Two nights ago, I mentioned the Chinese name for 'Crest'


Mandarin Jiajieshi

Cantonese Kaaikitsi

'good clean person'

which disappointed me because it didn't sound much like Crest.

I like sinifications that sound like the foreign original and have relevant meanings:

Sinification Not meaningful Meaningful
Not phonetically similar -- -+
Phonetically similar +- ++

Examples of the four types:

1. Not phonetically similar/Not meaningful: I can't think of any. Why make up a name that has no relation to the original and makes no sense?

2. Not phonetically similar/Meaningful: These fall into two categories:

- Chinese pronunciations of Chinese character names from other languages: e.g., トヨタ Toyota (originally Toyoda, spelled 豐田) becomes Chinese 豐田 'flourshing field', pronounced Fengtian in standard Mandarin.

- Calques of foreign names: e.g., Microsoft becomes 微軟 Weiruan 'micro-soft'.

3. Phonetically similar/Not meaningful:

- e.g., Mandarin 匈牙利 Xiongyali 'chest tooth profit' for Hungary. Phonetic similarity may be more apparent when read in another Chinese language: e.g., Cantonese Hungngalei.

8.29.3:27: Maybe this isn't the best example. The Magyar are Hungarians in English because of a supposed connection with the Huns (Chinese 匈人 'Xiong people') who in turn have been connected to the 匈奴 Xiongnu < Old Chinese *hoŋna. The supposed connection between all three is apparent in Chinese: 匈牙利 Hungary, 匈 Hun, 匈奴 Xiongnu.

Czech and Slovak are Mandarin 捷克 Jieke 'swift overcome' and 斯洛伐克 Siluofake 'this Luo attack overcome'. The Luo River is in China, not Slovakia.

Names in this class tend to be transcribed with the same characters: e.g., si 'this' is used to represent a foreign s in Mandarin

斯里蘭卡 Sili Lanka 'Sri Lanka'

斯拉夫 Silafu 'Slav'

斯德哥爾摩 Sidegeerma 'Stockholm'

斯大林 Sidalin 'Stalin' ('this great forest')

斯堪地那維亞 Sikandinaweiya 'Scandinavia'

斯巴達 Sibada 'Sparta'

which are the first six names that came up as suggestions when I typed 斯 in Windows Vista. The suggestion box is helpful to avoid the tedious process of typing additional characters. Characters like 斯 are almost like Japanese katakana; a string of them is likely to be a foreign name.

4. Phonetically similar/Meaningful: These fall into two categories:

- Meaningful but not (obviously) relevant: e.g., martini becomes Mandarin 馬踢你 matini 'horse kicks you' (doesn't sound like a drink).

Mandarin 歐寶 Oubao [owpaw] 'European treasure' sounds like Opel and has a nice meaning but doesn't tell us that an Oubao is a vehicle.

The first two-thirds of Mandarin 奥地利 Aodili 'inner land profit' describe landlocked Austria but li 'profit' only serves a phonetic function.

- Meaningful and relevant: Mandarin 基因 jiyin 'basic cause' sounds like 'gene' and even implies what a gene is.

8.29.4:13: Many Chinese words can be converted into Japanese, Korean, or Vietnamese equivalents simply by reading their characters in those languages. But this is not the case with 'gene':

Mandarin 基因 jiyin 'basic cause' > 'gene' Predicted counterpart Actual words for 'gene'
Japanese 基因 kiin 遺伝子 idenshi 'bequeath transmit (noun suffix)'
Korean 기인 (基因) kiin 유전자 (遺傳子) yujŏnja 'bequeath transmit (noun suffix)'
Vietnamese cơ nhân (基因) gen, gien, di thể̉ (遺體?) 'bequeath body', chất di truyền (質遺傳) 'substance bequeath transmit'

The Vietnamese reading of 因 has nh- < *ʔj-.

The Korean word was probably borrowed from Japanese (though read in Korean). This is one of a number of Japanese-Korean pairs that have no counterparts in China or Vietnam. Korean still uses the graph 傳 that has been simplified to 伝 in Japanese.

The first two Vietnamese words are obviously from English gene or French gène.

I don't know for sure what the source characters are for the third Vietnamese word, which may have been coined after Vietnamese stopped using characters. I can't find it in Nguyễn's 1966 dictionary though it's in Bui's 1992 dictionary. This suggests it might have been coined after 1966. I assume the first character is 遺 on the basis of di truyền (遺傳) 'genetic', a term common to all four languages. But the Chinese, Japanese, and Korean counterparts of di thể̉ (遺體) mean 'remains of the dead', not 'gene'!

