Draft of New Latin-based Kazakh Alphabet.

I realize it’s just a draft, and may never become a reality (a point made by cliff arroyo in this recent Log thread: “The switch from cyrillic to latin seems to be one of these issues that shows up every few years and gets some press and then disappears”), but this Kazinform report includes a paragraph that baffles me:

The scientists rejected the idea of introducing diacritical marks (glyphs added to a letter, or basic glyphs) as they suppose that because of rare use, the specific sounds of the Kazakh language can disappear.

Any ideas as to what they might be trying to say?

Comments

  1. Eli Nelson says:

    It sounds like they think that if diacritical marks were used to write certain sounds (e.g. “ĝ” /ɢ/ vs. “g” /g/; this is a totally made-up example because I’m not familar with Kazazh phonology), people would usually leave them off in writing (write both as “g”) and then this orthographic merger could contribute to phonological merger of both as /g/.

  2. Ah, that makes sense. I mean, the fear doesn’t make sense, but the explanation of the paragraph does!

  3. What Eli said, except that in their heads I don’t think they’re bothering to distinguish orthographic mergers from phonological ones.

  4. David Eddyshaw says:

    It explains why English doesn’t distinguish /θ/ and /ð/, and why Russian has predictable stress, after all.

  5. David Eddyshaw says:

    To be more charitable, there may after all be examples where a traditional orthography has played a role in maintaining an obsolescent phonological distinction, by a sort of spelling pronunciation, except that in such a case the result would not be wholly spurious but just artificially preserved.

    I can’t think of any offhand, but I wonder, for example, if this is ever the case with Classical Arabic words borrowed into modern spoken Arabic languages? (Lameen will know…)

    Has it happened in Turkish since the Atatürk reform?

  6. The influence of orthography on phonology can sometimes be very strong.

    Swedish used to have a fricatives /ð/, /ɣ/ and /θ/, which orthographically could be distinguished from the corresponding stops by writing dh, gh, or th instead of d, g, or t.

    Later, phonologically, the fricatives completely disappeared, while orthographically dh and d merged into d, etc. Thus, we had situation where some written ds, gs, and ks were ‘silent’.

    However, under influence of the written language, some of the ‘silent’ consonants have re-emerged in the spoken language, now in the guise of stops!

    Thus, the choice of orthography has partially reversed the phonological process of the fricatives disappearing, instead effectively causing a merge of the fricatives and the stops into stops.

  7. David Eddyshaw says:

    @Avinor:

    A bit like the ‘t’ in ‘often’…

    That’s certainly analogous. I was thinking more of phonemic distinctions that would otherwise have been lost altogether: rather as if there were people out there pronouncing standard English ‘night’ as [nɪçt] because that’s how it’s spelt.

    By the way, the name “Kazinform” sounds splendidly Sovietical.

  8. Suprasegmentals are a different matter, but for what are supposed to be separate phonemes, the functional load of the interdental fricatives in English is extraordinarily low. There is archaic thigh/thy, and ether/either in some accents, and there are a few noun-verb pairs like mouth noun vs. mouth verb (which really should be spelled mouthe, but that’s English for you) and teeth/teethe, and that’s about it. It’s also the case that /ð/ is a dead phoneme: foreign words with [ð] like padre, skordalia are borrowed with /d/, and consequently the two never fully separated, unlike /s~z/, /f~v/. Indeed, the distribution is almost but not quite rule-governed still, provided you treat native words and loanwords separately.

    Scots do say /nɪxt/ sometimes:

    If you can say,
    “It’s a braw bricht moonlicht nicht”,
    Ye’re aw richt.
    That’s aw (or: Ye ken).

  9. On the other hand, I’ve seen complaints about Kazakhs and other Turkic speakers pronouncing Turkic words with soft consonants before front vowels under the influence of Russian reinforced by Cyrillic orthography.

    A more realistic worry about diacritic marks being left off due to technical limitations or carelessness than the loss of the distinction in the language itself is that it would cause a lot of uncertainty about the correct original spelling, as you would be familiar if you ever had to figure out if a certain Gunther from Germany was really Gunther or Günther.

    This is a genuine worry, and I understand the reason they insisted on not using diacritics, just as in South Korea’s official romanization of Korean introduced in 2000. Even now, too many systems make it difficult to use more than ASCII letters (e-mail addresses, anyone?). But I really wish that new Latin-based orthographies would insist on using diacritics to put more pressure on the technology to catch up to a world that is not limited to the ASCII standard.

    By the way, they seem to be collapsing Cyrillic х /χ/ and һ /h/ into Latin h. Neither sound is native to Kazakh so the functional load of this distinction is probably small, but the choice of h for Cyrillic х /χ/ leads to some ambiguous instances of sh and zh, which could represent ш and ж or сх and зх. Would it have been too difficult to use Latin x for Cyrillic х, as it seems to be left unused?

  10. David Marjanović says:

    Latin x is in fact used for Cyrillic x in Azerbaijan.

