Date: Wed, 28 Jun 1995 18:01:10 EDT
From: "Randy Nichols, ACA Pres." <75542.1003@compuserve.com>

                      CHINESE CRYPTOGRAPHY
                             Part I


LINGUISTICS

The book "Chinese Self-Taught,"  gave me my first clue that
I might be working with an exception to the framework presented
in my previous post on Xenocripts. (1)   Instead of an ordered
alphabet, Chinese is based on thousands of characters or
ideographs.  (2)   Chinese is a "pure character" language; it is
written using no phonetic or alphabetic symbols, only characters
with an intrinsic meaning. (3)  Each character itself can be a
word.  Words are more often composed of multiple characters.
(4)  Characters themselves may be simple or composed of several
parts.  There are 227 parts of characters, called radicals.
A radical may contain additional strokes beside itself which
have meaning.  (5)   Each syllable has a maximum of two parts to
its phonetic sound, an initial and a final.  Unlike English,
each syllable can be spoken in one of four tones. (6)   Each
tone corresponds to different written characters with different
meaning.   First Tone is level, Second Tone is rising, Third
Tone is falling-rising, Fourth Tone is falling.  A Fifth
Tone is neutral (short and unstressed) used only for spoken
Chinese.  (7)   Transliteration of Chinese into English adds
diacritical marks for tone.  To the untrained ear, the sentence
"ma-ma mah maa .ma? sounds like a string of ma's, but with the
correct tones, it means "Mother is scolding the horse, right?"

In English, we have homophones (here, hear).  We have homographs
(same word with different meaning and pronounced differently ).
We have homonyms (spelled and pronounced alike with different
meanings).  In Chinese, differences in tone, and difference in
characters also imply different meaning. (8) (9)

In discussing a language where the smallest written unit is
itself a word or morpheme, the difference between a code and a
cipher becomes unclear.   How do we order the characters?


STANDARDIZATION

Since there is no standardized order for all Chinese characters,
some dictionaries, or indexes to dictionaries, list the
characters in pure stroke count order.  Characters having the
same number of strokes are usually ordered by radical order.
(10)

Another system is the phonetic alphabetic order.  The
character's pronunciation is converted to Roman or equivalent
alphabet, and then put in alphabetical order.   Problems on
this method come from dialect (more than 40 each with its own
pronunciation) and the ability to Romanize the pronunciations.
Mandarin dialect has three systems: Wade-Giles, Yale, and
Pinyin.  (11)



Mainland China has a more or less standard dialect, Putinghwa
and a romanization of that dialect, Pinyin.  Ordering the
characters is done by tone order first, and radical order
second.

There are 227 radicals in Chinese, and there is a standard order
for them based on stroke count.  Every Chinese character has
only one radical, which appears in it somewhere.  Counting pure
strokes is not an easy task. (12)

Dr. August evaluated six (6) different Encoding schemes that
have been used for Chinese manual cryptosystems. (13)  They were
evaluated in terms of ambiguity, number of information units
encoded, and code strength.  He also considered the problem of
mapping to the correct decode from the same table used for
encoding.  Complementary solutions may exist.  Dr. August
suggests that only the Four Corner System and the Chinese
Phonetic Alphabet System lend themselves to manual cryptographic
treatment.  (14)   He also suggests that  Pinyin Romanization
encoding is well suited for machine-based strategic
cryptosystems.

Dr. Atkinson suggests a fractionation procedure of the CCC
(Chinese Telegraphic Code System) to encode characters.  A
frequency analysis by stroke count (similar to Xeno approach)
and then grouped by primary CT digit could be used for
substitution ciphers. (15)

In Part II, I will cover aspects of Chinese Cryptanalysis and
in Part III, some of the historical considerations.


============================================================

(1)  H. A. Gales, "Chinese - Self Taught," Padel Book Co., New
     York, 1946.

(2)  D. A. August, "Cryptography and Exploitation of Chinese
     Manual Cryptosystems - The Encoding Problem", Cryptologia,
     Vol XIII, No. 4, October 1989.

(3)  R. Atkinson, "Ciphers In Oriental Languages," Cryptologia,
     Vol. 9,  No 4. , October, 1985.

(4)  IBID, D. A. August.

(5)  OP. CIT. , p290.

(6)  Op. Cit. , p291.

(7)  Op. Cit.,  p292.

(8)  W. L. Chen, "Standard Chinese Telecodes by Pinyin," Lynn
     International Publishing, San Francisco, 1982.

(9)  Anonymous, "Chinese Primer", Chinese Linguistics Dept.,
     Princeton Univ., 1986.

(10) S. W. Chan, "Elementary Chinese," 2nd Ed. Stanford Univ.
     Press, Stanford, 1959.

(11) "The Pinyin Chinese-English Dictionary,", Commercial Press,
     Hong Kong, 1979.

(12) IBID, Atkinson, p375.

(13) IBID, August, p294 ff.

(14) C. Dougherty, et. al., Chinese Character Indexes, Vol. 5:
     Four Corner System Index, Univ. of California Press,
     Berkeley, CA, 1963.

(15) IBID, Atkinson, p378-379.