1. Most of the characters of Vietnam's mythical past have anachronistic Sino-Vietnamese names. One exception is 聖揀 Thánh Gióng 'Sage Giong'. 聖 is of course a Sino-Vietnamese title for 'sage' and not a name. ButGióng is an indigenous name with an unusual nom spelling 揀.

The Sino-Vietnamese pronunciation of 揀 is giản with -n, not -ng. Normally Chinese characters with Sino-Vietnamese readings  ending in -n are not used to write native Vietnamese words ending in -ng. But in this case and others like it, I wonder if whoever chose -n characters for native -ng words spoke a dialect with an [n] > [ŋ] shift: i.e., a dialect in which 揀 giản sounded like giảng which is closer to Gióng. Codas in central and southern dialects have undergone a chain shift:

[ɲ] > [n] > [ŋ]

[c] > [t] > [k]

If 揀 originated as a nonnorthern spelling, how was Gióng spelled in the north?

Nom is usually treated as a single body of characters even though it was in use for centuries throughout Vietnam. I'd like to see that body analyzed into geographical, dialectal, and chronological strata. Nom could tell us about when and where sound changes occurred: e.g., when and where -n and -ng characters were first confused (implying [n] > [ŋ]).


Last night I somehow got the idea that the Jurchen phonogram

<gon> [kɔɴ]

might be related to Chinese 恭 <REVERENT> (old and calligraphic forms) rather than 拳 <FIST> (as I wrote about two days ago). I don't know how 恭 <REVERENT> got into my head. I thought I might have seen it when I was scrolling through Wells' "" (2011), but it's not actually there.

In Late Middle Chinese, 恭 <REVERENT> was pronounced *koŋ (cf. Sino-Korean 공 kong) - a good match for Jurchen [kɔɴ]. No need to invoke Old Chinese as I did with 拳 <FIST>. The Jurchen character could simply be based on a Parhae variant of Late Middle Chinese 恭 *koŋ.

Tonight I was wondering if a machine could be 'taught' to find potential Chinese graphic cognates for Khitan and Jurchen characters. One could start training it with Khitan and Jurchen characters identical in shape with Chinese characters: e.g.,

Then one could introduce Khitan and Jurchen characters nearly identical in shape with Chinese characters: e.g.,

Ultimately, one would then 'feed' the machine Khitan and Jurchen characters that have no obvious Chinese graphic cognates and 'ask' the machine if there are any near-matches: e.g., for Jurchen <gon>.

One could even add a phonetic dimension to the search process and get the machine to favor potential Chinese graphic cognates with readings close to the Khitan or Jurchen readings (whenever known). BAIQUIEN

Thirty-five years ago today, I got my first exposure to 連環畫 lianhuanhua - a copy of The People's Comic Book translated by Endymion Wilkinson - at a school fundraising carnival, unaware that just a short walk away, the University of Hawaii would one day have a lianhuanhua collection.

Here's a Sinification I had to DuckDuckGo because I couldn't guess what the original was: 白求恩 Báiqiúēn.

Select the blank area below to see what the original is:

白求恩 Báiqiúēn is (Norman) Bethune.

Bái is a Chinese surname that sounds like Be-.

I was surprised by qiú [tɕʰjow] for -thu- [θuː]. I would have expected t [tʰ] or s [s] intead of q [tɕʰ] for [θ]. But eventually I realized that 求恩 <SEEK FAVOR> is a meaningful verb-object sequence as well as a loose approximation of -thune.

I had a vague memory of Wikipedia having an article on conventions for Sinifying foreign names. This wasn't it, but it was interesting nonetheless. It reminded me of the 佛菻 Fulin problem that I wrote about almost ten years ago. It also taught me the term graphic pejoratives for what I've called derography (derogatory spellings). JURCHEN FIST - BARRIER BREAKER

I've been slowly copying Kiyose's (1977) edition of the Sino-Jurchen vocabulary of the Bureau of Interpreters. This forces me to take a good look at characters and think about how they're pronounced.

Entry 43 is gonkeu 'mountain pass', a borrowing from northeastern Chinese 關口 *gonkeu 'id.' (lit. 'barrier mouth'):

<gon keu>

Note how the style of the two characters doesn't match. I can't find the second character in Jason Glavy's Jurchen font, in Jerry You's fonts, or in N3696. It seems to have been overlooked because it doesn't have an entry in Jin's (1984) Jurchen dictionary. It is identical in shape to N4631 1734 which is in Jerry You's Khitan large script font, so I've made an image of N4631 1734 in lieu of crafting a Glavy-style image.