    Rumor has it that young people have abandoned the k/q and g/ʁ distinction of Tatar because – being predictable from vowel harmony – it isn’t made in Tatar Cyrillic. The language has enough Arabic, Persian and now Russian loanwords, however, that the distinction has become phonemic: /mæqælæ/ “(Wikipedia) article” has an unpredictable /q/, and the Cyrillic spelling can only hint at it by lying about the vowels (мәкалә).

    The Uzbek Latin alphabet, imposed by the dictator very soon after independence, is strictly limited to ASCII and uses lots of apostrophes instead of diacritics. I find it a trainwreck. At least it’s not as inconsistent as the Turkmen one, though…

  11. @David E, could I unpack your It explains why English doesn’t distinguish /θ/ and /ð/,

    You mean both the voiced and unvoiced dental fricatives are spelled “th”(?) Old English spelling has both Thorn and Eth, but it seems they didn’t carry a phonological distinction anyway(?)

    Was there ever any spelling system for English that distinguished voiced vs unvoiced? And yet there continues to be a distinction. There’s a minimal pair (at least in my speech): thistle vs this’ll.

  12. David Eddyshaw says:

    @AntC:

    It was as true as the following clause (and for the same reason.)

  13. AntC: The obvious device is to use dh, and it has been proposed in various spelling reforms that have not caught on. At least one reformer uses d: unsurprisingly, he is an elderly man from NYC.

    David E: The U.S. Navy is big on Soviet-style stump compounds. Various bureaus are called BuMed, BuPers, BuSandA (“supplies and accounts), and BuWeps. Similarly, the Office of Naval Operations is OpNav, the Navy Military Personnel Command is NavMilPersCom, and people whose official title was “Assistant Communicator” were once known as AssComs. (Properly these are written in all caps, but I have used CamelCase here for clarity.) Ordinary acronyms are also commonly used, of course, like OOD ‘officer of the day’, POW ‘plan of the week, prisoner of war’, SCBA ‘self-contained breathing apparatus’ (an earlier version with underwater gave us the word scuba).

  14. > It explains why English doesn’t distinguish /θ/ and /ð/

    /ð/ is mostly used for a few high frequency lexemes, so people are unlikely to get confused. Also, I think one could ask the question of whether, in close-to-phonemic orthographies, people are more likely to adopt mergers or spelling pronunciations in the few cases where the orthography is not phonemic, due to the expectation that the correct pronunciation matches the orthography.

    The proposal mostly looks like a transliteration. To what extent is the current orthography phonemic? Any missed opportunities for clean-up?

  15. I pity US immigration officers who will have to deal now with names like Inzhw Zhwmagwlova

  16. Some Russian Latin-based transliterations use x for х and w for ш. And I dislike them.

  17. WP has a nice list of English spelling pronunciations: often, forehead, clothes, salmon and friends, comptroller, waistcoat, conduit (U.S.), victuals, medicine, Bartholemew, Anthony (U.S.), Everest (the mountain; the explorer had the FLEECE vowel), Arctic, hotel, Ralph. In French, we have oignon (standardly ognon), Montaigne (in which the i was traditionally silent), club (originally as if clab), and shampooing (traditionally champouïgne but now often champouin).

  18. Greg Pandatshang says:

    According to Wikipedia, Cyrillic у is a consonant in Kazakh, representing /w/. Therefore, a name like Жумагулова should not be possible in Kazakh (I’m making some assumptions about phonotactics here). I’m aware that such a name does exist, and I’ll assume it’s a Kazakh person’s name. Is it possible that this is partly Russified, i.e. that the proper Kazakh spelling would be different? Even still, there would certainly be some opportunities for confusion. Or maybe Wikipedia is wrong about this.

  19. Greg Pandatshang says:

    P.S. A little more research reveals that Bakhytzhan Zhumagulov’s name is spelled Бақытжан Жұмағұлов on the Kazakh Wikipedia. So, per the new proposed standard, it would be not Zhwmagwlov but Zhumaghulov.

  20. Oh, so they propose to use w for “у”, not for “ұ”.

    It’s much more sensible, but still would result in monstrosities like Ajsha Abdwllina

  21. the phrase “great Kazakh writer M.Auezov” will be written in new script as “uly qazaq zhazwshy M.Aewezov”

    not quite the level of Johanna Nichols’ excellent Ingush grammar, but getting close

  22. An interesting comment here on some of the monstrosities that this proposal with create:
    https://www.facebook.com/drobyazkoe/posts/1468544639900184

    I can’t understand why anyone would consider using anything else than a modified Turkish alphabet for other Turkic languages.

  23. Cyrillic у in Kazakh stands both for /w/ and the tense vowel [u] from combinations like /ʊw/, /ʉw/, /əw/, and /ɪw/ (as well as Russian borrowings). So it’s like й /j/ and и [i] /əj/, /ɪj/ rolled into one. One wonders whether they could have cribbed ў from Belarusian to stand for the /w/.

