Jonathan Morse sent me a TLS review by Ellen Jones of Ää: Manifiestos sobre la diversidad lingüística by Yásnaya Elena A. Gil. It begins:

Ää is a collection of essays in defence of indigenous languages, multilingualism and cultural plurality, written by a member of the Mixe community of the Sierra Norte of Oaxaca, Mexico. Yásnaya Elena A. Gil is a linguist and indigenous activist known for forging dialogues between communities, for excavating layers of racial and gender oppression and critiquing the effects of neoliberalism in Mexico. Here she writes in her second language, Spanish, which she began to learn aged about six, having until that point spoken only Mixe (known to speakers as Ayuujk). Hers is an unlaboured Spanish, suitable for recounting conversations with her monolingual, Mixe-speaking grandmother, as well as for doling out practical advice on how to stem the tide of linguicide (stay informed and spread the word, she insists: make sure you know, for instance, the difference between a language and a dialect, and which indigenous nations’ territory is split between multiple states).

Jones spends almost half the review complaining that the author doesn’t “interrogate the concepts ‘mother tongue’ and ‘native language,’” which seems churlish considering this may be the only English-language review the book will get. She does not, however, explain what most concerned me: what the devil does “ää” mean? I will spare you the trail I took through the Google Labyrinth, but I finally came out with the answer, courtesy of Rodrigo Romero-Méndez’s dissertation A Reference Grammar of Ayutla Mixe (Tukyo’m Ayuujk), whose Acknowledgements include: “Special thanks go to Yásnaya Elena Aguilar Gil, a dear friend and colleague, who convinced me to work on a Mixe language (instead of a Zoque language) and particularly in her own community, Ayutla Mixes.” On page 242 we find this example of the adposition mëët ‘associative’:

ës tsuj ää ayuujk mëët
and beautiful mouth word with
‘with good speech’

So there we have it: ää is ‘mouth’ (and ä, according to §2.3, Table 2, p. 27, is a low back vowel). For more about the Mixe languages, see Wikipedia; you can see a three-minute clip of René González Pizarro speaking his dialect of Mixe here (make sure to turn on the closed captioning). That page includes a discussion of his rhetorical style, beginning:

What the English translation doesn’t capture, however, are the poetic qualities of René’s speech. Take the second line, for example: ja’ tu’uk aa mäjtsk aa nkajpxa’any nyaka’any translated as ‘I’m going to share a few words.’ This sentence demonstrates nicely two features that typify eloquent Mixe. (Both of these features are also characteristic of skillful speech across Mesoamerica.)

Note that there the word for ‘mouth’ is written aa, not ää; dialect difference? (Thanks, Jonathan!)


  1. David Eddyshaw says

    Jones spends almost half the review complaining that the author doesn’t “interrogate the concepts ‘mother tongue’ and ‘native language,’

    You can’t really expect someone who hasn’t had our advantages to understand the subtleties of repressive colonial discourse without help from a friendly Westerner!

  2. The discourse deficiencies will be divided amongst the peasants!

  3. 249$? But why!?!?
    I do not understand.

  4. Ah, i am an idiot. It was Mexican peso! Still it is a 1000 of Russian roubles.

  5. It’s 229 Mexican pesos, which is about $11.50 in USD.

  6. Still it is a 1000 of Russian roubles.

    As Keith says, not much over ten dollars. By US standards, they’re practically giving it away — if it were published by a university press here, it could easily be ten times that.

  7. Ah, i am an idiot.

    You missed a chance to say “Ää, i am an idiot.”

  8. Trond Engen says

    I thought this post would be about Finnish long vowels.

  9. David Eddyshaw says

    Something for another day …

  10. ᴞ ḧ

  11. Dmitry Pruss says

    Last photo from Finland on my Facebook was PÄÄSISÄÄNKÄYNTI

  12. Rob Solheim says

    Is the author’s first name taken from Russian ясная, I wonder? Is this a common name in Mexico?

  13. “ Is the author’s first name taken from Russian ясная, I wonder? Is this a common name in Mexico?”

    I bet it is taken from Russian and no, it is not a common name in Mexico.

  14. — Fagradalsfjall in Iceland is erupting now. Let’s go and walk right up next to the pāhoehoe!

    — ʻAʻā.

    (Joke I heard somewhere recently.)

