. 'Yard' is represented by an old Chinese character (8036) given a new
reading. The character for ri (sato-hen) 'a distance of 2.44 miles' was
altered by addition of kuchi-hen and pronounced mairu, 'a land mile'.
Addition of san-sui 'water' produced a character read notto or kairi 'nautical
mile'. The word pondo 'pound' (1231c) was represented by using the
character for kin 'catty' (4362) with kuchi-hen. But not all 'pounds' are
alike, so an existing kanji (8039) was adopted to denote the British monetary
measure.
Another interesting technique for kokuii formation consists of the
invention of a single character which takes on the pronounciation of a regular
Chinese compound -- in fact, becomes a substitute for it.
Five examples of this are:
jinrikisha, kuruma (347) jinricksha
bosatsu (3015) Boddhisattva
neehan (3022) Nirvana
This content downloaded from 147.251.102.99 on Wed, 30 Sep 2015 12:01:33 UTC
All use subject to JSTOR Terms and Conditions
47
hanzoo (5178) a type of water dipper
with a hollow handle
kanioo (11450) a bill (amount due)
All of these are "recognized" as kokuii in Ueda's Daijiten, but those for
bosatsu, neehan and kanjoo are also used in China. The presence of polysyllabic
on-yomi for single kokuii is a distinct departure from the treatment
of individual kanii.
Grammatical Function
By far, kokuii were created to represent sounds of words used as nouns
-- 87% of the total. An additional 9% are verbs, with the remainder being a
miscellany which includes the interjection appare (12058), the name kume
(8642) and the personal pronoun maro (14686).
Semantic Classification
For the criterion of determining semantic classification, I have used
the rather arbitrary item of the radical under which the kokuii are listed.
Based on this loose standard, slightly more than one-half of the kokuii
belong to five radical groups, fish (38), tree (25), mouth (11), bird (10)
and metal (9). The proportion of kokuii to kanii under any given radical
generally does not vary significantly from the overall ratio of these two
types of characters. Although the 38 fish characters are 13% of the total
under that radical, the 25 tree characters form only 3% of the 925 listed in
Ueda. The extreme case is with mi-hen 'body' where 6 of 31, or about 20%
are kokuii.
Compound Formation
Kokuii are rarely used as the initial element in compounds. Ueda lists
only 17 which are so used; only two, tsuji (11928) 'crossroad' and sasa
(8420) 'bamboo grass' appear as initial elements in more than five compounds.
In all 17 cases, the pronounciation is kun-yomi.
In contrast, four kokuii appear only in single compounds and only with
on-yomi. These are the koo of ankoo (14144, 14314) 'the angler; frogfish; a
sea-toad', both characters in aikyoo (14342, 14341) 'a fish similar to the
ayu', and the u (13605) of udon 'noodles'.
Stroke Count
Most of the kokuii are relatively complicated in terms of the number
of strokes required to write them. The range is from two strokes for shime
(80) 'closed' to twenty-nine strokes for tsuki or katakumi (5554b) 'a tree
of the species Zelkowa'. More than half consist of 10 to 18 strokes. The
average stroke number of about 13.1 is not appreciably higher than the average
for the 1,850 tooyoo kanji (about 10.4). It is, however, nearly double the
average of 726 simplified Chinese characters.
This content downloaded from 147.251.102.99 on Wed, 30 Sep 2015 12:01:33 UTC
All use subject to JSTOR Terms and Conditions
48
Kokuji in Chinese Dictionaries
Several kokuji appear to have been exported to China and now are
included in Chinese dictionaries -- most importantly, perhaps, those inventions
to denote measures in the metric system. Two kokuji mentioned
earlier, pondo and mairu, are used in Chinese, but with entirely different
meanings. The character for pondo is the accepted abbreviation for t'ing
'to hear', and mairu is pronounced li and used as a final particle and in
transliterating names. Three kokuji for medical and anatomical terms have
been accepted by the Chinese, gan (7667) 'cancer', sen (9517) 'gland', and
sui (9556) 'pancreas'.