    As I understand it, the proposed Latin alphabet keeps the asymmetry by letting w do double duty for /w/ and [u] while dividing the labour between j /j/ and i [i]. There is the precedent of Welsh, I guess (though in that language i also does double duty as consonant and vowel), not to mention the Perso-Arabic script for Kazakh. An existing system for writing Kazakh in the Latin alphabet that does use diacritics (as shown in the Wikipedia article on Kazakh alphabets) also seems to use w for both the consonant and the vowel.

    The correspondence chart not included in the linked Kazinform article shows that ә, ө, and ү will be written ae, oe, and ue respectively in this proposal. It is telling that even here they didn’t go for ä, ö and ü, three of the most widely supported non-ASCII letters, for these rather frequent sounds in Kazakh.

  24. Greg Pandatshang says:

    I can see the appeal of going full ASCII, and, as a full ASCII proposal, this doesn’t seem so bad. Clearly, as long as they’re going with digraphs, they should use “kh” rather than simply “h” for /x/, which will resolve some of the ambiguities of their other “_h” digraphs. I have no idea what’s going on with Cyrillic “у”. Is “zhazwshy” really pronounced /ʒazwʃə/, or is that the tense vowel? Jongseong Park’s comment appears to say that Cyrllic “у” only represents a vowel as the second element in a diphthong (which makes it orthographically comparable to vowel “w” in English).

  25. The original Cyrillic alphabet is already kinda difficult. Centralized etymological i is written <i>, while what is phonetically i, and phonologically i+yod is written <и>. Anyone moderately familiar with Ukrainian would propose exactly the opposite solution.

  26. marie-lucie says:

    JC: In France I have always heard and said (but never seen written) “champoin”, a word of two syllables in which the written sequence “pooing” recalls the word “poing” ‘fist’ (pronounced like “point” ‘dot’), where “oin” is a nasal diphthong. I have never run into “cham-pou-ïgne”, a clumsy attempt to imitate English pronunciation, rather than anything “traditional” (since the product has long been known, from a time when knowledge of English among the general population was quite limited). In a hair salon the person (usually an apprentice) who shampoos a client’s hair before the hairdresser cuts and styles it is (or perhaps was) called une champouineuse, so there must be a verb champouiner which supposes a nativized French base champouin (pronounced identically with my “champoin”).

    (After writing this I turned to the TLFI which writes hybrids like “shampoing”, “shampooineuse” and others but does not report anything like “champouïgne”).

  27. Alex M.

    Thanks for the link. anggharw is truly horrible.

    Looks like new Kazakh Latin wants to beat Welsh at unpronouncibility of spelling

  28. -I can’t understand why anyone would consider using anything else than a modified Turkish alphabet for other Turkic languages.

    I kind of like Azeri Latin. At least it got some style and unique feel. The rest are pretty ugly.

    So far, the Turkmen Latin alphabet was the ugliest, but new Kazakh Latin is even worse

  29. Montaigne (in which the i was traditionally silent)

    So it was pronounced as if Montagne?

  30. m-l: I didn’t mean to suggest that the written forms champouïgne, champouin represented actual French spellings; they are simply my respellings of the IPA forms, respectively /ʃɑ̃puiɳ/ and /ʃɑ̃pwɛ̃/. The presence of the English ending -ing, though not actually used in English in this word, certainly suggests a borrowing from an English base something like /ʃampuiŋ/.

  31. @Greg Pandatshang: Cyrillic у stands not only for [w] as in ауа [ɑwɑ] but the vowel [u] as in алу [ɑlu], which is phonemically a vowel-glide combination such as /ʊw/ or /əw/.

    If they’re going to go full ASCII and use digraphs for high-frequency sounds anyway, then I agree that they might as well use kh for х /χ/ and leave h for һ /h/, which preserves the distinction between the two as well as resolving most digraph-related ambiguities.

    @SFReader: The similarities with Welsh extend to y representing [ə].

    The Azerbaijani and Crimean Tatar Latin alphabets are basically extensions to the Turkish alphabet, and it would have better for the others to have gone the same direction. But that’s what you get when you have script reforms overseen by governments that don’t coordinate with each other.

    @Languagehat: Warnant’s Dictionnaire de la prononciation française gives the pronunciation of Montaigne as ‘mɔ̃-tɛ(-)ɲ(ə), pfs -ta(-)ɲ(ə)’.

    The name seems to be in fact an older spelling of montagne. The pronunciation [mɔ̃tɛɲ] is therefore a spelling pronunciation that has nevertheless become the usual one today. The ign in older spelling usually became gn in today’s French, though it remains in oignon (at least until the new official spelling ognon takes over).

  32. The name seems to be in fact an older spelling of montagne

    Which accounts for the English spelling mountain (and similarly fountain, captain, bargain), which look as if they should rhyme with detain, retain, sustain rather than with Downton, Sahaptin, jargon (at least for those of us with the Weak Vowel Merger). If I had my way I’d chainge [sic] all unstressed -ain to -en.

