99P llleRl T WalUOlllW Dl1 l CSl IWDV Dfiil$GJ L 'IDGJ u Cfll l lWCfll II OOl ul Wu IE ku Ill ' J Non • - Responsive J • BASIC PATTERNS OF CHINESE CODES AND CIPHERS • •• William T Mau ••••• • 1 CAN YOH MAKE QIIT THE NAME fPU7 7 le 1111 11011nmfl 01n 1m1 1011·71•• e 11e• 1 u 1111 1 lenn Em rv a D I 15 u 18 4 D J L J SM Z 22 I 111 111111 r N MS M trtt a str Declassified and Approved for Release by NSA on 03-01-2021 pursuant to E O 13526 MDR-109366 BASIC PATTERNS OF CHINESE CODES AND CIPHERS William T Mau B4 0 5 ATG-AXI 0900 IUIIOI 0502 0503 0504 0505 0 508 0507 0508 0509 t 0510 0511 0512 0513 0514 0515 0516 0517 0518 0519 JI C '1J rJJ 1P sf NJ l1JJ 7J J ' Ji ATG ATH ATI ATJ ATX ATL ATM ATN ATO ATP f J J J Kb Jh NJ ATS ATT ATU ATV ATW AT X ATY ATZ ATQ ATR • 0520 0521 0522 0523 0524 0525 0526 0527 0528 0529 1h 1 Mi- AUH - AUi AUJ th 1 J I· AUA AUB AUC AUO AUE AUF il # h lb I' 1211 AUK AUL AUii AUN AUO AUP AUO AUR AUS AUT • 0530 0531 0532 0533 0534 0535 0536 0537 0538 0539 0540 05 41 05'Z SQ J hJ JJ ilJ os s 0546 0547 0548 in AVC AVB 11 '7J 1t lt - 0544 0549 frl AYO AUW AUX AUY AUI AVA 0550 0551 0952 0553 0554 0555 0556 0559 0558 0559 AUU AUV C li AVE AVF AVG AVH AV AVJ AVI AVL-AVM AYN 0560 C561 0582 0563 0564 0565 0566 0567 0568 0569 f§ llf ii I 11 81 fl lfi jtai AVO AV P AYO AYR AVS AVT AVU AVV AVWA'IX 0570 0571 057Z 0573 os1 0575 0578 0577 0578 0579 C I I AVY AVI AWA AWB AWC AWD AWE l f AWF WG AWH 0510 0581 0582 0583 0584 0585 0586 0587 0588 0589 -Hr ff- if- if -- 14 AWO·AWP AWQ AWA AWi AWJ AWK AWL AWM AWN 0590 0591 0592 0593 059' 0595 0596 o I osri 0598 1$ rt 1' t5 F AWS JIWTlAWU AWV AWW AWX AW'I AWZ AlCA AlCB flhsn is a cods not a co k7 Whsn it's Chinue plain tszt As many but not all readers knor l in ordinal f com-municati the Chinese use a foia-digit cods to NpNasnt the thousands of cha raotel'B of thsi r written Za iguage That code is the Standard T6legmp1tia Coda S'PCJ a page of which is eho Jn at left Sinoe C1tinsse charaote s oamwt be sent by te tegruph this set of digi tat equiva 'tents was BBt up to make teZecOllfllUnications possible 'rhe fm u -digit groups are what Ile usually see in traffic ths trigrophs below the charaote rs are an alternate set of equivalents NNly IU1Sd Sotr1e vereion of this codll ha8 been in use in China ewr sines ts l efll 1PhY introduaed thsl'6 The Chinsse Nationalists uas t u oldsr ncrc11 Chinese Telegmphic codll also known as the Hing Cods from thB Chinese words on the cover of the book MING NA plain aodll which contains older folffl8 of many oharactez s and wmch reads from right to left mt'h8r than left to pight The STC book consists of 100 pages ion each of which is a 10 % 10 nutm i i th single-digit l'oi l and column coordinate11 In eaah cez i is one ChiMBS ohaNcts r The aharacuN ars arranged in radioal stltoks ords r tha most ooMIIOn dictionary ordll r in Chinese use wi-th mf0%'tunatsly some outof-Ol'dll r e oeptiona The basic charaate rs oacupy the fil'st 9 pages t'h8 Nmaining pagsa oontain shol't foms additional aharactsl's and special tables inal uding date tt-s and r wtks of punctuation and the Latin alphabet 1« 1s The l OOt o r bass of a C1ti114Jse aha raoter is the ''radical T1isre are 1 14 radiaah most of which aN charoateN by themaelv11s i e they have a 111Baning But thei r meaning is e tsnd11d or modified by ths addition of one ol' more strokes to fOI'fn anothsri charaDtsl' The numbsl' of Ndioals wae sstabl ishsd about 1660 and thsizo diatiO llll'll ozoder is fu ad by ths nunberi of et rokse i t takes to rn i te eaDh 0'118--f ram a single stroke to l - - - In STC sach Mdi caZ appears as a section hsading and the charaote s r ihich fo llau it aN 1U'N11 1Bd baaicatty in asoendi ng PV lbe r of et rokea added to the basia root CarpaH this wl th the English a1 p' iab6t 1anth it8 26 lette s only