Kun-yomi and on-yomi
Nearly all kokuji bear only kun-yomi. Some exhibit the strange phenomenon
of having multiple on-yomi (such as hanzoo mentioned above). A few have
single on-yomi, such as gan and sui. Others have both Chinese and Japanese
readings. Examples of the latter are doo / hataraku (485) 'work' or 'to work'
(which combines freely as second element in compounds), byoo (12495) 'rivet'
which may also be read ribetto or kashime; and joo or gojoo (11061) 'the
command of a superior' which may also be pronounced oose. In view of the
numerous exceptions to the expectation that kokuji should have only Japanese
readings, it seems difficult to derive any generality except that, for the
most part, one may expect Japanese readings most of the time.
Other kokuji
As stated in the introduction to the Lehmann and Faust listing, the
authors included characters "which are generally regarded in Japan as kokuji,
although some may not fall strictly within that category."7 Certain other
characters were deliberately excluded which were:
". . .created by them (the Japanese) merely for purposes
of amusement, such as erebeeta gaaru, meaning 'elevator
girl' and containing elements meaning 'woman', 'up', and
'down'; and hamu raisu, meaning 'ham and rice', and made
up of the radical for 'rice' and an ideographic element
resembles the katakana symbols for ha mu."8
To be sure, I have no quarrel with such distinctions. What is interesting
is that the principle used to invent such a diverting character as
erebeeta gaaru has the sanction of previous use in recognized kokuji. By way
illustration, the following may be cited:
tooge (2579) 'mountain pass'. Consists of mountain radical
plus 'up' and 'down'.
kase (4951) 'a reel for spinning thread'. Composed of tree
radical plus 'up' and 'down'.
kamishimo (10699) 'a male dress worn during Tokugawa Era, divided
into upper and lower portions, both of the same color'.
Consists of clothing radical plus 'up' and 'down' (the
characters for which may be read, respectively, kami and shimo.)
This content downloaded from 147.251.102.99 on Wed, 30 Sep 2015 12:01:33 UTC
All use subject to JSTOR Terms and Conditions
49
kohaze (13162) 'a clasp' (for leggings or tabi). Composed
of leather radical plus 'up' and 'down'.
Additionally, among the characters included by Lehmann and Faust which
were "generally regarded by Japanese as kokuii" (but omitted by Ueda), a few
mimic precisely the principle of hamu raisu. For instance, the katakana
symbol ra is added to the ship radical to denote ranchi 'launch' (not
'lunch'). A similar example is the use of ship radical with sui 'water', to
denote the more usual three-character compound suiraitei 'torpedo boat'.
Conclusion
Kokuji are the result of an attractive creativity on the part of the
Japanese, and have served several pruposes: The designation of certain concrete
objects or living things with a high degree of specificity; the
denoting of new forms of measures; the simplification of the writing effort
(in the case of single kokuji for Chinese compounds); and the providing of
amusement.
The invention of kokuii is certainly an ancient activity, as Douglas
Mills has pointed out. A few of the kokuii "recognized" by Ueda are noted
as having appeared in the Manyooshuu, Shinsen Jikyoo, and Wamyooshoo. For
the future, however, it seems that the recognition of kokuji is much less
promising.
Only three kokuii are approved for use on the list of 1,850 tooyoo
kanii. These three are doo / hataraku (485) 'work, to work', hatake (7506)
'a dry field', and tooge (2579) 'mountain pass'.
FOOTNOTES
1. Hiroshi Tsukishima, Kokugogaku Yoosetsu, Tokyo, 1959, p. 81.
2. See the extremely interesting note by Douglas E. Mills "Some
Characters of Japanese Origin in the Ninth-Century Japanese Dictionary
Shinsen Jikyoo," Journal of the American Oriental Society, 87/3, 1967, pp.
297-300.
3. W. P. Lehmann and Lloyd Faust, A Grammar of Formal Written Japanese,
Harvard, 1951.
4. A. N. Nelson, The Modern Reader's Japanese-English Character
Dictionary, Tuttle, Tokyo, 1963.
5. Shin Kanwachuu Jiten, Kinrei-sah, Tokyo, 1955.
6. Eugene Ching, "The Simplification of Chinese Characters," Papers of
the CIC Far Eastern Language Institute University of Michigan, Ann Arbor,
1963. p. 55. The average stroke number of the 726 Chinese characters which
were simplified was reduced from 14.1 to 7.3.
7. Lehmann and Faust, op. cit., p. 4 of appendix.
8. Ibid.
This content downloaded from 147.251.102.99 on Wed, 30 Sep 2015 12:01:33 UTC
All use subject to JSTOR Terms and Conditions