<gon> [kɔɴ],

a transcription character for northeastern Chinese *gon-syllables (觀冠館 as well as 關), resembles Chinese 拳 <FIST> (see old forms here) and even vaguely sounds like its Early Old Chinese reading *NI-kron. Later readings of 拳 do not have o-like vowels:

If the character were 'invented' c. 1119 according to the conventionally accepted scenario, why modify Jin Chinese 拳 *küen [kʰɥen] to represent the syllable gon [kɔɴ]? There was no shortage of  Chinese characters with *gon-like readings (e.g., the aforementioned 關觀冠館) that could have been phonetically more appropriate models for a Jurchen character <gon>. I think Jurchen <gon> was inherited from an older tradition going back to a time when 拳 had *o in Chinese:

stage 1
stage 2
stage 3
stage 4
拳 Early Old Chinese *NI-kron 'fist'
Serbi script
(graph shape unknown)
*<gon>-like reading (and other non-Chinese-based readings?)

Parhae script

(graph shape unknown)

*<gon>-like reading (and other  non-Chinese-based readings?)

Khitan large script


Jurchen large script


Following Janhunen (1994: 114), I regard "the Khitan and Jurchen 'large' scripts [...] as parallel, rather than successive, developments" of the Parhae script, so I do not think the Khitan large script <gon>-like character from the epitaph for the 太師 Grand Preceptor (1056) as written in Jin (1984: 17) is ancestral to Jurchen <gon>. The two, however, should share a Parhae ancestor.

The problem with the above scenario - besides the fact that the hypothetical Serbi and Parhae ancestors of <gon> are not attested - is the huge gap of over a millennium and a half between Early Old Chinese c. 1000 BC and the (unattested!) Serbi script of c. 400 AD.

It's not entirely implausible, though, that some archaic nonprestige dialect of northern Chinese preserved *o as late as 400 AD. The 8th century Old Japanese phonogram 支 ki (earlier read *ke) reflects an Old Chinese *kie whose initial had palatalized to *tɕ- in prestige dialects long ago. The practice of writing *ke as 支 originated from the Korean peninsula centuries earlier and must have started at a time when some northern Chinese dialect still preserved *k-. (支 ki in Taiwanese and other Min varieties in the south preserves the original initial to this day.)

Jurchen <gon> has variants that do not look much like Chinese 拳 <FIST>:

in line 2 of the monument commemorating the victory of Emperor 太祖 Taizu of Jin over the Khitans in 1114 (1185; the earliest attested form)

~ on the bottom of ne 11 of the monument recording the names of successful candidates for the degree of 進士 jinshi in 1224 (Jin [1984: 17, 199] writes this character two different ways, and without seeing a photo or rubbing of the monument, I don't know which is correct)

(12.5.1:13: Jin and Jin [1980: 301] have the form with ㄴ in their hand copy of the text of that monument. I can't even find the character in this rubbing.)

The most 拳 <FIST>-like form

is first attested as a transcription of the first syllable of 觀音 *gonin 'Guanyin' in line 11 of the monument commemorating the foundation of 永寧寺 Yongning Temple (1413).

12.6.21:06: APPENDIX 1: Modern Chinese o-reflexes of 拳 Early Old Chinese *NI-kron 'fist' from Xiaoxuetang:

東鄉 Dongxiang
kʰuon 24
Pu-Xian Min
仙游 Xianyou
Eastern Min
福清 Fuqing
德慶 Deqing
西岸 Xi'an

Their k(ʰ)- does not reflect the original root-initial *k-; it is from a later fused *g- < *ŋg- < *ŋk- < *NIk-.

I do not know which of those forms is native and which is borrowed.

Moreover, I do not know the details of the phonological histories of those varieties, so I cannot be certain that their -(u)o- directly preserves Old Chinese *-o-. Late Old Chinese *-wɨa- or Middle Chinese *-wɨe- could have fused into -(u)o- later.

I have excluded reflexes like 弋陽 Yiyang Gan ɕʰyon 13 with fricative and affricate initials because they are less conservative-sounding: i.e., because their initials are no longer velar. But who knows, maybe their rhymes are more conservative than their initials.