The huge 1993 English-Vietnamese dictionary compiled by the

Trung tâm Khoa học Xã hội và Nhân văn Quốc Gia Viện Ngôn ngữ học

(中心科學社會 và 人文國家院言語學; và is the only native word] )

'Center Science Social and Humanities State Institute Linguistic'

'Linguistic Institute of the National Center for the Social Sciences and the Humanities'

lists the fourth word which is unique to Vietnamese and has Vietnamese modified-modifier word order ('substance genetic' rather than 'genetic substance'. STATUS TWINS HUT

While researching last night's post, I discovered Thomas Chan's 2001 MA thesis on Cantonese characters. From a non-Cantonese perspective, these characters have unusual uses or are totally alien. For example, last night I mentioned 結/纈 which have a special reading lit 'knot' in addition to their usual reading kit. In Mandarin, neither character has an l-reading: 結 jie and 纈 xie. Although lit is a strange reading from a Mandarin perspective, it is ultimately cognate to the other readings:

Character Old Chinese Cantonese Mandarin
*kʌ-lit kit jie
lit (none)
*N-kʌ-lit (none***) xie

Not all unusual Cantonese readings have Mandarin cognates. Not far below lit in Chan's (2001: 158) list of Cantonese characters is 孖 ma 'twin'. Cantonese ma normally corresponds to Mandarin ma, but the Mandarin reading of 孖 is a completely unrelated zi (cognate to Mandarin 子 zi 'child'?; 孖 is a drawing of two 子 children).

孖 puzzled me many years ago when I first saw it on the sign of Waimalu Chop Suey in Hawaii. I later figured out that


Ct Waimalou

lit. 'status twin hut'

was a Cantonese transcription of Waimalu*. A Mandarin speaker who didn't know that 孖 was ma in Cantonese might read it as Weizilu, assuming that speaker knew the rare character 孖**.

*8.28.0:14: Written from right to left in the sign:


餐廳 tshantheng is 'restaurant', not 雜碎 'chop suey'.

**孖 isn't among the c. 8,500 characters in the 1971 edition of the pocket dictionary 新华字典 Xinhua zidian (New Chinese Character Dictionary).

8.28.00:28: If the Vietnamese still used Sino-Vietnamese pronunciations of Sinified names: e.g.,


Ct Wasingteun 'Washington' (there is no ton in Cantonese)

> Viet Hoa Thịnh Đốn

then 位孖廬 'Waimalu' would be Vị--Lô in Vietnamese.

I assume the Vietnamese reading of 孖 corresponds to Mandarin zi rather than Cantonese ma. I have yet to see a Vietnamese reading derived from a Cantonese-specific reading.

8.28.0:40: Chan (2001: 49) mentioned a Cantonese reading tsi (ji in his romanization) for 孖 that is cognate to Mandarin zi. Is tsi extinct? I can't find it in A Chinese Talking Syllabary of the Cantonese Dialect which only lists ma (maa in its romanization).

***8.28.0:46: The early Cantonese dialect that was the source of Sino-Vietnamese had a reading like *ɦiet < *get < Old Chinese *N-kʌ-lit which was the source of SV hiệt. This *ɦiet should theoretically have become Cantonese hit, but I can't confirm such a reading. It may have gone extinct as more people read the uncommon graph 纈 as kit by analogy with its common phonetic 結 kit. MY TEETH NEED A GOOD CLEAN PERSON

Who is this 'person'? It's actually a product I use every day. I saw its name in a Chinese ad on Wednesday.

Hint 1: The Chinese characters for it are


'good clean person'

Hint 2: They are pronounced Jiajieshi in Mandarin, which doesn't sound like the original English word. Their Cantonese reading Kaaikitsi is closer, especially if the second k is replaced by an r.

Select the space below for the answer.

佳洁士.is the Chinese name for 'Crest'. The second character is a simplification of 潔. I don't know why an l-character wasn't chosen as an approximation of -re- in Crest.

8.27.1:09: When I was trying to think of a Cantonese transcription of Crest that more closely resembled the original, it occurred to me that 洁 kit is homophonous with 結 kit 'knot' ... which can also be read lit:

English C re s(t)
Cantonese kaai kit si
Chinese characters

Was 結 kit/lit replaced by a semantically more relevant homophone 洁 kit 'clean'?

I was able to Google 佳结士 with 结 'knot' which I presume is a misspelling for its homophone 佳洁士 with 洁 'clean'.

lit 'knot' also has another spelling 纈 distinct from 結 kit/lit (but with 結 as phonetic).

kit and lit are both from Old Chinese *kʌ-lit. The Zhongshan dialect form khə-lit 'knot' (Chan 1984: 302; found in Sagart 1999: 103) comes close to preserving this original form (though I suspect Zhongshan i is not a direct retention from OC *i, but was 'recreated': i < *ie < *e < *i).

結 Mandarin jie [tɕje] is also from *kʌ-lit:

jie [tɕje] < *kie < *kiet < *ket < *klit < *kʌ-lit TANGRAPHIC RADICALS 15: GAX FROM TAX

David Boxenhorn pointed out that last night's analysis derived

tax = due + bil

which would have been a better post title.