    A bizarre sort of spelling pronunciation is represented by Dutch oubollig ‘old-fashioned, dated’, a native word which is quite phonetically pronounced /auˈbɔləx/, but is often pronounced /u-/ under the impression that it is one of the many French loanwords in which the spelling ou for /u/ is preserved.

    m-l: The English word shampoo in its modern sense ‘hair cleaning liquid’ doesn’t show up in the OED until 1897; before that it is shampooing liquid or something similar. (To make things worse, the original meaning of shampoo (v.) in English was ‘massage (the scalp or otherwise)’, and it doesn’t get its modern meaning ‘wash the hair’ until the 1850s. Per the TLFi, shampooing in the modern sense (rather than as a nomen agentis, a sense also available for English shampoo) appears in 1883. So apparently the two words in their modern senses are independent developments, or at most related by semantic transfer.

  33. Ah, that makes sense. I mean, the fear doesn’t make sense, but the explanation of the paragraph does!
    I tried to see whether the Russian Version is clearer or more linguistically sophisticated, but it isn’t.
    In any case, I can understand the fear somewhat – a major share of Kazakhs and most non-Kazakhs in Kazakhstan don’t speak Kazakh natively and learn it only in school, based on writing, and it’s often taught and learnt badly. So there is a danger that learners might ignore features that look optional and that teachers might not bother sufficiently to correct that; with digraphs at least you don’t have any excuse to leave anything out.

  34. That makes sense too.

  35. marie-lucie says:

    JC: shampooing

    Of course the French usage started from English, but my point is that it must have started from the written not the oral form. Seeing shampooing liquid on a bottle of the stuff, a French speaker ignorant of English would interpret this as a noun + adjective phrase (liquid being the obvious equivalent to French “liquide”), and from there go on to use shampooing as a noun. Such a person might hesitate between French “s” and “ch” for the initial sound, but will most likely assume that “pooing” should be pronounced like French “poing” ‘fist’, an unanalyzable word. The fact that the English word includes the suffix -ing would not occur to a person ignorant of the language.

    It is possible that since (some) knowledge of English is now widespread, that some people are using a pronunciation closer to the English one, but I have never heard it (I don’t spend much time in France, but I do get my hair done when I am there!).

  36. marie-lucie says:

    a major share of Kazakhs and most non-Kazakhs in Kazakhstan don’t speak Kazakh natively and learn it only in school, based on writing, and it’s often taught and learnt badly. So there is a danger that learners might ignore features that look optional and that teachers might not bother sufficiently to correct that

    This is a problem with some indigenous languages of Canada (at least). In one case I am familiar with, about 100 years ago a missionary had devised a spelling system which included several diacritics. More than half a century later, some old people still used it occasionally but rather haphazardly (there was not much use of writing in the communities, but a hymn book was produced so that the hymns – original or translations – should not be forgotten). A linguist later visited the area and helped an old native clergyman produce some texts in a more linguistically accurate orthography, which was declared official and later used in school programs. This new one also used diacritics, but not the same ones, or not with the same meanings. As a result there is much confusion about pronunciation among younger people learning the language as L2, especially about differentiating velars and uvulars, the latter using underlining, which is not obvious to do on Facebook, for instance.

  37. Bob Hoberman says:

    David Eddyshaw wrote, “I wonder, for example, if this is ever the case with Classical Arabic words borrowed into modern ‎spoken Arabic languages?‎” Yes, there is exactly such a case. Arabic has two letters, ظ and ض, that long ago represented two phonemes, probably /ðˤ/ and /ɮˤ/ respectively. These have fallen together in all known modern Arabic vernaculars. The reflexes in each dialect parallel the reflexes of /θ/ and /ð/: dialects that still have interdentals also have /ðˤ/, dialects that shifted interdentals to stops have /dˤ/, etc. But many dialects* make a distinction in Modern Standard Arabic between ظ and ض, as /ðˤ/ or /zˤ/ versus /dˤ/, and the same goes for words borrowed from MSA into vernacular speech, or words with a classical flavor.

    (*Well, “dialect” isn’t the right word because we’re talking about MSA, but in each dialect community there is a pretty uniform way to pronounce MSA. I don’t know a better word.)

  38. David Marjanović says:

    in close-to-phonemic orthographies, people are more likely to adopt mergers or spelling pronunciations in the few cases where the orthography is not phonemic, due to the expectation that the correct pronunciation matches the orthography

    Absolutely.

    Which makes me wonder how long the phonemic vowel length distinction of Turkmen is going to last. It was ignored in Cyrillic, and it’s ignored in Türkmenbaşy’s Latin.

  39. I kind of like Azeri Latin. At least it got some style and unique feel. The rest are pretty ugly.

    Me too. A frustrating thing is that It would fit Persian quite well, but no-one Romanises Persian in that way, we’re stuck with finglish and no distinction between /æ/ vs. /ɑ/, collapse of orthographic ق vs غ, <kh> for /x/ and so on.

