Talk:Mongolian language/Archive 1

changes
Hi Whimemsz!

I've just seen the changes of the Mongolian language side, and they are not all for the better. On the one hand it now (as intended) focuses exclusively on Halh (thy Khalkha) and the only word that was intended to be a transcription of the Mongolian script is transcribed in a wrong way (in  the second  would have to be a Greek gamma). This is not really justified as the number of speakers in Inner Mongolia is either equal or even surpasses the number of Halh speakers. (You would of course propose that Inner Mongolia is divided into Ordos, Chahar etc., but that still leaves the fact that the majority of those 5.7 million speakers do not speak Halh.) But most curiously, reading the IPA of the Cyrillic letters, I am rather reminded of the Inner Mongolian pronunciation. For as to phonemes, the Halh dialect has no voiced labial or dental plosives, and the same is true of the affricates. The contrast is instead achieved by aspiration, such that the allegedly voiced phonemes go unaspired: &#1076; is /t/, &#1090; is /th/ etc. (The h of th should be put up.) For more details see “Svantesson et alii (2005): The phonology of Mongolian. New York: Oxford University Press” that deals extensively with Halh, but with diachronics of Mongolic as well.

The worst thing about the side is that we still don’t learn anything about phonology (except a hint at phonemes: you list just one single /g/; you wouldn’t be understood, e.g. bag “team”, baG “small”), morphology, syntax, vocabulary etc. Most basically, it would have to be said that Mongolic has vocal harmony, that it is agglutinating (suffixes only) and that word order is SOV. There are lots of interesting aspects as well, e.g. case system, postpositions, verbal system (verbal nouns, converbs, finite verbs; mode, evidentialis) and of course word classes (I wouldn’t trust Poppe in that respect), but I know quite well that no book available (at least as the English reader is concerned) is reliable enough. If this didn’t discourage you, you might try to use “Juha Yanhunen (ed.) (2003): The Mongolic language. London: Routledge” as a basis. If I am to find the time, I might deal with these subjects myself, but for the time being, that three most basic aspects of grammar and a change of a bunch of IPA entries should suffice. If you provide me with your email, I could mail you the transcription list of Svantesson et alii.


 * Thank you for your observations. I have to say right from the start that I know hardly anything about Mongolian -- the little I do know comes mainly from a Lonely Planet guide, which I'm sure isn't the most reliable of sources. But I saw that there was hardly any information at all on this page, so I decided (perhaps mistakenly), that it would be better to add SOMEthing. I'll try to see if I can pick up some of those books soon, and try to fix the article as soon as I can. I don't really know what to do in the meantime, so if you or anyone else who knows Mongolian wanted to help fix my screwups, that would be fantastic... Thanks again for the references and the advice! --Whimemsz 02:12, Jun 1, 2005 (UTC) ((By the way, you can just use my Wikipedia userpage to get my email address))


 * I think this page and the discussion could be useful : French Wikipedia, though there are many mistake in the pages.--Henri de Solages 02:12, 6 November 2005 (UTC)

I’m just wondering what good there might be in adding „useful phrases“ to a lexicon entry. Anyway, they’d rather be consistent with either the transcription or with phonetics. Neither is the case. E.g. something like “bayarlaa” is said, but the transcription is “bayarlalaa” and the API-standard would look slightly different as well. There are a good number of mistakes as well, e.g. “oroslar” (orosoor), “a lot – tom” (“a lot” is ikh, “tom” is “large”), “down – doo” (which is correct, strictly speaking, but “doo” can never be used without one of the suffixes –r and –sh). So I think it would be best to altogether abandon this section.


 * And in any case, why are the phrases at the very bottom of the article? Shouldn't they be just before the references and external links sections? Being at the very bottom sort of detracts from the overall layout, at least for me, since I feel like sections like "references" and "external links" and so forth should be at the very bottom... --Whimemsz 01:16, 1 January 2006 (UTC)

Observations / Sergey Radchenko
The article is a bit weak both on the theoretical side, and in practical examples. On the theoretical side, it would be nice to see more discussion of the links between Mongolian and Buryat / Kalmyk languages, more thorough analysis of grammar and writing. It would be useful also to compare Inner Mongolian vocabulary with Khalkh Mongolian, as there are substantial differences. It should be pointed out, too, that in Inner Mongolia very few people will actually speak Mongolian, at least in towns. Of course, their road signs are in Mongol bichig, but who can read them?