  15. David Marjanović says


    jäääärne “ice edge”

  16. William Boyd says
  17. jäääärne

    Winner of the Ää Äwärd for 2021!

  18. Trond Engen says

    From Wikipedia on Mixe languages:

    Mixe phonology is complicated and little documented.[citation needed] There is a palatalized series of all consonant phonemes (as in Russian, Polish, or Gaelic) and possibly a fortis/lenis distinction in the stop series, the recognition of which however is obscured by a tendency of allophonic voicing of consonants in voiced environments. Syllable nuclei are notoriously complex in Mixe, varying in length and phonation. Most descriptions report three contrastive vowel lengths.[3] There are multiple values of phonation, one being the typical one (what some phoneticians call “modal voicing”). The other types of phonation have been variously termed checked vowels, creaky voice vowels and breathy voice vowels. Some Mixe variants are vowel innovative and some, notably North Highland Mixe, have complicated umlaut systems raising vowel qualities in certain phonological environments.

    We may assume that the phonemic length contrast (or one of them) is represented by the double vowel. One might imagine that the umlauts are used for raised vowels in North Highland Mixe, but that’s probably too much too ask for, and if the raising is allophonic, the contrast wouldn’t even (or necessarily) make sense to a native speaker. Maybe they’re used for a “value of phonation”.

    Also, it annoys me that the Wikipedia article omits the two Mixe languages of the Veracruz province for no reason at all.

    When I say “for no reason at all”, I mean “for reasons that are not clear to me but I suspect to be political”.

  19. Trond Engen says

    This is obviously handled in Rodrigo Romero-Méndez’s dissertation. I’ll report back.

  20. Trond Engen says

    There’s a lot more to be said about the vowel system, but for the basic notation.

    3.2.1 Vowels and their allophones

    As expressed above, vowels contrast in quality and length, in addition to forming complex nuclei. Here, I will present the allophonic variation without taking into consideration length or laryngealization, because the allophones dealt with will be also found in long or laryngealized vowels. Nonetheless, it is worth pointing out that long vowels in general are more resistant to phonological processes.

    First, I will describe the vowels and then the most common allophones.

    1. /i/ high front unrounded vowel.
    a) /niː/ [niː] nii ‘hot pepper’
    b) /tiː/ [tiː] tii ‘what’

    There are only a few words whose phonological representation has /i/. It is found in many words, but in most cases it is the result of the vowel metaphony. I will come back to discussing its phonological status in §5.2.

    2. /e/ mid front unrounded vowel.
    a) /nek/ [nekʰ] nek ‘humid’
    b) /tek/ [tekʰ] tek ‘(small) lizard’

    3. /a/ low front unrounded vowel.
    a) /kaː/ [kaː] kaa ‘very much’
    b) /tam/ [tam] tam ‘like’

    4. /ʌ/ open-mid back unrounded vowel.
    a) /hʌw/ [hʌw] jäw ‘feel!’
    b) /kʌk/ [kʌkʰ] käk ‘basket’

    5. /ɤ/ mid back vowel with inconsistent lip rounding
    a) /tɤː/ [tɤː]~[toː] too ‘rusty’
    b) /akʂɤn/ [akʂɤn]~[akʂon] akxon ‘very’

    While older speakers usually produce this vowel as unrounded, it tends to be produced as a mid back rounded vowel among younger speakers. As far as I can tell, the variation is related to different idiolects, not to any phonological context. In the previous examples, it is more likely that an older speaker would produce an unrounded vowel while a younger speaker would produce a rounded one. In order to avoid further confusion, I will systematically represent this vowel as unrounded [ɤ], but bear in mind that it could also be pronounced as rounded [o] in any case.

    6. /u/ high back rounded vowel
    a) /ku/ [kuː] kuu ‘when’
    b) /jʔuk/ [j̰ukʰ] y’uk ‘his dog’ (FrogA-85)

    7. /ɨ/ mid-high central unrounded vowel
    a) /hɨpk/ [hɨpʰkʰ] jëpk ‘corn (on the cob)’
    b) /mɨh/ [mɨh] mëj ‘big’

    Perhaps, /ɨ/ is the vowel that presents the most variation, particularly between stressed and unstressed syllables. However, even among stressed syllables, it can be articulated as a high central unrounded vowel or as a mid-high unrounded vowel.