ti io of 1 17tioh ha1X1 mscming '1a6n thsy stand alons Feb-Mar CRYPlOLOG •Pagel fOP 8Mllt1Jf WA NP 80811 1' UJ IRlk Thie articte e oami nes certain features of Chinese-Zangua ge 11'1 Ptographia aods systems Each type is treated separateZ y but in actuaZ pNctice a aods system often inoo'l ws or moN of the mBthods of encryption discussed beZ oi In addition P O'l' more methods for encoding values not in the IIOCabuZa ry oan be used in a singZ aods system I BASIC CODE PATI'ERNS Total Strokes Order As we all learned in early training codes can be one-part or two-part One-part codes are so formatted that one book suffices for both encoding and decoding In two-part codes the order of plain and code equivalents is so mixed that two books are required one for encoding and one for decoding A modified one-part code is one in which the regular pattern has been complicated in SOllle way Most Chinese codes are one-part But one of the most important questions facing the codebreaker as he looks at a new Chinese code is this Assuming that it is one-part and that the values therefore occur in logical order rather than scrambled what is that logical order In European languages it would be some form of alphabetic order but Chinese has at least five possible ways of arranging a one-part code Radical Stroke Order The codes used by the People's Republic of China PRC may be arranged in the usual radical stroke order which parallels the STC book but other systematic orders--even other radical stroke syste s--are possible Compotmds can be inserted between single characters For ex111ple a given row of ten values in the STC book is 0100 0101 0102 0103 etc through 0109 'The corresponding values in a radical stroke order code might be 0100 0100 0226 0101 0102 0102 7022 etc Phonetic Order An alphabetic order for Chinese-language values can be achieved by one of several systems of phonetic representation of characters Stransely the system usually seen is not the Pinyin introduced in 1958 with which we at NSA prefer to wOTk for convenience' sake but the older National Phonetic system desisned in the early 1920's in imitation of the Japanese kana STC books published in China contain these 37 phonetic symbols in the otherwise blank cells 9720--9756 'Thus while the svmbols in the names of Mao Ze Dong Shanghai and Beijing Peking Peiching would occur in the same order in a Pinyin listing as in English alphabetical order Bei Dong Hai Jing Shang Ze under the National Phonetic system they would be listed in the order Bei Mao Dong Hai Sbang Ze This systn is known frOII the first four syllables in it as the Bo-Po-Mo-Po system The same na11es in ight be ordered by the mmber of strokes with which they are written using the number of strokes in the new circa 19S8 short forms where applicable Thus arranged they would read Shang 3 strokes • Mao t c J 4 strokes Bei Dong $f- # 5 strokes Jing c- t ' S strokes 8 strokes Ze 8 strokes Hai 10 strokes Characters with the same nWllber of strokes may be arranged within that category by either phonetic or radical order Sentential Category Order Many PRC codes especially Dlilitary codes are in the form of charts rather than books Dimensions vary but the 9 x 9 -trix is most comon The usual number of matrices in a code of this type is fro• six to nine Such charts frequently use popular phrases and sentences to fill in the rows of the matrices Categories such u time year month houT and points of the compass iast south west nor are listed in adjoin ng ceITsTn other 111 tr1ces Of the characters used in our nlllles exa11ple above Dong the literal meaning of which is east and Bei literally north would be found with south and west 'The sentential cateaory code is often referred to as IIOdified one-part because it ls one-part requires only one book to has