I initially thought Yiyang Gan ɕʰ- was a typo for Yiyang Gan tɕʰ-, but Xiaoxuetang lists three other  characters read ɕʰyon: 穿權棬. All have [tɕʰ-] in standard Mandarin. If ɕʰ- is a typo, it's not an isolated one. On the other hand, Xiaoxuetang has 124 characters read with tɕʰ- in Yiyang Gan and 0 characters read with ɕʰ- and rhymes other than -yon. So either ɕʰ- has an extremely restricted distribution or it is a typo for the readings of one homophone set (穿權棬拳).

APPENDIX 2: 12.9.22:18: A history of 拳 'fist' from Old Chinese to modern standard Mandarin:

1. The root of 'fist' is *kron 'roll', which does not seem to occur by itself and therefore has no characters. It is also in 卷 *CI-kron-ʔ 'to roll'.

2. A prefix *NI- was added to this root.

The prefix has to have a nasal initial *N- to account for the later voiced initial (see steps 6-8 below).

The prefix has to have a high vowel *-I- to account for the later vocalism (see step 4 below).

The prefix seems to be a nominalizer: 'roll' > 'rolled thing' > 'fist'.

But Baxter and Sagart's (2014: 54) *N(ə)- was not a nominalizer; it converted transitive verbs into intransitive verbs.

Perhaps *NI- was *mI-. Baxter and Sagart (2014: 55) reconstruct an *m- that converts verbs into agentive/instrumental nouns and an *m- for body parts. But neither of these *m- (= *mI- in my system?) are good fits: an agentive/instrumental noun from 'roll' should be mean 'roller', not 'fist', and the body part prefix is added to nouns, not verbs.

3. *o broke to *wa: *NI-kron > *NI-krwan

I follow Starostin (1989) who posits *o-breaking at the 'Classical Old Chinese' stage immediately after the 'Preclassical Old Chinese' stage which is the earliest stage in his reconstruction.

4. A prefix *NI- with a high vowel triggered vowel bending in the following syllable: *NI-krwan > *NIkrwɨan with *a partly bent up to match the height of the unknown high vowel *I

5. The high vowel was lost:*NIkrwɨan > *Nkrwɨan

6. *N assimilated to *k- (if it wasn't already velar): *Nkrwɨan > *ŋkrwɨan

7. *k- assimilated to *ŋ-: *ŋkrwɨan > grwɨan

8. *ŋ- was lost: *ŋgrwɨan > *grwɨan

9. *-r- was lost: *grwɨan > *gwɨan

10. *-a- fronted: *gwɨan > *gwɨen

11. *-ɨ- fronted: *gwɨen > *gwien

12. *-wi- fused into *-ɥ-: *gwien > *gɥen

13. The level tone developed two allophones: one in syllables with *voiced initials like *g- and another in syllables with *voiceless initials like *k(ʰ)-.

14. *g- aspirated and devoiced: *gɥen > *gʱɥen*kʱɥen > *kʰɥen; the allophones of the level tone became phonemic after *g- and *kʰ- merged into *kʰ-

15. *kʰɥ palatalized: *kʰɥen > quán [tɕɥɛn]

What I wrote as *e might have been [ɛ] all along, but I have chosen a simpler symbol since there was no contrast between */e/ and */ɛ/ in diphthongs.

Some of the relative chronology is unclear: e.g., 14 and 15 must have followed 13, but I'm not sure whether 13 followed 12 which must have followed 11. (H)EARING HIPPOS

The Wiktionary entry for ear says its Persian cognate is هوش hush 'intellect' which surprised me because I expect Persian h- to correspond to English s-: e.g., Persian هفت haft : English seven. Wiktionary reconstructs 'intellect' at the Proto-Indo-Iranic level as *Hā́wšiH 'ears; understanding' and at the Proto-Indo-European level as *h₂ṓws 'ear'. Following Beekes, I interpret Proto-Indo-European *h₂ as *ʕ, and I interpret Proto-Indo-Iranic *H as a glottal stop (cf. Beekes' /ʔ/ < *H in Avestan). None of those sounds should correspond to Persian h which is from Proto-Indo-Iranian and ultimately Proto-Indo-European *s- (e.g., 'seven' which was *septḿ̥  in Proto-Indo-European). h- in Persian hush is irregular like the h- in Greek ἵππος híppos 'horse' < Proto-Indo-European *ʔéḱwos (cf. Persian اسب‎ asb 'horse' which has no h-). (The i of híppos is also irregular.)

Could the h- of hush and híppos be by analogy with h-words with similar semantics? But what would the models be? And how did h- appear in both the Kurdish and Persian forms of the word? h- seems to be an innovation in Middle Persian hōš (Old Persian ušiy has no h-). Did Kurdish acquire h- through contact with Persian?