How many other hidden 'messages' are in tangraphic analyses?

tax in turn is the source of the bil within gax in 1021:


1021 1lhew 'miscellaneous, mixed' (bombaegax; gax = ciebil) =

0993 1lhew 'to herd, graze' (bombaecie; phonetic) +

5851 2nieʳ 'all, various' (taxgak; tax = due + bil; semantic)

Is bil really semantic in 1021? Why not abbreviate 5851 as the unique radical tax instead of the more common bil found in at least 262 other tangraphs? Why isn't 5851 just tax instead of taxgak with a gak that doesn't do much to clarify its meaning?

The Tangraphic Sea derives 0993 from 1021, though I am suspicious of complex-to-simple analyses:


0993 1lhew 'to herd, graze' (bombaecie) =

1021 1lhew 'miscellaneous, mixed' (bombaegax) +

5578 2nwiə 'to herd' (voxpok; vox = biobaebum) +

1514 1lɨə 'suppress' (hilbalcieces)

gax appears in only two other tangraphs. 4578 is just a variant of 0671 with gax instead of jai.

2zəuʳ 'message, letter' (; analysis unknown)


4578 = 0671 1sa 'to swell; to choke' (gaxjom = jaijom) =

1972 2rieʳ 'swell' (baebaabilbaebil; baebaabilbae = gax or jai; semantic)

2696 1dʒi 'narrow' (qaljom; semantic) TANGRAPHIC RADICALS 14: EATING VARIOUS HIGH TAXES

The only tangraph containing the radical tax (that's just David Boxenhorn's three-letter code; its meaning is unknown) is 5851:


5851 2nieʳ 'all, various' (taxgak; tax = due + bil) =

5856 2ɣa 'in' (teacok; tea = due + three extra strokes ㄱㅡㅡ) +

4617 2lhiụ 'swallow, eat, take' (bukbelbil; buk = 'mouth', bel = 'surround') +

0428 1dziụ 'tall; high' (boagak; looks like 'speech' [why?] + 'high')

Were 3/4 of tea (= due) and bil (meanings unknown) really fused into tax? None of the three source tangraphs sound like 2nieʳ or have relevant semantics except for 'high' (implying 'many' > 'all, various'), so I think this analysis is dubious. If so, what is the real analysis of 5851 and why was a unique radical (tax) created for it? THE GOLDEN GUIDE: LINE 78: TANGRAPHS 386-390

78. I translated the object-verb sequence 3912-0123 as a single verb. Although the structure is

(Subject + Object + Verb) + Adjective

'The walking of the Khitan is slow'

my translation is meant to parallel that of line 77 ('The brave and strong Tangut go on parade').

Tangraph number 386 387 388 389 390
Li Fanwen number 4361 4389 3912 0123 1762
My reconstructed pronunciation 1tʃhɨə 1tã 2bi 0dʒɨi 2lwəi
Tangraph gloss first half of 'Khitan' transcription tangraph step, pace to walk, go slow; lazy
Word Khitan
Translation The Khitan walk slowly.

386: 4361 has an unexpected affricate corresponding to the stop of 契丹 *khɨttan, the Chinese name of the Khitan. Since I don't know of any Tangut-internal affrication, I assume that the affricate reflects palatalization in the Khitan dialect that 1tʃhɨə-1tã was borrowed from.

4361 was analyzed in terms of the tangraph that obligatorily follows it plus a phonetic:


4361 1tʃhɨə (first half of 'Khitan') (boxbeehel) =

4389 1tã (transcription tangraph) (boxyinbaxbelcin) +

1374 1tʃhɨə 'that' (beehel)

387: Why do both 4389 and 4361 have 'wood' on top? Did the Tangut associate that element with the Tangut?


4389 1tã (transcription tangraph) (boxyinbaxbelcin) =

4250 1si 'wood' (boxdexdexcok) +

1439 1tã 'unlined garment' (yinbaxbelcin; loanword from Chn 單 'single', as in 'single-layer')

388: 3912 (tone and analysis unknown) consists of doe 'leg' plus the rare radical zeo.

389: Although the analysis of 0123 is unknown, it is a source graph in at least six analyses that derive five radical (sequences) from it:

0123 0dʒɨi 'to walk, go' (bamdexdoodex) >

1. bam in 0086 1ʃwɨõ 'patrol' (bamdumdexcok)

2. hun = bamdex in 0360 1bɨu 'crawl, creep' (herbaehun)

3. 2/3 of wou = bamdexbil in 0681 1ʔieʳ 'to fall, sink' (wougoi)

4. yok = bamdexdoo in 0676 1vɨe 'to go' (yokqos)

5. doo in 0787 1kã 'drive' (hundoocin)

6. dexdoodex in 4347 1tʃhɨo 'to go' (boxdexdoodex)

How many of these parts are actually from 0123?