  40. David Marjanović says:

    Azeri sounds a lot like Persian, too. And like Hungarian. Much less like Turkish than I expected.

  41. ‘The dialect of Turkish used in Azerbijan is not very unlike that spoken at Constantinople; but in the latter city the pronunciation has been so refined, polished and effeminated, as one may say—while in Persia the original harsh, vigorous accent has been preserved—that the two nations are scarcely comprehensible to each other.’

    I entertained myself now and then in Istanbul by making some local friends acquainted with the above, from Lady Mary Sheils in 1856. Even within Turkey I understand the deep Anatolian accent sounds more manly than does that of Istanbul.

  42. I am native speaker of Kyrgyz and honestly I am quite satisfied by our current Cyrillic alphabet. I guess at some point our government could decide to switch to Latin alphabet, but I don’t see any reasons to do it except political ones. I actually experimented with latin alphabet for Kyrgyz(I am not a linguist btw), so my impression it can be done, but just from the aesthetic point of view I don’t like Turkish-like Latin. I would prefer Czech-like variant. However given that most people in Kyrgyzstan won’t like diacritics, realistically their number should be minimal. So after some experimentation I found some possible solution with 28 letters(25 out of 26 ASCII letters and 3 diacritics), but it requires some “creative” usage of letters(like q for ч or x for ш).

  43. Qyngyz Kan wouldn’t approve

  44. Which makes me wonder how long the phonemic vowel length distinction of Turkmen is going to last. It was ignored in Cyrillic, and it’s ignored in Türkmenbaşy’s Latin.

    It’s even worse in (Cyrillic-based) Chechen. Vowel length is not indicated either, and е, о, оь can stand, respectively, for /e/, /e:/, /ie/, /ie:/; /o/, /o:/, /uo/, /uo:/; /ö/, /ö:/, /yö/, /yö:/.

    One variant of Latin-based script (in Russian):

    Noxčíŋ Skandinâvicà
    https://nxskandinavica.blogspot.com/2016/10/blog-post.html

  45. David Marjanović says:

    Oh, looks like there’s a /n/-/ŋ/ distinction that’s ignored, too: нохчийн мотт comes out as noxčíŋ muott. Wikipedia cites a Turkish-based spelling, noxçiyn mott, that ignores it, too.

  46. Nicky: Thanks, it’s great to get the point of view of a native speaker!

  47. David Marjanović says:

    /n/-/ŋ/ distinction

    Nope, that’s nasalization of the preceding vowel.

    I just finished reading the whole thing (to the extent that my Russian vocabulary lets me). The idea is very interesting: it’s tailored to the Swedish/Finnish-with-Saami keyboard layout, so every letter can be typed with recourse to nothing more special than the AltGr key.

    Also interesting: the use of h to mark the ejectives, as opposed to the unmarked aspirates. I’d have done it the other way around; are the aspirates more common than the ejectives?

  48. are the aspirates more common than the ejectives?

    All voiceless plosives and affricates are strongly aspirated and, I believe, occur more often than ejectives. So the choice makes sense.

  49. Greg Pandatshang says:

    Speaking of spelling pronunciations infecting the spoken language, I seem to recall reading somewhere that initial /h/ became unconditionally silent in all major dialects of English in the last few hundred years, but then was restored because of spelling pronunciations. Can’t remember what the source for that was. Is there any truth to this rumor or is it just scurrilous gossip?

  50. Eli Nelson says:

    I have also heard this, but I wonder a bit about the scope of the change (was it really 100% at one point, or was it just more common than in modern speech, with [h] still possible in emphatic positions?). I wonder if Piotr is following this thread and could explain.

    From Old English to Standard English,” by Dennis Freeborn, quotes D.G. Scragg (“A History of English Spelling”, 1974) as saying that

    All the regional dialects of Southern English lost the phoneme /h/ during the Middle English period.

    I have the impression that aspirated plosives usually imply the presence of [h], although I don’t know if this only holds when aspiration is contrastive. Also, restoration of a completely lost phoneme due to spelling-pronunciation seems a bit implausible to me, and it didn’t happen in any of the Romance languages except Romanian, which I understand has /h/ in Slavic loanwords as well.

  51. Eli Nelson says:

    I guess since that quote specifies “Southern English”, it seems possible that some English speakers would have had some familarity with hearing a consonant in h-words from Scots.

  52. Couldn’t “Southern English” just mean the southern half of England, though? On the other hand, I don’t have the impression that /h/ dropping is any less common in the North. Wikipedia has a map of /h/ use in England (“based on Upton and Widdowson, 2006”) that shows it being preserved only in Norfolk, Suffolk and Essex in the east; in Durham, Northumberland and part of Cumberland in the far north; and in Somerset and parts of Dorset, Wiltshire and Hampshire in the southwest. I guess those survivals (especially the one in the east) might have been enough, combined with spelling pronunciation, to re-establish /h/ in prestige speech. They would seem well placed to explain the retention of /h/ in the US, although it is strange that there’s not a trace of British-style dropping to be found here (hu-, wh- and reduced forms notwithstanding).