so• eone who has memorized the sentences and has the chart in front of hi • C IIPARATIVE ORDER OF nm CHARACTERS IN nm NAMES ''Mao Ze Dong Shanahai Beijing'' STC and Radical Stroke SHANG 0006 J Pin- Nat'l Total Sentential yin Phon Stroke BEi Bl I SHANG 3 MAO JING 0079 t DONG MAO BEi 0S54 It HAI DONG st JIN HAI DONG S JING JING 8 MAO SHAHG MAO 2639 3029 HAI 3189 SHANG SHANG ZE 8 BEI ZE 3419 Peb-Mar • CRYPTOLOG Page 2 flP 91i81HR' Wllilll t ZE MAO 4 DONG BEI S ZE HAI 10 ZE DONG HAI JING EXAMPLE Of SENTENTIAL ARRANGEMENT 1 2 3 4 5 10 t l l· I i IJ It 11 • 1A II •• ' - K 12 11 t t t l 46 t 13 If t If If j r ff ti - J 6 ij t · or dinomes Soaetimes the code groups for digital values are used with a flag group to spell the digits of an STC group A device used in pre-Co1111unist codes was a table of radicals to be added to the character represented by the preceding code group Sincei the spell tables can represent the Chinese language fully they give the effect of a cipher within a code · I _ I T t '••·· ' ------· I _ -- A chart containing all the initials and finals that compose National Phonetic representations of Chinese characters peraits the user to spell the so D d of a llissing character Reconstruction of a chart such as the one which follows greatly assists the bookbreaker•s recovery of values in the basic system For our own convenience we at NSA write the cell contents in Pinyin instead of using the actual phonetic s 'llbols ·- Une 10 AH units are to pay attent ian t o concealing their equipment and stN igt hening air defense Do l'ICt e poae targsts Line 11 Our advano• ia bloal gd Request artillery aupporot to help in aompZ eting our mission Line 12 Ths siti ation has suddenly ahanged CaauaZ tiee are heavy Une 13 four unit is direated t o aompZ ete aombat pz epai atio is at onoe 114 -132 105 116 l fl If 10 _ ENCODED TEXT Phonetic Representation 10 20 10 I io 40 so eo IC 20 30 M J H C 70 ' IN I IG L IH 9G G R IO 60 ual 70 10 lu ao ul Q A Ing an D X 0 la I 40 % al Uo N CH I Request your-unit increase support u G u• ·- -· -- un lu ua PROVISION FOR ADDITIONAL VALUES In practice the reconstruction Qf codes is facilitated tiy the fact that the code vocabu lary is not all-inclusive some of the characters in a aiven message aay not be in the code vocabulary at all and the code clerk must have1 some means of encoding the absent characters 'Jbere are several ways to deal with lllissing characters and each ll ltes the bookbTeaker•s task a bit easier Among them are the phonetic spell table phonetic variation character construction imd dissection and enciphered STC I Another wa y to use sound rather than meaning is phonetic variation A special flag group indicates that the group which follows is not to be read as its true value but that the character intended is a homonym of the plaintext equivalent of the group sent For ex11111Ple in the phrase l' J '1r ---- 8490 the flag group 8490 indicates that the next group Zhong is being used instead of the intended character 'f' fit cf7 which is also pronounced Zhong but which is not in the code vocabulary Often one I digit in the flag group will change to reflect Chinese codes often contain a subsystem fOT a specific tone aaong the the four tones that spelling out characters which do not occur in the intended character has The flag thus 110re the code vocabulary Most collUIIOn are phonetic f exactly identifies the intended character B1DOng its ho110n 'IDS values or substitution tables for STC monomes Spell Tables Feb-Mar CRYPTOLOG Page 3 Another Chinese cipher not infreauently encountered is repaginated STC The repagina Using flag groups to select only a part of tion may be merely an end-around shift of the the preceding character the part itself being page-nWDber sequence or it may involve a random a separate