  21. Yes, I am sorry. I always pay attention to books prices in various countries. But 250 for Mexico looked a way more outrageous than what I used to call outrageous. Then I realized it is pesos. As for 1000 roubles – the question is whether it is accessible for an average Mexican.

  22. I’ll bet Yásnaya Elena Aguilar Gil was named after Tolstoy’s estate.

  23. I thought the same. In Russian the word actually has connotations, associations and sound that could make it absolutely appropriate for given names, but it is an adjective (a long form of it). There are morphosyntactic issues that make it unlikely that a Russian would use this form as a name. But they were not embedding this name in Russian:)
    Once I met a Brazilian with a common Russian name, and it was not by chance: her parents were Communists.

  24. John Emerson says

    We can presume that Stalin Colinet’s parents were Communists, I think. Stalin, if I can speak of him so familiarly, was one of the very few Dominicans ever to play in the NFLand also one of the very few graduates of a NYC HS ever to do so.

  25. PlasticPaddy says

    Yasna is also a female name in South American Spanish. But it is not clear what the origin of this name is.

  26. Jasna is a South Slavic female name, occasionally found in Poland, Czechia (as Jasňa) and probably Slovakia as well. It might have been carried over to South America by Slavic immigrants, Croats more specifically.

  27. jäääärne

    jää +‎ äär.

    jää-äär (genitive jää-ääre, partitive jää-äärt)

    The edge of a stretch of ice.

    Usage notes
    This term is spelled without the hyphen (“jäääär”) in a number of non-Estonian texts, in which the word is presented as a curiosity; the unhyphenated spelling is nonstandard.


  28. Lars Mathiesen says

    Danish had Graaand until 1947, now gråand (‘mallard’). I have tried to come up with words with either a-å or å-å, but words ending in -a are mostly female given names, and words starting in å- are not common either. (Also it was Graa-and in the dictionary at least).

    Maybe Smaaaager would work — ‘petty usury,’ you might find yourself wanting to express that outside of word games. Or Lavaaand, ‘lava spirit’.

  29. Trond Engen says

    Blaaaas “Blue hill”? The problem is that you don’t have hills. But you have rivers: Graaaa “Grey river”.

  30. January First-of-May says

    Back in 2015, I once tried to figure out what was the alphabetically first place name in the world; when I stumbled on Aabenraa, Denmark, I thought that was it, but alas it apparently isn’t.

    (Best I could tell at the time it was either Aaartali, Bangladesh, or Aabadiye, Lebanon. Of course in some forms of alphabetical order numerals come before letters, and it wasn’t very clear whether e.g. “Aaba” would come ahead of “Aabadiye” either. There are several places just called “A” or “Aa”, and of course a lot of places called Å, but that isn’t really at the start of the alphabet, is it?)

  31. Aabadiye

    Alas, this one starts with a ʿayn, which is the eighteenth letter of the alphabet.

  32. John Cowan says

    Something could be done with Småålande ‘the uninhabited islands between Sweden and Finland’, perhaps.

  33. Trond Engen says

    Of course, a nymph or a nixie is an aaaand. The nixies of Graaaa would be Graaaaaande. The little nixies would be smaaaaaande.

  34. Trond Engen says

    You could also argue for Graaaaaaaanden “The Grey River river spirit”, but that may come through as contrived,

  35. Lars Mathiesen says

    What I’d really like was a compound that could split either way, a-aa and aa-a. But a/å doublets are rare if they even exist, I can’t think of one, and for two such to combine ‘both ways’ and make sense is not likely.

    In the meantime, have sukkerrør and sukkerroer (sugar canes and beets) — proving that respelling æ and ø as ae and oe is not always safe and resolvable by context. ((En) roe has two syllables, (et) rør is idemplural).

  36. Athel Cornish-Bowden says

    Until it broke when I dropped it on the floor I used to drink my coffee after lunch from a cup that announced its price on the base as something like 2461$. However, the $ were Chilean pesos, as I bought it in Santiago airport.

  37. Athel Cornish-Bowden says

    Yasna is also a female name in South American Spanish. But it is not clear what the origin of this name is.

    The Jasna that I knew in Chile spelt it with J but it looks like the same name, especially as the J was pronounced as Y. She was from a Croatian family in Punta Arenas, the southernmost city in the world, which has a substantial Croat population.