    Funnily enough, I didn’t know that /h/ dropping was so long established until I saw some videos about the reconstructed Shakespearean pronunciation used at the Globe, and was surprised to hear them talking about two ’ouse’olds in Verona.

  53. The evidence for universal /h/-dropping in Southern English varieties is pretty strong:

    1) All the words with strong and weak forms (pronouns especially) are h-less in all varieties when weak: (h)e, (h)is, (h)im, (h)er, (h)ave, (h)ad. Hit, the neuter singular pronoun, has become it in both strong and weak positions in all but a very few varieties.

    2) Originally [h~x~ç] were allophones of /h/, and when the last two (used in coda position) went, the first [used in onset position] was not far behind. In addition, the OE clusters /hr/, /hl/, /hn/ became /r/, /l/, /n/ at about the same time [x~ç] was lost or even a little earlier.

    3) The only surviving instances of /h/ were in the clusters /hj/, /hw/; the former was lost in some accents, the latter in almost all.

    4) So it was spelling pronunciation that restored /h/, and inconsistently so, as with the pronunciation of herb, which is one of the very few /h/-less words in American English (which is “una lingua tedesca in bocca irlandese”) but has restored /h/ in other Englishes.

  54. David Marjanović says:

    The story goes that Sir Walter Scott made fun of “the eclipsed manner of the Queen’s English”.

    Romanian, which I understand has /h/ in Slavic loanwords as well

    That’s [x]. And has Romanian really put it into inherited Romance words? Google Translate says “I have” is just am

  55. Eli Nelson says:

    I believe it is used in learned loanwords from Latin.

    Wikipedia, and this Wordreference thread, says that [h] and [x] are allophones in Romanian, similar to the situation in Hungarian: [x] is used in pausa or before other consonants, but [h] is used before a vowel. https://forum.wordreference.com/threads/does-romanian-have-the-consonant-h.2563692/

  56. More on Chechen orthography and phonology:

    A proposal for a Chechen orthography
    by Erwin R. Komen
    http://erwinkomen.ruhosting.nl/doc/2007_CheOrthography-Komen.pdf

    Chechen stress and vowel deletion: An optimality theory approach
    by Erwin R. Komen
    http://roa.rutgers.edu/files/923-0807/923-KOMEN-0-0.PDF

    Chechen intonation
    by Erwin R. Komen
    https://www.eva.mpg.de/lingua/conference/2011_caucasian_languages/pdf/Komen.pdf

    Ingush alphabet (in Russian):
    http://serdalo.ru/2597-alfavit-ingushskogo-jazyka.html

  57. -A proposal for a Chechen orthography

    Interesting. At least it’s readable, though one needs to get used to it.

    Especially w for I (glottal stop). Who else does that?

  58. A proposal for a Chechen orthography

    That’s a very convenient little introduction to Chechen phonology and orthography, and it leaves me wondering how anyone ever manages to learn Chechen (other than as a native language).

  59. It’s convenient that the Latin alphabet has both k and q to represent the difference between velar and uvular stops, but one disadvantage of the Latin orthography that Komen doesn’t mention is that unassimilated Russian loanwords would have to be left in Cyrillic, which creates a terrible appearance on the page.

  60. Why would they necessarily?

  61. how anyone ever manages to learn Chechen

    Nothing hard in particular: Georgian/Kurmanji consonants + Norwegian vowels. Initial geminates should be hard for those not born speaking nnapulitano but they don’t seem to be terribly common.

  62. I’m not talking about the phonology, I’m talking about the way it’s written. Ancient Egyptian may have been an easy language to speak, but hell if I can learn it from its written representation.

  63. Especially w for I (glottal stop). Who else does that?

    If I’m interpreting correctly, that’s similar to eye-dialect “miwk” for the way old school Londoners pronounce “milk.”

  64. Hat: If you mean why would it look awful, I can’t exactly say; one problem, however, would be that you’d double-take every time you hit an unassimilated loanword and had to switch the interpretation of the shared Latin/Cyrillic characters. Serbian does Latin or Cyrillic, but never both in the same text. Otherwise, there would have to be a diacritic-free transliteration of Russian unique to Chechen (assimilated loanwords would be written in the Latin Chechen orthography).

    In addition, script shifts create instant illiteracy. Better to sensibly reform the Cyrillic orthography on the lines Komen indicates, and live with the k/q problem. Alternatively, use one of the existing Cyrillic letters for /q/, such as Ӄ/ӄ (multilingual), Қ/қ (multilingual), Ҡ/ҡ (Bashkir), or the adopted Ԛ/ԛ (Kurdish). The first is the most distinctive, but is used only in Siberian languages; the second is more widely used.