What got me thinking about ears was a stand-up comic on the radio joking about boxen as a plural of box. That made me check what the Old English plural of box was (boxas) and look into Old English declension in general. Here's a colorful summary. Via Wikibooks I found earan 'ears' as an example of a real Old English n-plural.

Then I started thinking about French plurals and via Wikipedia found Mickael Korvin's nouvofrancet proposal to spell all plurals with -s, among other things (like respelling the -ais of the proposal's name as -et). Is there a book like Robbins Burling's Spellbound (now only $14.67 US on Amazon!) on French spelling reform?

ADDENDUM: 12.3.23:39: Today I realized that English ear and hear are near-homophones. I'm afraid to look for a folk etymology 'deriving' one from the other. To my surprise, Wiktionary derives hear from a Proto-Indo-European compound *h₂ḱh₂owsyéti < *h₂eḱ- 'sharp' + *h₂ows- 'ear' + *-yé- (denominative suffix) + *-ti (3rd person singular suffix). The h- is all that is left of *h₂eḱ- (and it is a remnant of *ḱ-, not *h₂- which I interpret as *ʕ-). ARIRANG KOREA

Tonight I discovered the TVK2 channel which airs content from Arirang TV whose onscreen logo alternates between English and Korean. I can't find the Korean logo online. It looks something like this:


"Something", because the letter ㅇ has the same shape and size at both ends of the logo which is almost symmetrical. So all the hangul letters are on the same line, whereas in normal hangul, <ng> would be under <r.a>: 랑, not 라ㅇ. Are linear hangul logos 'in' now, or is this logo an outlier?

If the Khitan small script had survived into modern times, how would it have been computerized? Hangul blocks represent syllables, but Khitan small script blocks represent words (including inflected forms) which are far more plentiful than syllables in any language. In pre-Unicode days, the KS X 1001 encoding of Korean only allowed for 2,350 out of 11,172 possible modern Korean hangul syllables. There must be more than 2,350 or 11,172 possible Khitan small script blocks. Unicode and sophisticated character-combining fonts can handle the Khitan small script now, but how would computers thirty years ago have handled them? Would pre-Unicode computerization have popularized linearization of the Khitan small script?

Back to Korean: I saw this episode of Gangnam Insider's Picks on TVK2 which mentioned Guardian: The Lonely and Great God at 17:00. The Korean title is

쓸쓸하고 찬란하神 – 도깨비

ssŭlssŭl-ha-go chhallan-ha-shi-n - tokkaebi

'lonely-be-and resplendent-be-HON- - goblin'

= 'goblin that is lonely and resplendent'

The title is written entirely in hangul except for the honorific-adnominal suffix sequence -shi-n written with the homophonous hanja 神 <GOD>. I've never seen this kind of hanja wordplay in modern Korean before. (The use of hanja to write homophonous Korean words is, of course, a core practice of the extinct hyangchhal and idu writing systems.)

Here's that show title and much more in calligraphy. CZUCHRY

Today UPtv's GilMORE the Merrier 153-episode marathon of Gilmore Girls ended.

One of the show's stars is Matt Czuchry who "is of Ukrainian descent on his father's side." I was surprised to learn that his name is pronounced [ˈzuːkri] in English rather than [ˈuːkri] which is closer to the Ukrainian pronunciation of Чухрій <Čuxrij> as [tʃuxrʲij] (where's the stress?). I suppose [z] is a spelling pronunciation of Czuchry which looks like a Polish-style romanization. Did Czuchry's paternal ancestors come from western Ukraine?

I got the Ukrainian spelling of Czuchry from the Ukrainian Wikipedia (which unfortunately does not specify the stress). The Russian Wikipedia simply Russifies the English pronunciation of his name as Зукри <Zukri> [ˈzukrʲi] instead of Russifying his Ukrainian name as Чухрий <Čuxrij> [tɕuxrʲij].

While I'm on the subject of Ukrainian names, Wikipedia has a list of "somewhat comical" Cossack surnames. My favorite is Добрийвечір <Dobryjvečir> 'good evening'. Google shows that surname is alive and well in Ukraine today.

Tangut Yinchuan font copyright © Prof. 景永时 Jing Yongshi
Tangut character image fonts by Mojikyo.org
Tangut radical and Khitan fonts by Andrew West
Jurchen font by Jason Glavy
zAll other content copyright © 2002-2019 Amritavision