On the examples side, the "useful phrases" are fairly misleading. As one commentator noted, there is little regard here for spelling rules. But in any case, as long as the average tourist out there can benefit from it. I also found a few mistakes in the actual phrases, corrected those - but there are quite a bit more mistakes, so I just gave up on this for now.

Maybe it would be worth getting a native Mongolian speaker (preferably a linguist) to go through this article and offer corrections. Anyone knows of one?

Transcription for Mongolian
I have posted a request for opinions about a standard transcription table for mongolian on Wikipedia talk:Naming conventions (Cyrillic). Please join the discussion there. --Latebird 20:57, 10 March 2006 (UTC)

Recent changes
Recent changes

While adding the grammar part (except phonology), I’ve taken away the list of “useful phrases” the use of which didn’t become clear at all. I didn’t insert any hyperlinks, thus a helping hand to put this right would be welcome. Furthermore, my English is rather mediocre, and while this text has been checked by one doctor of Mongolian studies and linguistics, no native speaker of English has read it through so far.

I disagree with Sergey Radchenko that a comparision to Buryat/Kalmuck would’ve been helpful, as both are not to be regarded as parts of Central Mongolian, but as independent languages within the Mongolic language family. However, South Central Mongolian lexical items (as well as phonology and other grammar, for that matter) would have their merits, preferably with a little information about the development of the Proto-Mongolic sound system.

And either Buryat is a related language or a dialect! Thus we shouldn’t extend “Mongolian”’s geographical distribution to Buryatia, or at the least we shouldn’t talk about significant numbers. G Purevdorj 14:18, 13 June 2006 (UTC)

Cyrllic vs. Roman
This article about a language I'm interested in has been vastly improved lately with the addition of a wealth of grammtical, phonological, and syntactic information, and the deletion of the "useful phrases." I have a suggestion that someone might want to think about, though. This is the English langauge version of wikipedia. Most English speakers are illiterate (or in my case, nearly-illiterate) in the Cyrllic alphabet. I can agonize my way through most of it and yeah, it's probably good for me, and yes, I know this is the way Mongolian is written, but the bottom line is, I don't know the words I'm looking at. I have to take som much time to figure out what I'm looking at that it gets in the way of my grasping the grammatical examples. I would understand them so much more quickly if they were Romanized. No reason you couldn't display sample sentences in both alphabets. I realize wikipedia isn't all about me, but I'm sure i'm not the only one in this boat. If you follow my suggestion, thanks. If not, thanks at least for considering it. C.M.71.215.128.134 18:04, 14 June 2006 (UTC)

Cyrillic can be learned within a couple of hours and is certainly wholesome, but as this should also be a text for people with no particular interest in Mongolian, here you are. The downside is that I had to choose a more precise transcription that is less intuitive to English readers. G Purevdorj 22:51, 14 June 2006 (UTC)

Thanks! And I like the more accurate, less "Englishy" looking approach.

no /k/?
the phonology says that Mongolian doesn't have /k/, but it does have /g/... is that true?! I was under the impression that was impossible. --Krsont 19:27, 19 June 2006 (UTC)

[k] turned into [x] and, well, was absent from then on. (In some Northern dialects we still have an allophon [k] of /x/ with non-pharyngeal vowels.) Now in loanwords from Russian, [k] is reintroduced and might already be considered a marginal phoneme again. G Purevdorj 22:18, 20 June 2006 (UTC)

no /b/?
It also says Mongol lacks /b/. What about the city of Ulaanbaatar? or the fellow who gave me a ride to Khatgal: Batbayaar? C.M.65.102.39.98 20:56, 11 July 2006 (UTC)


 * The phonology section only describes the most prevelant dialect of Mongolian, Halh, which, from what I understand, does not have a voiced labial plosive (/b/). The letter for /b/ is pronounced /p/ in this dialect.  Other dialects do have the /b/ phoneme.--WilliamThweatt 21:19, 11 July 2006 (UTC)


