How do we know that Chapter IX of the Tangut dictionary Homophones lists Tangut characters that have z-like initials? Tibetan transcriptions include sibilants:

sl- (only twice)



H-z- (only once)

(g-)zh- (zh- = [ʑ])

rdz- (only twice; cf. the -rʒ- of foreign approximations of Czech ř in Dvořák)

c- [tɕ] (only once)

j- [dʑ] (only once)

(I am assuming that zh-, c-, j- were pronounced conservatively. In any case, those letters imply Tangut consonants other than l- or r-.)

Chinese 移 (as sole transcription and as fanqie initial speller) corresponds to Tib (g-)z-, H-z-. 移 was ?*ji in Tangut period NW Chinese but the diacritic might indicate a j-like consonant with more friction. Oddly 移 doesn't correspond to Tib zh- [ʑ]. Tangut period NW Chinese had no initial *(d)z-, and I'm surprised there are no Chinese diacritic + ?*(t)s- transcriptions corresponding to Tibertan sibilant transcriptions of Chapter IX Tangut characters.

Tangutologists at least as far back as Nevsky have reconstructed *z-like initials to account for the above transcriptions:

Nevsky 1926: z-, ž-

Nevsky 1960 (I: 138): z-, ź-, ńź-

Nishida 1964: ʁz-, ňž- (he also has ʁ- but no z- or ž-; I don't know of any language with Cz-clusters but without simple voiced sibilants)

Sofronov 1968: z̀- (with grave accent!; no z-)

Li Xinkui 1980: z-, ʐ-

Huang Zhenhua 1983: ɽz- (he also has ɽ- but no z-)

Li Fanwen 1986: z-, ʑ-

Gong 1997: z-, ź-

Arakawa 1999: z-

Arakawa (according to Kotaka's now offline "Tangut in Four Weeks"): z-, zz-

(Can anyone verify where Arakawa first reconstructed zz-?)

Tai 2008: ʁz-, ʁź- (but no ʁ-, z-, or ź- which seems unlikely)

So far, I don't know of anyone else who proposed ɮ- or a Czech-like ř-. I thought my ʐ- was also unique but I had forgotten about Li Xinkui's reconstruction. I have never seen his 1980 mimeographed book 西夏语辅音系统研究 (A Study of the Tangut Consonantal System) and am quoting Li Fanwen's (1986: 126-127) table of his reconstructed Tangut consonants.

Next: How did Czech develop the exotic sound ř? TANGUT LIQUIDS: THE DVOŘÁK HYPOTHESIS (PART 1)

Chapter IX of the Tangut dictionary Homophones lists Tangut characters which have initial liquids and voiced fricatives in Gong's reconstruction:

l- lh- r- z- ź-

If Gong is correct, this chapter mixes two types of consonants whereas other chapters only have a single type (e.g., labials in Chapter I). Moreover, why aren't z- and ź- in the same chapters as their affricate counterparts dz- and dź-?

Chapter VI: ts- tsh- dz- s- (but not z-!)

Chapter VII: tś- tśh- dź- ś- (but not ź-!)

The Tangut classification of consonants was modelled after the Late Middle Chinese consonant classification system which groups the fricatives *z- and *ʒ- with their voiceless counterparts:

齒頭音 'tooth head sounds': *ts- *tsh- *dz- *s- *z-

正齒音 'proper tooth sounds': *tʃ- *tʃh- *dʒ *ʃ- *ʒ-

Why would the Tangut move z- and ź- to a different category if they otherwise closely followed the Chinese model? The obvious answer is that the consonants that Gong reconstructed as voiced fricatives were somehow more similar to the liquids l- lh- r- of Chapter IX than to the sibilants of Chapters VI and VII.

In the past, I have proposed that Gong's z- and ź- should be reinterpreted as ɮ- and ʐ-. ɮ- is lateral like l- and lh-, and if lh- was a voiceless fricative [ɬ], then ɮ- would be its voiced counterpart. ʐ- might be like standard Mandarin r- which varies between [ɽ] and [ʐ] depending on speaker. It would contrast with Tangut r- which would be a flap [ɾ] as in Japanese or Korean.

I've been looking at Czech lately and it occurred to me that Gong's ź- and my ʐ- could be reconstructed as the ř of Dvořák. Czech ř is an unusual sound that Janda and Townsend transcribe as [r͡ʒ] and describe as

a unit phoneme, which means that it is simultaneously a dental and a palatal [? - do they mean simultaneously a liquid and a fricative?; Wikipedia considers it to be alveolar]. ř is usually the last sound acquired by Czech children and its mispronunciation constitutes over 50% of their speech defects (Palkov‡ 1994: 347, 350).

