Circassian hanzi: an experiment

A forum for discussing linguistics or just languages in general.
Post Reply
dva_arla
cuneiform
cuneiform
Posts: 184
Joined: 25 Oct 2019 21:03
Location: Realm of Ideas

Circassian hanzi: an experiment

Post by dva_arla »

Foreword

Circassian hanzi was an idea that dawned upon my mind a year ago, when looking up some materials on Circassian and Old Chinese concurrently. Circassian derivational morphology is awfully reminiscent of Old Chinese, and the language is composed of mostly monosyllabic stems and roots, so why not? Though after pondering upon the idea for some time (weeks?), I dismissed the endeavour as a potentially futile one (i.e. it will join the ranks of my half-baked projects); however, as I came to re-appraise some of my past ideas the idea seemed, to my eyes, yet again appealing.

Disclaimer: my attempt at a transcription of Circassian is hardly a serious one -- call it an experiment if you will -- merely done for purposes of leisure and education. It is not to be construed as advocacy of its imposition in society at the expense of the scripts currently in use. I wouldn't see anyone, especially the Circassians themselves -- who have not much to do with the Chinese culturally, historically, or politically, approving to accept its wider use. And thus, to quote Jan van Steenbergen's justification for Polish Cyrillic, "This project, therefore, is primarily a thought experiment, my answer to the question if such an orthography would be possible at all."

How does this work?

The word that first came into mind when I started with the whole thing is the word адыгабзэ ~ адыгэбзэ* 'Circassian (language)', which may be analysed as a compound consisting of two elements: adıǵa 'Circassian' + bza 'language'. It wasn't so hard to come up with the character 語 for the latter of the two elements. The former, though, proved to be more of a challenge. I wasn't sure if the ethnonym should be broken down into even more minute roots -- Shagirov's etymological dictionary doesn't include an entry on the word; neither can anything on the etymology of the word be made out of Smeets, and therefore I was therefore compelled for the time being, to accept Wikipedia's word on the matter:
The Circassians refer to themselves as Adyghe[37] (also transliterated as Adyga, Adiga, Adige, Adığe, Adyge, Adygei). The name is believed to derive from Atyghe (Adyghe: Iатыгъэ, romanized: 'atığə) meaning "high" to signify a mountaineer
and therefore assigned the character 嵩 (meaning 'lofty' and containing the semantic component 高 which means the same thing, save that the latter is more commonly used in Chinese and Japanese) for adıǵa. Thus for адыгабзэ ~ адыгэбзэ we have the ideographical transcription 嵩語.

The 'okurigana'

Non-sinitic languages which have adopted Chinese characters (Japanese, Korean, Zhuang, Hmong, Bai, Yi, Vietnamese) or made an imitation thereof (Khitan, Jurchen a.k.a. Old Manchurian) would have to, sooner or later, adapt the characters to better suit their language. This, of course, does not immediately happen, at least for the languages belonging to the former group. The written languages of the speakers of the language had not for a long time been their native tongues, but Classical/Literary Chinese -- pronunciation adopted to the phonological constraints of the native language (Vietnamese and Zhuang seems to have retained the pronunciations comparatively faithfully). Then the Koreans came up with a rather cumbersome system (which the Japanese copied). The prologue to the anecdote about the sword and the spear (which I hope to transcribe in Circassian ideograms) would be so annotated if adopted to an Englishman's use:



有 fin
鬻 a, go to b
盾 start

矛 go to a
者 b, go to c

to be read in the order indicated by the annotation: 'Sword-and-shield-selling Snghra'ese bloke exists.'

I had quite some fun creating annotations, so let's try it out on Latin next. The line is from Aeneid; courtesy of UTexas's Linguistic Research Centre. Try to figure out this one for yourselves.

Arma virumque canō, Trōiae qui primus ab ōris
* >> > < << 0 -0-

The system was, obviously, rather cumbersome; it was only intended as a reading aid of sorts. When the need and desire for vernacular literature did arise, all of the Sino-Xenic orthographies came up with the same idea (though at different times): pick up Chinese characters which sounded roughly the same (this, in fact, had been done since much earlier by the Chinese, though sporadically, when they needed to transcribe foreign names and toponyms [e.g. 烏弋山離 for Alexandria]). Some of the characters eventually became so overused that they tended, in Japanese orthography, to simplify at a much faster rate than the rest of the script. Thus was the so-called 'men's hand' and 'women's hand' born. In 1442 King Sejong came up with the idea of creating one for his nation; the scholar-bureaucrats were rather pissed of about it. The rest is history. (pinch of salt: the prevalence of the usage of hangul in the subsequent years is a bit higher than what many cursory, secondary sources would suggest)