 * Hmm, my impression was that /b/ was a pretty prominent phoneme. Not saying you're wrong, I'm just surprised, but you seem pretty confident in saying that.  So, Ulaanpaatar? C.M.65.102.39.98 19:07, 12 July 2006 (UTC)


 * Well, I am trained in linguistics, but will admit I know very little about Mongolian. However, I do know that the situation is very complicated by many factors including the many different dialects, the isolation of the communities of speakers of the different dialects, the separation of Mongolia into "Outer Mongolia" which is heavily influence by Russian and "Inner Mongolia" which is heavily influenced by Mandarin Chinese, and the use of different writing systems over the centuries and in different communities.  With all that in mind, ancient Mongolian did have the /b/ phoneme.  That is why it is present in placenames and in the "alphabets".  Over time, /b/ has come to be pronounced as /p/ in some dialects, as an intermediate between /w/ and /b/ (bilabial continuant) in others, and retained as /b/ in yet others.  As for prominence, speakers of dialects which keep the /b/ (or something close to it) may indeed outnumber those of dialect which do not -- however, the "standard" or "prestige" dialect is based on Halh, which, according to all my sources does not have /b/.--WilliamThweatt 21:13, 12 July 2006 (UTC)


 * I don't actively speak the language, but I hear it a lot. From my non-linguistic experience, the b in eg. Ulaanbaatar is proncounced very similarly to the pronounciation in English. That's also the comparison that you'll find in most teaching literature. I may not understand the difference correctly, but my impression was that it's rather the /p/ that is missing. --Latebird 07:23, 13 July 2006 (UTC)


 * I guess there are two reasons why /p/ frequently appears to you like [b]. Firstly, if you’re German or the like, your own mother tongue almost always combines the features voicedness and aspiration to differentiate between /b/ and /ph/ (meant to be one phoneme), and there is even data suggesting that aspiration is more important. Thus, you would map [p] onto /b/. Secondly, Svantesson et al. (“The phonology of Mongolian”, 2005) didn’t take into account any contextually motivated voicing. Therefore,  might have [b], because it is preceded by a (voiced) nasal, but has [p]. G Purevdorj 23:39, 18 July 2006 (UTC)


 * Judging from what little Mongolian I have heard, the Mongolian б is a voiceless lenis, . This is the same sound as the one written b in Pinyin and (in many cases) the Revised Romanization of Korean. It is also the word-initial allophone of /b/ in many Englishes, the only realization of /b/ in southern German, and (as far as I can tell) the only realization of /p/ in Spanish.
 * Most languages that have voiceless lenes have an aspiration contrast (indeed, Mandarin, Korean, and to a lesser extent English do, and so does Mongolian). Still, voiceless lenes and unaspirated fortes are not the same. Languages like French, Russian, and Japanese have unaspirated fortes ([p t k]); these don't sound the same as, and probably shouldn't be transcribed as such in a phonetic transcription. Southern German, among a few other isolated cases, even has a phonemic contrast between voiceless lenes and unaspirated fortes.
 * However, Mongolian word-final г really is [k], not [g˚]… at least in the one case where I was able to listen closely enough. David Marjanović 16:55, 9 June 2007 (UTC)


 * Karlsson (Rhythm and intonation in Halh Mongolian, 2005: 60, 153) provides the possible realizations of the word-final clusters tg [tg, tk], ɮg [ɮg, ɮk, ɬg, ɬk] and ʦg [ʦg]. The phoneme must be /g/, because you will never hear [k] when /g/ becomes intervocalic after a suffix is added. G Purevdorj 23:14, 9 June 2007 (UTC)

Palatalization
Why are the palatalized consonants considered separate phonemes? Is there any difference between a palatalized consonant and a consonant + y? i.e. is "ty" a palatalized t?  If so, isn't this just a cluster?65.102.39.98 19:07, 12 July 2006 (UTC)


 * Palatal consonant is an article here that may answer your question. Briefly though, "palatization" is used in linguistics to describe the "place of articulation" (the palate, as opposed to dental, aveolar, retroflex, etc).  A palatal /t/ is articulated with the blade of the tongue against the palate, while a dental /t/ is articulated with the tongue against the back of the upper teeth.  This is a single sound (phoneme).  In linguistics, "ty" would indicate two seperate phonemes pronounced in sucsession (written in IPA as /tj/)--WilliamThweatt 21:20, 12 July 2006 (UTC)