  65. the way old school Londoners pronounce “milk.”

    All Londoners, you mean. L-vocalization has been spreading all through England and even beyond.

  66. Hat: If you mean why would it look awful, I can’t exactly say

    No, if I’d meant that, I’d have said “Why would it necessarily?” I meant “Why would [unassimilated Russian loanwords] necessarily [have to be left in Cyrillic]?” And I’m still wondering.

  67. Because that’s how you write Russian, at least in Russia. Another point that I didn’t realize before: the Latin script is associated with separatism.

  68. @JC: Like Hat, I don’t really understand why that would be necessary. Is there precedent for that happening in any of the historical Latin-script languages of the USSR? (Or in… Karelian, the only one which Russia seems to have currently?)

  69. Because that’s how you write Russian, at least in Russia.

    But you wouldn’t be writing Russian, you’d be writing Chechen. We don’t write самовар in English, we write samovar.

  70. What English does has nothing to do with what a minority language does. English assimilates loanwords almost instantly.

  71. Regardless, it’s still not clear why a Russian loanword would have to be written in Cyrillic. It seems unlikely.

  72. Greg Pandatshang says:

    Ancient Egyptian may have been an easy language to speak, but hell if I can learn it from its written representation.

    By the way, I’ve been reading Loprieno’s Ancient Egyptian: a linguistic introduction lately (h/t Conlangery Podcast). It’s a lot of fun and I highly recommend it, but it does induce a bit of a headache in me to look at the example sentences written in transliteration, i.e. consonants only, no vowels, which is all you get for almost all of them. I mentally transmute them into the old-school legible Egyptological form, viz with “e” inserted wherever necessary. I sort of wish Loprieno had just gone ahead and done that himself. It’s not the way Ancient Egyptians would have pronounced the words, but, that said, a series of consonants with no vowels is also not the way they would have pronounced them.

  73. ‘All Londoners, you mean. L-vocalization has been spreading all through England and even beyond.’

    As an Irishman, with lots of English TV and literal English people to be had around me growing up, I understood the south-of-England accent all my life, but it’s only a few foreign languages and a good understanding of the IPA later that I’m happy to mentally place its phonemes. (The issue was the usual one that a lot of the semi-naïve writing on the subject from natives doesn’t quite get how their accent differs from everyone else’s frame of reference.)

    But now, yes, post-vocalic /l/ to [w] is standard there, and if they were as pleasantly Brazilian as they are in this when it comes to their un-conciliatory, un-Brazilian approach to negotiating with the EU, well, the next couple of years wouldn’t have the entertainment value (to put it optimistically) that they promise to have.

  74. David Marjanović says:

    Word-initial long consonants aren’t quite that rare; Swiss German has “geminates all over the word” which are articulated as such even when they can’t be heard (because they’re all voiceless plosives).

    Wikipedia, and this Wordreference thread, says that [h] and [x] are allophones in Romanian, similar to the situation in Hungarian: [x] is used in pausa or before other consonants, but [h] is used before a vowel.

    My take from that thread (after subtracting the annoying poster who didn’t understand, and didn’t try to understand, what people were talking about) is that there’s geographic-or-whatever variation. The sound file it links to has a completely unambiguous, very back-velar [x], but in the last post someone reports having heard a native speaker say the same word with [h].

  75. are articulated as such even when they can’t be heard

    Link doesn’t work; tell me the URL and I’ll fix it.

  76. David Marjanović says:

    Oh. No idea what I did there. The link is here.

    I once read that the Russian Federation have a law that all the languages native to its territory have to be written in Cyrillic. Does anyone know more about that?

  77. January First-of-May says:

    I once read that the Russian Federation have a law that all the languages native to its territory have to be written in Cyrillic. Does anyone know more about that?

    Never heard of that, and, as correctly noted by Lazar, Karelian is still not written in Cyrillic (nor have I heard of any plans to make it so).

    It could have been a brief old proposal, however.

  78. The Law on Languages of the Peoples of the Russian Federation says that the state language of the Russian Federation and state languages of the republics of the Russian Federation have to be based on Cyrillic alphabet. State language of the Russian Federation and state languages of the republics of the Russian Federation can have a different basis if it’s established by the federal law.

    So the law applies only to languages which claim “state language” status within their republic (Tatar, for example).

    Karelian always had Latin alphabet and the law doesn’t apply to them. Because state language of Karelia is Russian (Karelian has official language status, which is different thing)

  79. In practice, implementation of the law means that Cyrillic is used for state language purposes (for writing signs on state buildings or publishing laws), while Latin script can be used for any other purposes. That’s the situation with Crimean Tatar at present.

  80. But…why?

  81. Fears of separatism, I assume.

  82. David Marjanović says:

    A bit like Kublai Khan ordering the Tibetan monk ‘Phags-pa to create one script for all the (major) languages of the empire.

  83. IIRC, some Goidelic dialects still have initial ll/nn/rr alternating with l/n/r as base vs. lenited, but the geminates aren’t written as such.