    Sorry, I wrote the above before seeing that Alex K. had said something similar.

  38. Don’t apologize, there is no requirement or expectation of parsimony in a LH thread.

  39. Zooming out and in in google/yandex maps to get from my humble shithole to Aaartali does feel like flying.

    Especially that there is no Aaartali on Google maps… Hm. the place I landed in is
    আশারতলী .

    But Chittagong is a good place:

    It was historically settled by many tribal refugees from Burma Arakan in 16th century and now it is settled by the indigenous Jumma people. Today, it remains one of the most heavily militarized regions of the world, and one of the least developed parts of Bangladesh.
    Oops. Not this. This but not the “militarized” part. This:
    The tribal populations include the Chakma, Marma, Tripura, Tanchangya, indigenous Assamese, Keot (Kaibarta), Chak, Pankho, Mro, Murang, Bom, Lushei, Khyang, and Khumi,[3] and differ markedly from the Bengali majority of Bangladesh with respect to language, culture, physical appearance, religion, dress and farming methods

  40. Punta Arenas, the southernmost city in the world

    Or Ushuaia. Depending on whether you think 120k population qualifies as a city but 60k does not, and more importantly, on whether you are Chilean or Argentine.

  41. Athel Cornish-Bowden says

    My wife is Chilean, so I think that answers your question.

    Wikipedia gives 127,454 for Punta Arenas and 56,956 for Ushuaia. I tend to think of Ushuaia (which is only just in Argentina) as a tourist resort whereas Punta Arenas is real city. I guess I can’t claim that Puerto Williams is a city!

  42. Puerto Toro is not a city either…

  43. My relatives moved to a city with some 4 thousands of population, a former capital of a forgotten princedom. It is still waiting its first tourist to call it a “tourist resort”, I think.

    A city must have a city wall.

  44. A city has to have a cathedral, according to the British.

  45. A city has to have a cathedral, according to the British.

    I looked up Asgard in Wikipedia to disprove this and was terrified to see “enclosure” instead of “city” (Russian gorod).;-/

  46. An ancient Greek city had to have an agora and a theater, if I remember correctly.

  47. That’s just something I remember from my visit to St Davids (Tyddewi), which bills itself as the smallest city in the UK (population 1,600).

  48. Wikipedia’s list of northernmosts calls everything a city, beginning with Longyearbyen (pop. 2400). However, its list of southernmosts goes from “Town > 1,000 people” (Puerto Williams, pop. 2900) to “City > 50,000 people”. This is an outrageous inconsistency, which I shall expose in a letter to The Times.

  49. J.W. Brewer says

    I guess “southernmost city that has given rise to multiple rock bands whose music I like” is too subjective an accomplishment for wikipedia to note? (That’s Dunedin, N.Z., but if there’s a band from Punta Arenas I should know about, I’m willing to listen. The ’80’s Boston band the Del Fuegos I guess evoked extreme southernness with the name but I’m not sure they had ever personally been south of the Equator.)

  50. January First-of-May says

    the southernmost city in the world

    The southernmost town in the world, meanwhile, appears to be Villa Las Estrellas, one of only two civilian settlements in Antarctica (the other one, Esperanza Base, is farther south but is apparently classified as a research station).
    Weirdly, Wikipedia does not mention either of the two on its list – not even on the linked “southernmost settlements” page. (Of course McMurdo is larger and further south than both, but it is also just as obviously not a town.)

    the smallest city in the UK

    Meanwhile the smallest city in Russia is Chekalin, pop 866 (est 2021); it bills itself to tourists as Russia’s smallest city (now apparently true again by 2021 estimates, though for a few years that spot was technically taken by the planned settlement of Innopolis), but there aren’t many tourists, and the city barely survives.

    I posted somewhere (…apparently in a Discord chat) a few years ago about L’Île-Dorval, Quebec, as being supposedly the world’s only city with an official population of zero; that might have been the case as of 2001, but the 2016 census gives it a population of 5, and I’m not sure if it’s even (still) the world’s least populated city. (It apparently lost city status in 2002, but regained it back in 2006.)

    It is said (the accounts vary) that Zashiversk, in a remote part of what is now Yakutia, lost its city status in 1890, some years after the authorities in Yakutsk finally noticed that the place no longer had any inhabitants.