 * The "phonetic description" part of Palatalization will help you more. BTW, WilliamThweatt, you mean [], not //, because you're talking about sounds, not phonemes. David Marjanović 17:00, 9 June 2007 (UTC)

Morphology
I've read a little bit about linguistics, but I know too little of that (and nothing of Mongolian) to contribute anything of substance to this article. However, I believe that the morphology section needs to be revised: it seems to contain a talk-page-style argument. Seeing a "(still?)" and similar notes doesn't make it appear very accurate to me Nyttend 20:44, 29 August 2006 (UTC)

I dislike to speak about abstractions roughly depicted by one single example. However, I shall refute the single point that you directly critizised. Grammaticalization may turn proper nouns to demonstrative pronouns to personal pronouns. Now, it isn't clear at all if the Mongolian demonstrative pronouns still necessarily exhibit any demonstrative meaning or if they have already become (distal and proximal) personal pronouns. There are linguists who argue for the latter, such as Sechenbaatar. Yet, this position has not been that well established yet. So I wrote down the more conservative opinion and inserted the "(still?)" to make sure that nobody too readily relies on this data. Another point that you "might" have critizised is, for example, "(probably perfective, otherwise past)". We do not have any sufficient analysis of the Mongolian aspect system. Uwe Bläsing provided one for Kalmyk, and though this has not been adapted to Mongolian proper, some similarities between the two languages clearly show that some points of the nowadays commonly held beliefs about certain Mongolian suffixes are highly likely wrong. Thus it had to be put more ambivalently in order not to imply knowledge that doesn't exist (if we set certain scientific standards). G Purevdorj 14:09, 31 August 2006 (UTC)

Altaic Controversy
According to major reference works, everbody agrees on Turkic, Tungus, and Mongolian families. These language families have some common words and they are typologically similar. Whether they are genetically related or not is the subject of ongoing research. The similarities could be result of long-term contact.

All languages are influenced by languages they are in contact with. According to the standards set by linguists, languages that make up a family must show productive-predictive correspondences. The shape of a given word in one language should be predictable from the shape of the corresponding word, or cognate, in another language. Turkic, Tungus, and Mongolian satisfies all these similarities.

I do not agree with the idea that the Altaic controversy necessitates total abandonation of the Altaic.

The family name "Altaic" is a commonly used terminology to label these languages. Turkic, Tungus, and Mongolian are still Altaic regardless of exact status of Altaic is. The dispute is a minor issue among a small circle of specialists. Furthermore, even these specialists use the term Altaic to label, especially, for these three language groups. e104421 11 September 2006, 18:40 (UCT)


 * Three things:
 * It is very hard to assume good faith for your edits when you interpret "take it to the talk page" to mean "force my change through again after it's been reverted by two different people, then go make a post on the talk page".
 * Also please be aware of the three reverts rule if you are not already so. I will not revert you again and tempt you to make a third revert which would get you banned, but someone else may.
 * I consider references to specifically-named scholars and their work (e.g. Starostin, his Etymological Dictionary of the Altaic Languages, and the critical reviews it received) to be more reliable than a generalist reference like Ethnologue.
 * As for the actual issue at hand, I'll leave that to the rest of y'all. cab 01:18, 12 September 2006 (UTC)


 * The first time you deleted the Altaic Languages and the Ethnologue reference, I assumed good faith, reverted it, and summarized in the edit comments the reason for the reversion. When you came back and made the same edit again, with no comment on the talk page, not only about deleting but also Altaic controversy, just erasing the Altaic languages link and also Ethnologue reference, my good faith went out the window. There is already an informative paragraph at the Altaic Languages page about the controversy, that's the reason why i already put that link there in the table. I also aware of the three reverts rule. First you removed the link and the reference without any warning by reverting the article.e104421