  84. Greg Pandatshang says:

    Is it really phonetically [lː], [nː], [rː] vs. [l], [n], [r]? As a notable Goidel, to wit the Tom Cruise character from Far And Away, famously observed, I’ve often wondered about that.

  85. Some Chechen/Ingush-related books can be found here:
    http://zhaina.com/language/

  86. @Greg Pandatshang: I think devoicing is involved. Whether it’s replaced actual gemination or only supplements it, I don’t know. I’m referring to East Goidelic (Scottish and Manx). Any comments by someone that actually knows?

  87. They changed it again.
    https://docs.google.com/viewerng/viewer?url=http://www.akorda.kz/upload/media/files/d9bc81021d59a7eaa9835fdae9069532.docx

    Example of the alphabet approved today:

    Y’i’ki’pedi’i’ag’a qos’ keldin’iz (Welcome to Kazakh Wikipedia)

    And now the new Kazakh Latin alphabet is the ugliest Latin alphabet in existence.

  88. Squiffy-Marie van 't Blad, Dutchman-at-large says:

    And now the new Kazakh Latin alphabet is the ugliest Latin alphabet in existence.

    It may be ugly but at least it will be awful to type! (At least in a world full of “smart quotes”.)

  89. Probably also the only one that doesn’t employ plain c but has c’ with a perversely ugly diacritic.

  90. Nursultan Nazarbayev’s ingenious idea was that if you use one special character to perform all kinds of diacritic duties, everything will be easy to type on a computer. But I suppose Nazarbayev’s personal experience as a typist is limited. Someone else always types for him.

    Ja govoril, čtoby ne bylo krjučkov i lišnix toček. Čtoby srazu v komp’juter možno bylo postavit’.

    Excuse my Romanised Russian. I’m not taking any risks.

  91. Only using ‘c’ with a diacritic and not having it on its own does seem a bizarre idea – perhaps they had a Maltese adviser?

  92. David Marjanović says:

    Nursultan Nazarbayev’s ingenious idea

    And Islam Karimov’s before him.

    perhaps they had a Maltese adviser?

    At least proper names with c are common in Malta. For instance, the government is headed by one Joseph Muscat who has been in plenty of EU-related news lately.

  93. What do you think about the new decree?

    http://online.zakon.kz/Document/?doc_id=33613600#pos=0;0

    “Y’i’ki’pedi’i’ag’a qos’ keldin’iz (Welcome to Kazakh Wikipedia)”

    Oh my goodness!

  94. David Marjanović says:

    What do you think about the new decree?

    It looks as if very little thought went into it.

    Oh my goodness!

    Actually worse than O’zbek:

    Har kim yoza oladigan
    ochiq ensiklopediya
    Vikipediyaga xush kelibsiz!

    30-oktabr 2017, dushanba
    Oʻzbekcha Vikipediyadagi maqolalar soni: 129 088

    Maqola yarating

  95. It makes Kazakh look like the “Random Apostrophe” alien / fantasy languages that plague SF and fantasy novels…

  96. More Kazakh Latin horror stories:

    I`y`li`i`a men I`u`ri`i` ku`i`ey` men a`i`el attandy.
    “Julia and Yuri became husband and wife”

  97. January First-of-May says:

    It makes Kazakh look like the “Random Apostrophe” alien / fantasy languages that plague SF and fantasy novels…

    I’m totally sniping Y’i’ki’pedi’i’ag’a if I end up writing a SF novel anytime soon 🙂

  98. “Humans who tried to pronounce it choked and died in agony…”

  99. 🙂

  100. Greg Pandatshang says:

    Any interest from Language Hat regulars in pooling our resources to field a candidate for president of Q’a’zaq`st`an in the next elections? With a platform based purely on orthographic concerns, of course. Nazarbayev can have the presidency back for all I care once we have our K’a3ak’ romanisation reform securely in place.

  101. David Marjanović says:

    in the next elections?

    Elections?

  102. Oh, they have elections in Kazakhstan. It’s just the kind of election that the president cannot lose.

  103. Greg Pandatshang says:

    Sure, Nurs Naz has been in power since 1991. But I assume that’s just because he hasn’t had to face a candidate with better Roman orthographic ideas yet—just imagine the debates!—at least not one with all the political connections and influence of the LH commentariat behind him.

  104. https://tengrinews.kz/kazakhstan_news/apostrofyi-mogut-izmenitsya-budet-proishodit-latinitsey-330424/

    From what I’ve read in this article; it’s about how the use apostrophes, according to the Director of the State Language Development Fund, could be changed after a period of approbation. So it looks like there is some chance that the proposal could be changed. Hopefully they’ll choose the digraphs choice.

  105. Not to mention the fact that in Facebook, I see lots of hysteria and outrage over the alphabet choice from the Kazakhs. They all seem to be also against the use apostrophes and prefer a Turkish-used variant.

Speak Your Mind

*