  51. I’d like to know the record for “northernmost cheerful band”.

  52. Athel Cornish-Bowden says

    Until I went to Tromsø for a PhD exam a few years ago, the most northerly place I had ever been to (not counting flights and airports) was Söderboda, in Sweden, of which the name means “southern village”. Until you get to the north pole there is always somewhere further north, in this case Norrboda.

  53. David Eddyshaw says


    Welsh WP patriotically says it’s the smallest city in Wales. Nobody cares about the UK …

    The soft mutation is peculiar. I suppose it’s one of those things like Hywel Dda.

  54. PlasticPaddy says



    iv) The first consonant of the second element of a compound word (that tells us who the first element of the word refers to) is often softened when the first element is a feminine noun singular.

    This sometimes occurs even though the first element is masculine. For example:

    Tŷ (house) + Dewi (name of a saint) becomes Tyddewi.

    Cae (field) + Dafydd (name of a person) becomes Cae Ddafydd.

  55. David Eddyshaw says

    Well, yes: it’s because the name has gone from being a well-behaved noun phrase to being a (sort of) compound, and there are plenty of other examples, especially with names of people (to say nothing of Arglwydd Dduw.) It’s still peculiar.

    I wonder if it has anything to do historically with placenames being often used in the locative? (It having just occurred to me that the phenomenon with masculine personal names like Hywel Dda “Hywel the Good” might be a generalisation of a vocative pattern …)

  56. ää is ‘mouth’

    Tenuous False Cognate of the Day: Proto-Samoyedic *aŋ ‘mouth’ > Proto-Nenets-Enets *ŋʲäŋ > Proto-Enets *ŋʲäʔ > pre-Tundra Enets *ääʔ > Old Tundra Enets /eːʔ/ (> modern /eʔ/).

  57. Proto-Samoyedic *aŋ ‘mouth’ > Proto-Nenets-Enets *ŋʲäŋ

    How does that work — Proto-Nenets-Enets just sticks ŋʲ before initial vowels?

  58. Basically yes, originally just *ŋ but palatalized before front vowels; the famous “Michael Jackson sound change” (also occurring in Nganasan).

    In Tundra Nenets this even remains productive: ‘airport’ is ңэ̇ропорт /ŋa͡eroport/, ‘electricity’ is ңэлетричества, ‘oblast’ is ңобласть, etc.

  59. Proto-Nenets-Enets just sticks ŋʲ before initial vowels

    Curiously, Initial ㅇdoes the same, which is not pronounced, though.

  60. Fascinating!

  61. Trond Engen says

    My son suggests it’s a physiological response to the Siberian winter. It’s what happens when you inhale before talking, try not to pull cold dry air into your lungs, and hold the constricture a little into an initial vowel.

    But that doesn’t explain Michael Jackson.

  62. What’s it called again when you append a sound to the beginning of a word specifically…? 🙂

  63. J Pystynen:

    The Wikipedia has this on the prosthesis of /ŋ/ in Samoyedic:

    In Nganasan, Nenets and Enets, PS vowel-initial words gain an initial /ŋ/ via rhinoglottophilia (which may be subsequently palatalized to /nʲ/). This is occasionally found in other Samoyedic languages as well, usually with the exception of Mator.


    Is there any more explicit account available of exactly what might have happened? By prosthesis of */h/? Or by nasalization of the vowel after initial /ʔ/ and that nasalization somehow reanalysed as an onset /nʲ/? (Do initial nasals nasalize following vowels in Samoyedic significantly?) Or nasalization of a vowel preceding /ʔ/ in sandhi and then metanalysis/mis-segmentation?

  64. I would not think thru prothesis of /ʔ/, Samoyedic languages are the one part of Uralic that has been very happy to add glottal stops around (aside from Livonian, as hatted recently) and yet we don’t see any word-initial ones in the Samoyedic languages that don’t have prothetic *ŋ. “Rhinoglottophilia” was apparently added here by an anonymous editor just last fall, I’ve never heard that used for this before. (Maybe some confusion with word-final *n > /ʔ/ in Nenets and Enets?)