390: 1762 and its source tangraphs (one of which is its near-mirror image) have circular analyses.

What do tun 'skin' and pik 'hand' have to do with being slow or lazy? Why is 'skin (vertical line) hand' (1762) synonymous with 'not hand skin' (2965)?


1762 2lwəi 'slow, lazy' (tunbaepik) =

2965 1lwõ 'slow, lazy' (ciapiktun; tun = 'skin'; cognate to 1901 below and possibly 1762?; -w- from *P- prefix) +

3743 2kwəụ 'slow' (feobaepik; feo = 'fear', short for wom [see below], pik = 'hand')


2965 1lwõ 'slow, lazy' (ciapiktun) =

1901 2lõ 'slack' (ciapikdao; cia = 'not'; cognate to 2965) +

1762 2lwəi 'slow, lazy' (tunbaepik)

3743 has no known analysis. It is an abbreviation of 3726:


3726 2kwəụ 'slow' (wombaepik) =

6034 1dziəʳ 'speed' (womdex; variant of 2520 feodexbua; wom 'speed' = feobua) +

1762 2lwəi 'slow, lazy' (tunbaepik) THE GOLDEN GUIDE: LINE 77: TANGRAPHS 381-385

77. The name for the Tangut finally appears in the Golden Guide. I initially thought it was a surname* in this line, but looking at the following tangraphs in this line and the next reveals that 'Tangut' is an ethnonym. Finally, a sentence with a subject and a verb rather than just a list of surnames!

Subject (Noun + Adjective 1 + Adjective 2) + Verb

I'm going to violate my own rules for simplified romanization and spell 2miə-2nɨaa as Minya (cf. Classical Tibetan Mi-nyag) rather than Myna. The late KB Kepping used the spelling Mi-nia.

Tangraph number 381 382 383 384 385
Li Fanwen number 3752 3296 1959 0142 3852
My reconstructed pronunciation 2miə 2nɨaa 1kiʳ 2lị 1dʒɨe
Tangraph gloss 1st half of 'Tangut' 2nd half of 'Tangut' brave; violent brave; violent; strong to go; to parade
Word Tangut
Translation The brave and strong Tangut (Minya) go on parade;

381: 3752 looks like 2544 'sage' plus 4855 'one of two', but neither of those tangraphs appears in its analysis:


3752 2miə (1st half of 'Tangut') (geohae) =

2344 2mi 'Tangut' (geobul; cognate to 2miə?) +

1771 2siẹ 'wisdom' (helhae)

382: Are the Tangut really the 'Great (Che')on', or was 3437 chosen at random from the 33 other tangraphs with por?


3296 2nɨaa (2nd half of 'Tangut') (giipor) =

2893 2khwe 'big' (giiwun) +

3437 2ʔõ 'the surname On; 2nd half of the surname Che'on' (geopor; see line 51)

383: Could qee = bamix be derived from the left side of Chn 敢 'dare'?


1959 1kiʳ 'brave; violent (gaeqee) =

2716 1rieʳ 'skillful, ingenious' (gaedumcin) +

0142 2lị 'strong' (bamhixdex)

384: No analysis of 0142 (see above) is unknown. Could the entire tangraph be based on 敢 'dare'? Could 2lị be borrowed from Chn 力 'strength'?

385: Was zeo (= basdexbaedex) really abbreviated as qor (= basbaedex)?


3852 1dʒɨe 'to go; to parade' (qordal) =

3912 2bi 'step, pace' (doezeo)

5250 2daa 'tour, walk' (taedexdal)

There are 9 tangraphs with qor. All analyses ultimately point back to zeo in 3912 as the source of qor.

qor tangraph Source of qor
3852 zeo in 3912
3858, 3992, 4822, 5187 = 5597, 5891 3852 < zeo in 3912
3952, 4014 3992 < 3852 < zeo in 3912

Why was zeo abbreviated in those tangraphs but not in 3883


3883 1bi 'prestige' (zeofol) =

3912 2bi 'step, pace' (doezeo; phonetic) +

3354 1ɣɪ 'power, strength' (dexfoltun)

which is nearly homophonous with 3912? 3912 and 3883 are the only two tangraphs with zeo.

*There are Chinese with the surname 華 'China', Japanese with the surname 大和 Yamato ('Japan'), and Koreans with the surname 韓 Han ('Korea'), so it's not surprising that 'Tangut' was a Tangut surname.

(Actually, the Korean surname is taken from a Chinese surname whose sinograph was later used to transcribe the names of early states on the Korean peninsula. As far as I know, there are no Vietnamese with the surname 越 Việt.)

Tangut fonts by Mojikyo.org
Tangut radical font by Andrew West
All other content copyright © 2002-2010 Amritavision