The Vietnamese and Zhuang, whose languages are, in terms of morphophonology and syntax, not too far off from Chinese, came up with a different idea: Vietnamese flipped the characters around (to match Vietnamese world order), but then created new characters based on the classic principles of Chinese glyph design (Japanese did so to, but to a much lesser extent). These might've as well worked for Korean and Japanese, but less effectively than a phonetic co-system would, given the polysyllabic nature of both languages (though imho an abugida would suit Japanese better; in fact, I have come up with one, which deserves a thread of its own)

The presence of plenty of prefixes and suffixes in Circassian would thus require such a system. A defective phonetic system -- one not differentiating voicing, aspiration, etc. -- would suit our purposes just fine. The thing is that interdialectal correspondences of consonants can get, due to certain sound changes, rather tedious. Take the plosives, for instance: due to certain sound changes, Temirgoy and Kabardian, two dialects among many which have attained the status of a literary variety, ended up with three-way correspondences of voicing p~p, p~b, b~b; k~k, k~g, g~g, etc. One cannot predict with absolute certainty the voicing of the initial consonant of a Kabardian word corresponding to a Temirgoy word starting in a voiceless consonant, and vice versa. Thus we eliminate voicing from the features indicated by our phonemic sub-system, which is created with one chief guiding principle (amongst others) in mind: Which grapheme a certain phoneme from a certain variety corresponds to must be predictable from that variety alone. In other words, no two cognate phonemes may be mapped to a different grapheme. This principle serves to absolve us of the hassle of checking up the etymological dictionary just to ascertain the qualities of the diaphonemical consonant. Sever the conveniences of modern communicational technology and the Americans and Britons would be at a loss to how the vowel in the phonetic set BATH is pronounced in the other side of the Atlantic!

One can see it as an antithesis to the diaphoneme: while the diaphoneme aims for, in the words of Justin Rye, the 'Highest Common Denominator', our co-system aims for the basest. Making a diaphonemical system, based on Proto-Circassian, is not unfeasible: project each word in every Circassian dialect to their Proto-Circassian source, and design the phonemic 'co-system' accordingly (not unlike how the letter a in bath accounts for two differing qualities in two different prestige varieties of English), but then we'll end up with a huge amount of phonemes (the table in the Wikipedia article on Proto-Circassian counts 66 consonants), more than we would like to deal with in this little, fun experiment of ours.

The phonemes of Proto-Circassian (neatly serving as a dia-phoneme to all current Circassian dialects) are mapped, as follows, to each grapheme of our system, provisionally denoted in Latin capitals.

Code: Select all

P: *b  *p  *pʰ  *pʼ 
T: *d  *t  *tʰ  *tʼ 
Q: *q  *qʰ  
QU: *qʷ  *qʷʰ  
KU: *ɡʷ  *kʷ  *kʷʰ  *kʷʼ 
Ç: *ɡʲ  *kʲ  *kʲʰ  *kʲʼ *d͡ʒ  *t͡ʃ  *t͡ʃʰ  *t͡ʃʼ  *t͡ʂ  *t͡ʂʰ  *t͡ʂʼ *ʒ  *ʃ  *ʃʰ  *ʐ  *ʂ  *ʂʰ  *ʑ  *ɕ   *ɕʼ 
ÇU: *d͡ʑʷ  *t͡ɕʷ  *t͡ɕʰʷ  *ʑʷ  *ɕʷ   *ɕʷʼ 
C: *d͡z  *t͡s  *t͡sʰ  *t͡sʼ 
S: *z  *s   
L: *ɮ  *ɬ  *ɬʼ  
X: *ɣ  *x   
XW: *xʷ   
Ẋ: *ʁ  *χ   
ẊW: *ʁʷ  *χʷ   
Ħ: *ʔ  *ħ   
ĦU: *ʔʷ    
M: *m    
N: *n    
I: *j    
U: *w    
R: *r    
Diphthongs and triphthongs (based on Kuiper's chart):

Code: Select all

		ST	ÇT	
PC				
PS				
PÇ			ÇÇ	
				
PL				
PX	TX		ÇX	
				
	TKU		ÇKU	
	TXU		ÇXU	LXU
				
PẊ	TẊ	SẊ	ÇẊ	
PẊU	TẊU		ÇẊU	
	TĦ		ÇĦ	LĦ
				
PST				
PSKU				
PÇĦ				
STẊ				
STẊU	



The grapheme, as is evident from the chart, contains information solely of the place and manner of articulation. Note that Ç and ÇU are mapped to both affricates and fricatives of diverse places of articulation. Justification provided on demand.