Some non-Czech speakers, including other Slavic language speakers, approximate ř as a sequence like [rʒ]: e.g., Dvořák is Дворжак in Russian.

However, ř is unique to Czech as far as I know - I can't find it in the UPSID database which lacks Czech - and I would prefer to reconstruct a less exotic consonant like ɽ or ʐ.

Next: Why should z-like consonants be reconstructed for Chapter IX initials? ANIMATE INANIMATES

I've been using Janda and Townsend's free Czech grammar from the Slavic and East European Language Research Center (SEELRC), a site I've barely begun to explore.

On Saturday, I talked about three plural endings for Czech masculine animate nouns. You might be surprised to learn which nouns are considered 'animate'. Janda and Townsend list several classes of animates on p. 16. Only the first is obvious:

1. "all male (or grammatically masculine) living creatures capable of motion (i.e., non-vegetable)"

2. "In the case of very small creatures, such as bacil 'bacillus', mikrob 'microbe' animacy is open to interpretation."

3. "[F]acultative aniimates, nouns which refer to inanimate objects but display animate endings"

3a. "Animal names used to refer to inanimates": e.g.,

kon’íček 'little horse; hobby'

3b. Names of things that look like human males: e.g.,

sněhulák 'snowman', maňášek 'puppet'

3c. Names of things that are homophonous with human males: e.g.,

talián 'Italian; Italian sausage'

3c'. Things that are similar to things that are homophonous with human males: e.g.,

vuřt 'wurst' (a sausage like talián 'Italian sausage')

3d. Things associated with male behavior:

"drinking of alcohol": pan‡ák 'shot'

"smoking of cigarettes": čmoud 'drag from a cigarette'

"card-playing": ferbl 'name of a card game'

"sports": kraul 'crawl (swimming)'

"mathematics": n‡ásobenec 'multiplicand'

That last one is really macho!

I hope the SEELRC Czech grammatical dictionary "under development" will indicate which inanimates are animate because that information isn't in any dictionary or glossary I have. PROM NIGHT

is not a title you'd expect to see on my blog, but it is the original title of the movie known in Czech as Maturitní ples, the mystery phrase from my last post. Although I keep thinking it sounds like 'maternity place', it actually means 'leaving examination ball' which is not exactly the same thing as an American prom (not that I've ever been to one!). I was shocked to see maturitní ples and its colloquial Czech equivalent maturiťák translated as 'senior prom' in Janda and Townsend's Czech grammar. Although I don't know anything about Czech culture, translations like that one seem too American to be true. MATURITNÍ PLES

is the Czech title of a recent American movie. What does it mean? Answer tomorrow. Its colloquial Czech equivalent is maturiťák. I learned both terms from p. 102 of Janda and Townsend's Czech grammar.

Here's the key to last night's quiz:

English words in Czech:

džinsy < jeans (dž is pronounced like English j)

sendvič < sandwich (Czech has no vowel like the a of sandwich and has no w; Czech č is pronounced like English ch)

šejkr < shaker (for mixing drinks; Czech š is pronounced like English sh)

trénink < training (Czech has no native words ending in -ng)

Other European borrowings in Czech:

autostop 'hitchhiking'

fotoaparát 'camera' (photo-apparatus)

fotograf 'photographer'

inženýr 'engineer' (Czech ž is pronounced like zh)

Native Czech words:

bratr 'brother'

host 'guest' (! - cognate to Russian гость gost' which didn't undergo the Czech g > h shift) ANGLICKÁ SLOVA V ČEŠTINĚ

Robot (which I mentioned in my last post) is the only Czech borrowing into English I can think of*, but Czech has borrowed a lot of anglická slova 'English words'. Can you identify the originals for

džinsy (a double plural; -y is a Czech plural suffix; I'm surprised it's not džínsy with a long vowel)




The last three are taken from p. 17 of Short's Teach Yourself Czech.

These words aren't from English but you still might be able to guess what they mean - or not!


bratr (native to Czech - what's the English cognate?)



host (cognate to English host, but native to Czech; Czech h is from earlier g, which is why modern Praha corresponds to English Prague - English preserves some earlier, now obsolete names: e.g., Peking for Beijing)


Answers next time. Until then, you can find more puzzles to solve at the Slovník cizích slov (Dictionary of Foreign Words).