    Speculatively this could be a Paleosiberian substrate feature, since in Eurasia northwest of Sino-Tibetan, word-initial /ŋ/ is found only in northern Samoyedic, Tungusic, Chukotko-Kamchatkan and Nivkh (and, apparently, Dolgan — due to Samoyedic loanwords I bet). This kind of fits also with how westernmost dialects of Tundra Nenets, in substantial contact with Komi / Russian / Sami, have again lost those cases of initial /ŋ/ that were not palatalized to /nʲ/ (but then the significance of this is surely undermined by the same also happening in Enets).

    An old theory also was that already Proto-Uralic had initial *ŋ- and it was just lost elsewhere, but this is refuted by developments like Proto-Samoyedic *wota > Tundra Nenets /ŋōdʲā/ ‘berry’ (with *wo- > *o- > /ŋo-/; contrast Selkup *kotə) or PSmy *jïntə > TN /ŋin/ ‘bow’ (with *jï- > *ï- > *ŋï- > /ŋi-/; contrast Nganasan /dʲintə/) which allow dating *ŋ-prothesis as fairly recent in relative chronology.

  65. Thank you, J Pystynen! Very interesting.

  66. J Pystynen! If you see this…

    word-final *n > /ʔ/ in Nenets and Enets

    What exactly is the relationship between the initial nasal of Nenets ńeneć’ and the apparent lack of initial or glottal stop(?) in Enets enet’ (form correct?)? Is this an example of the prosthesis described above? From what you said above, if I understand correctly, the prosthesis of the nasal also occurred in Enets.

  67. Yes, that’s a prothetic ń-. This term meaning ‘true people, true ones’ (before Soviet times not an ethnonym in Enets, but just the word for ‘people’) is derived from Proto-Samoyedic *änå ‘real, true’; no wider Uralic cognates known but original lack of anlaut is seen also in a Selkup cognate āni̮ľ ‘indeed’.

    The model I’m following is from Sammallahti (Finnisch-Ugrische Forschungen 41, 1975) which posits for Enets the following history:
    1) epenthesis of *ŋ before all word-initial vowels;
    2) palatalization of *ŋ to *[ŋʲ] before front vowels;
    3a) some kind of dissimilation of *ŋʲä to *ŋʲa in Forest Enets (for secondary *ä from *a)
    3b) this now marginally phonemic *ŋʲ merging with *nʲ;
    4) common Enets loss of unchanged word-initial */ŋ/, including remaining *[ŋʲ].

    Steps 1), 2) and 3b) are paralleled by Nenets, 4) by some cases of loss of intervocalic *ŋ. From a purely Enets perspective there might not be much reason to assume a general epenthesis stage, but it does seem to help in providing a natural path adding up to the much more specific epenthesis of *a- to Forest /na-/ versus Tundra /e-/ (e.g. F naddo ~ T eddo ~ Nenets ńāmt° ‘horn’ < *ämtə < *amtə).

  68. David Marjanović says

    Curiously, Initial ㅇdoes the same, which is not pronounced, though.

    I doubt this was ever supposed to indicate [ŋ], though. I’ve long suspected it is what it looks like: a zero, for the zero initial of Chinese philology.

  69. Curiously, initial ㅇdoes the same, which is not pronounced, though.

    As far as I understand it, the Modern Korean use of ㅇ results from the graphic conflation of the original null initial consonant letter ㅇ and the archaic letter ㆁ used to represent the velar nasal, including the theoretical velar nasal initial of Sino-Korean words. The words written with a velar nasal initial at the time of the invention of Hangul were probably pronounced with no initial at the time—in fact, that was one of the reasons for the creation of Hangul in the first place, to remind people that it had been there. The Wikipedia has a nice summary of this view:

    The explanation of the letter ng also differs from the traditional account. Many Chinese words began with ng, but by King Sejong’s day, initial ng was either silent or pronounced [ŋ] in China, and was silent when these words were borrowed into Korean. Also, the expected shape of ng (the short vertical line left by removing the top stroke of ㄱ) would have looked almost identical to the vowel ㅣ [i]. Sejong’s solution solved both problems: The vertical stroke left from ㄱ was added to the null symbol ㅇ to create ㆁ (a circle with a vertical line on top), iconically capturing both the pronunciation [ŋ] in the middle or end of a word, and the usual silence at the beginning. (The graphic distinction between null ㅇ and ng ㆁ was eventually lost.)