character is known as character dis- scramble of the page numbers Either shift or scramble may extend to the coordinates of row section Character construction on the other or column or both on each page hand involves flag gToups that instruct the recipient to take part A of character X and Local transposition within a group or the add it to part B of character Y to create the insertion of nulls can disguise the basic STC intended chaTacter An e xaaple of dissection group Thus with transposftion cabd 6153 0132 might break out to Take the right-hand side of 0932 0171 ® becomes 5613 3012 HAI $ f- 11 which gives us MEI An 3092 7011 Nulls can stretch each arou to five example of character construction would state diiits Insertint a O between band c in the Take the left-hand side of HAI and add iroups of our example gives us 61053 01032 01071 In practice transposition and it to the character JING j which gives 09032 insertion of nulls are often c011bined so that in the example given the original message would us LIANG becoe 56013 30012 30092 70011 Enciphered STC While it might be argued that the additive and repaginations belong under codes rather Flags might also indicate the presence of than ciphers we include them under ciphers STC groups within the text of coded messages the sue baSic extensive vocabularY is Such a flag would indicate that the group which because used Chinese codes selectively restrict this follows is not a value in the code but is the basic vocabulary STC group representing the character intended This would be equivalent to inserting plaintext Systems such as those discussed here have so the STC group is often enciphered by an addibeen used by the PRC in its c ications tive or by transposition Knowing some of these nethods certainly lightens the work of the bookbreaker As another writer CIPHER SYSTEMS has said Chinese codebooks are nothing more than a compilation of written characters exThere have been some substitution ciphers pressed numerically for the purpose of telellhich assigned cipher equivalents to phonetic graphic cOlllllUllications The real clue to the values but enciphered Chinese plain text in structure of a code lies in the arrangement of NSA parlance is usually some enciphel'lllent of these characters STC ffeP liiiQAE'l llllRf Additive substitution and transposition are all used An additive cipher may be as slGFle as adding a four-digit constant stutter to each group tu£1ing 6153 0132 0932 0171 I into 7264 1243 1043 1282 by the addition of 1111 it may be as secure as a one-time running key in which a different four-digit group is added to each plain STC gToup or it may be some intermediate method Character Construction and Dissection c-ft lj - t f I Non - Responsive F -11 r•••CR1Y ulM • Paze 4 I Non - Responsive I • i an you maie oul llie name A real-life puzzle eul mitted by GLENN EMERY Pl6 ABCDEF GHIJKLMNOP QRSTUV XYZ We know that in a certain non-English codebook the roman alphabet occurs as shown on a code page consisting of a 10 x 10 matrix see above The initial dinoae of a code group represents the page the final dinome column and row respectively A message is received which contains the following groups in llid-text 3742 3792 3732 3767 3742 3767 3732 3775 3772 Parentheses have been observed in other messages setting off groups which represent special categories of infoniation supplemental to the main body of the code and context indicates that the parenthetical groups in this message represent the name of a powdered 11ilk available in the Southeast Asian market So it is suspected that the groups are two words using the roman alphabet whose paae has been renUlllbered as 37 What are the words and which column and row coordinates can be recovered Solution next 110nth I 51 CDC I IS Feb-Mar • CRYPTOLOG Page 18