Now the question remains; what sort of form shall this co-system assume? Instead of taking the trouble of creating a script anew, let us co-opt a currently existing one. This script here, created by a Circassian and his son (let's call the script Daur's script in their honour) based on ancient Circassian runes would be an ideal option. It has, however, yet to be included in Unicode. Latin is out of the question for reasons of aesthetics. The choice, obviously, falls onto two scripts whose compatibility with handzi has been proven by hundreds of years of use: hangul and kana, each with their own advantages and shortcomings as we shall see in a while.

Code: Select all

P	비	ふ	Ç	이	し	X	기	は		L	리	り
T	디	た	ÇU	오	しわ	Ẋ	ᅙᅵ	ほ		M	미	ま
Q	키	こ	C	지	つ	Ħ	히	ひ		N	니	な
QU	쾨	こわ	S	시	さ	XU	괴	はわ		I	ᅌᅥ	や
KU	꾀	かわ				ẊU	ᅙᅬ	ほわ		U	오	わ
						ĦU	회	ひわ		R	ᄙᅵ	る
										A	ᅌᅵ	あ
The assignment of hangul jamos onto the script may seem rather erratic, and for the most part do not conform to the pronunciation of the latter in Korean; but I had to what I can in order to fit all of the diphthongs in (more detailed explanation on demand). No such consideration are involved, for obvious reasons, in deciding which kana to allocate to each phonemic set (though at least one point of reason in the design: the o- set reserved for the uvulars); all digraphs and the labialised vowels are written in digraphs. Note that all hangul jamo are always accompanied with a horizontal line to the left, be they followed, or not, with a sound in the dialects (Proto-Circassian had a (C)(C)CV phonosyntax, i.e. its syllables were always open). In Korean this would represent the vowel /i/; however in our system this is a graphical feature used in order to conserve the shape of the consonants which have no bearing upon the pronunciation.

Ejectivisation, uniform across accents, may be optionally indicated with the handakuten (゜) in kana, and in hangul by the use of a device to be determined later.

Vowels:

Traditional analyses of Circassian vowels posit a vertical, three-vocalic system consisting of the vowels *ā (orthographic а), *a (э), and *ə (ы). More recent findings, however, have shown that it is quite possible to eliminate *ā and *ə, when the former is analysed as a sequence *aa and the latter as a mere 'filler' (an overview may be provided on demand). We are thus left with the vowel a, which is optionally denoted in the Hangul-based system with a stroke to the right of the line (ᅡ), and in kana with the dakuten.

*ā is to be written as the sequence [previous vowel + ᅌᅵ/あ], thus being analysed as underlyingly the sequence *aa.

We are at last fully equipped to write. One example before I take a wheeze of fresh air (Cyrillic transcription based on Adyghe):

書리 書り  тхылъ 'book' : etym. root тх ('to write') + лъ (derivative suffix with the base meaning 'to lie')
書리蓋 書り蓋 тхылъыпӀ 'paper' : тхылъ + пӀ ('cover', 'case', 'shell'). The senses covered by the ideograph and the Circassian root congrue amazingly.

디사ᄙᅡ見니 / たさら見な / тызэрэлъэгъун!
Conlangs in progress:
Modern Khotanese
Modern Gandhari
?? - Japonic language in the Mekong Delta
Locna - Indo-European language in N. Syria
Wexford Norse
A British romlang, &c.
dva_arla
cuneiform
cuneiform
Posts: 184
Joined: 25 Oct 2019 21:03
Location: Realm of Ideas

Re: Circassian hanzi: an experiment

Post by dva_arla »

Links to online dictionaries, for reference:

Chemgui -> Russian
Kabard -> Russian
Russian -> Kabard
Conlangs in progress:
Modern Khotanese
Modern Gandhari
?? - Japonic language in the Mekong Delta
Locna - Indo-European language in N. Syria
Wexford Norse
A British romlang, &c.
dva_arla
cuneiform
cuneiform
Posts: 184
Joined: 25 Oct 2019 21:03
Location: Realm of Ideas

Re: Circassian hanzi: an experiment

Post by dva_arla »

In order to better accommodate the consonant clusters, I have decided to give the kana-based system a quick overhaul: the cluster onsets get their own kana!

P- へ T- て S- せ Ç- し L- れ

(cf. their homophonic counterparts, used standalone or as the coda of a cluster)

P は T た S さ Ç ち L り

Contrast:

はつ  蠅 б1адзэ (PACA)
へちつ 蚤 бжыдзэ (PÇECA)
Conlangs in progress:
Modern Khotanese
Modern Gandhari
?? - Japonic language in the Mekong Delta
Locna - Indo-European language in N. Syria
Wexford Norse
A British romlang, &c.
Post Reply