    Note also Ki-Moon Lee and S. Robert Ramsey, A history of the Korean language, p. 117, on the letter ㆁ :

    The character readings in the Tongguk chŏngun dictionary of 1447 were certainly artificial. For example, among its initials the Tongguk cŏngun contains geminate consonants (ㄲ, ㄸ, ㅃ, ㅉ, ㅆ, ㆅ) to represent “wholly muddy” [voiced] sounds, as well as a glottal stop (ㆆ) and a velar nasal (ㆁ). These were prescriptive pronunciations intended to “correct” the readings of Chinese characters then in use in Korea.

    The Hunminjeongeum Haeryebon “Explanations and Examples of the Correct Sounds for the Instruction of the People” (the royal explanation of Hangul written in literary Chinese) of 1446 says this of the letter ㆁ :


    ㆁ is a velar sound like the initial of the word/character 業

    (業 “work, karma” is now Mandarin , from Middle Chinese /ŋɨap̚/ or the like (Pulleyblanks reconstructions for ease of reference), illustrating the loss alluded to in the Wikipedia entry. In other languages, you find Sino-Japanese (Go-on reading) ごう (< earlier ごふ gopu or the like), Sino-Vietnamese nghiệp “karma”, and among the more conservative Sinitic languages, Chaoshan Min ngieb8 (sound file here), Meixian Hakka ngiab6, etc. or the like (Mandarin ). Modern Korean has 업 eop [ʌ̹p̚].)

    Contrast this with the null ㅇ :


    ㅇ is a throat/laryngeal sound like the initial of the word/character 欲

    (欲 “desire, wish” is Middle Chinese /juawk̚/ (Pulleyblank again), Mandarin , modern Sino-Japanese yokuよく, Sino-Vietnamese dục (d here is Hanoi /z/, Saigon /j/).)

    Another symbol was also invented for the glottal stop.:


    ㆆ is a throat/laryngeal sound like the initial of the word/character 挹

    (挹 “to ladle out, scoop out, bale out, decant” is Middle Chinese /ʔip̚/, Mandarin , Sino-Vietnamese úp, modern Sino-Japanese yū ゆう, ō おう, modern Korean 읍 eup. Lee and Ramsey, p. 117, say of this letter: “like the geminates, the letter was created for the representation of Chinese character pronunciations”.)

    And note also Lee and Ramsey, p. 150, on the phonological system of Late Middle Korean and the sound [ŋ] at time of the introduction of Hangul:

    The distribution of the velar nasal ng (ㆁ) was more restricted. It did not occur in word-initial position, though it was apparently found in morpheme-initial position in 쇼ㆁ아지 syongaci ‘calf’ (sywo 쇼 ‘ox’ + diminutive -ngaci ㆁㅏ지) and the polite marker -ngi- ㆁ as well as in a few Sino-Korean words such as sange 사ㆁ어 ‘shark’ (sa 사 鯊 + –nge ㆁㅓ 魚). But there is little reason to believe that ng has ever occurred in word-initial position in Korean.

    (When you read this quote, in the words 쇼ㆁ아지 “calf” and the others, the 쇼ㆁ syong syllable block and the other syllable blocks containing ㆁwill probably fall apart in posting and not be properly composed as Korean syllable block should be, because electronic typographical support for the letter ㆁ is poor.)

    As I understand it, the peculiar modern use of ㅇfor zero at the beginning of a syllable and /ŋ/ at the end doesn’t really result from the loss word-initial /ŋ/ in the history of Korean. Rather, the minuscule graphic difference between the velar nasal ㆁand the zero initialㅇ began to be ignored because the use of ㆁ in initial position was a learned prescription not corresponding to the usual phonetic reality, which was the zero initial. The actual zero initial letter ㅇ could be used instead, both to represent the zero initial and velar final.

    الله أعلم

  70. J Pystynen, thank you for that explanation of the Samoyedic nasal initials!

  71. John Cowan says

    (It having just occurred to me that the phenomenon with masculine personal names like Hywel Dda “Hywel the Good” might be a generalisation of a vocative pattern …)

    As is the case with the names Charles, George(s), Yves, and James (now in English only), which maintained the -s of masc. nouns in the nom. sg. due to their vocative use.

Speak Your Mind