*There are two more I didn't know about in "From robots to soap operas: 100 years of Czech-English borrowings": pistol and howitzer! ĎÁBEL-ISH DOUBLE DIACRITICS

One of my proposed Czech orthography changes in the last post involved the consistent use of ˇ atop vowels to indicate palatal consonants or consonants followed by [j]. I mentioned that this method wouldn't work with final palatals that have no following vowels to bear ˇ. But there's another problem. Czech uses the acute accent to indicate long vowels: e.g., ďábel is [ɟaabel]. Here's how the word would be respelled using my proposals from the previous post:

1. ďábel (no change)

2. dǎ́bel (with an acute atop ˇ - this might not even display properly in your browser - yuck!)

ǎ́ is reminiscent of Vietnamese which is full of double diacritics e.g., ắ - the breve ˘ (curved unlike the Czech háček = ˇ) indicates vowel quality and the acute indicates the tone.

3a. djábel

3b. dyábel

To avoid diacritic stacking, long vowels could be written doubled:

1'. ďaabel

2'. dǎabel

3a'. djaabel

3b'. dyaabel

Taking the nondiacritic approach of 3b' to the extreme, I could replace other Czech letters with digraphs:

č > cy (so Čapek would become Cyapek)

Čapek popularized the word robot (< Czech robota 'work'; more here) and almost used the term laboři from Latin labor instead of roboti 'robots'

ř > ry (so Dvořák would become Dvoryaak - but ř sounds more like rzy!)

š > sy

ž > zy

Czech has two letters for long u [uu], ú and ů. The latter is from old ó [oo] which shifted to uo and then to [uu] (but written with a little o on top as a reminder of its origin). ů could be written as uo to reflect its older pronunciation or could be written as uu to reflect is current pronunciation.

(Polish underwent a similar change not reflected in spelling: old [oo] became [u] but is still written ó.)

Czech got a new ó [oo] from foreign loanwords. I'll provide an example in my next post. DIÁBEL

In my last post, I asked,

Can you guess what ďábel means? Hint: Think of the ' in ď as a little i.

If that superscript i is enlarged, the result is diabel which is close to French diable and Spanish diablo, both 'devil'. The correspondence

Czech ď : foreign di

is also found in another word I mentioned yesterday:

Kanaďan 'Canadian'

ď roughly represents the sound 'dy' (the IPA voiced palatal stop [ɟ] to be precise). This sound is written as plain d before ě and i. The spelling of the palatals [c ɟ ɲ] (approximately 'ty dy ny') depends on the vowel:

vowel \ consonant c- ɟ- ɲ-
back vowels: ť-, ď-, ň- (not n’-!) a ťa ďa ňa
o ťo ďo ňo
u ťu ďu ňu
front mid vowel: ˇ atop e e
front high vowel: regular t-, d-, n- i ti di ni

[ti di ni] are written as ty dy ny. The i/y distinction refers to the previous consonant, not the vowel:

ti [ci] 'to you' : ty [ti] 'you'

If I could redesign Czech orthography, I'd like palatal consonants to be indicated by a single consistent marker on the consonant: e.g., (bold indicates changes from the current orthography)

c- ɟ- ɲ-
a ťa ďa n’a
o ťo ďo n’o
u ťu ďu n’u
e ťe ďe n’e
i ťi ďi n’i

[ti di ni] would be ti di ni without ’.

(I would prefer ˇ atop t d n but my font automatically turns ˇ into atop t and d.)

: pe, etc. would become p’e : pe, etc.

Alternately, I could mark the palatal consonants with a single consistent mark on the following vowel: e.g.,

c- ɟ- ɲ-

[ti di ni] would be ti di ni without ˇ.

: pe, etc. would remain : pe, etc.

However, this method has a fatal weakenss: it would not have any means to indicate palatals not followed by vowels.

An even simpler solution would introduce digraphs with -j- (or -y-?)

c- ɟ- ɲ-
a tja dja nja
o tjo djo njo
u tju dju nju
e tje dje nje
i tji dji nji

[ti di ni] would be ti di ni without j or y.

: pe, etc. would become pje : pe or pye : pe, etc.

Final palatals would be written as -tj -dj -nj or -ty -dy -ny (since the latter would no longer be needed for [ti di ni]).

ty dy ny would be similar to Hungarian ty gy (not dy!) ny for palatals and would avoid the ambiguity of medial -Cj- clusters which could represent either single consonants [c ɟ ɲ] or [t d n] followed by [j] in compounds. I can't find any current Czech compounds with -t-j- -d-j- -n-j- in Janda and Townsend's Czech grammar, but I can't think of any reason why they wouldn't exist.

Tangut fonts by Mojikyo.org
Tangut radical font by Andrew West
All other content copyright © 2002-2009 Amritavision