 * The page as it was in its original state did not advocate "total abandonation" of the Altaic. (Nor did any other pages you have been editing to advance your pro-Altaic bias). It merely stated the Language family as "Altaic (disputed)", which is perfectly understandable since "A language family is a group of genetically related languages" and the validity of Altaic as a genetic grouping (rather than a sprachbund, areal grouping, or whatever) is disputed in many reliable sources. We do not want to hide the fact that this dispute exists or imply that Altaic as a genetic grouping is undisputed; nor do we want to devote undue space to the controversy, since it's already written about elsewhere on Wikipedia. So the single word "disputed" is in my mind, the best solution, rather than a link to an external source. No one disputes the fact that Mongolian is referred to an "Altaic language". But many people also refer to Persian as a "Middle Eastern language" along with Arabic and Hebrew, or talk about the hanzi-using languages as "CJK(V) languages". That does not make "Middle Eastern language" or "CJK(V) language" a valid genetic classification. cab 11:04, 12 September 2006 (UTC)


 * The main problem is related with the "(disputed)" tag, cause this leads misunderstanding. There exists Altaic Languages but whether they are genetically related or not is the subject of ongoing research. The family name "Altaic" is a commonly used terminology to label these languages. This "(disputed)" tag causes confusion such as "--- language's being Altaic is disputed". For this reason, i'm supporting to remove the "(disputed)" tag. The discussion should be done either in the Altaic Languages page or maybe better in the Altaic hypothesis page. That's why i put the link to Altaic languages in the table. I'm not trying to promote anything just to prevent misunderstanding. e104421


 * The current indexing solution is not the best I could think of, but it is hardly worth disputing, as long as the "(disputed)" remains. As to the previous comment: Of course scientists (including those opposed to Altaic) sometimes use the label "Altaic" to refer to a sprachbund or the like, but they do so with an eye on the ongoing discussion. And if scientists use such a label they are aware of its faults - for those who propose a sprachbund for this area wouldn't limit it to members of Mongolic, Turkic etc - but nonscientists probably aren't. So it is not a question of how to define "Altaic" so that it is a scientifically useful label - if genetically it isn't, it won't be in any other respect as well. G Purevdorj 12:27, 25 September 2006 (UTC)


 * I think "controversial" is better than "disputed", in the light of the Etymological Dictionary of the Altaic Languages by Starostin et al. (2003). See Altaic languages. David Marjanović 17:04, 9 June 2007 (UTC)
 * I just made the change. The previous state of affairs amounted to claiming that Altaic exists because the Ethnologue says so, and that is ridiculous. The evidence for Altaic is very good, but the Ethnologue people haven't done any research on it! David Marjanović 17:10, 9 June 2007 (UTC)


 * Thanks for editing that cruel link. However, as I stated above, the evidence for Altaic is very weak (it's not my fault that the people editing the Wikipedia Altaic page trust in Starostin's dictionary more than any expert does), and so erecting a link is the wrong idea: it must immediately become clear that "Altaic" is "controversial" (or, as I would suggest, "disputed"); that means, it would be a good idea to write just that: "Altaic (disputed)" (within the box). Unfortunately, I don't know how to do that, so if anyone could do this edit? G Purevdorj 23:51, 9 June 2007 (UTC)

History?
There should be at least a section about the development of the language, Old Mongolian etc. 91.148.159.4 15:08, 15 May 2007 (UTC)

Sound system
There are a few oddities in the section on the phonology. For example:
 * What sense does it make to analyze ө and ү as /o/ and /u/ when the sounds [o] and [u] don't even exist in the language? Why not simply call the phonemes /œ/ and ? Or have I missed something?
 * I'd like to see some evidence that в is really [w] and not [v] in any kind of Halh. The three people from the university of Ulaanbaatar I have heard speaking used [v]. Furthermore, while is easy,  is difficult, if not impossible to keep apart from . David Marjanović 17:21, 9 June 2007 (UTC)

It’s not that straightforward to justify any name for a phoneme that can be realized from [w]/[v] to [f], and one should be a little bit cautious before calling this an “oddity”. Karlsson 2005 somewhere notes an example of /w/ that has become [f] in an interconsonantal position. Besides, Svantesson et al. (2005: 20) note correctly that “[w] is less rounded than e.g. English [w]”. I’ve taken the time to listen to a few Mongolian songs. Here’s what I found:


 * Үүрээ, Мөнхбат:: Чамайгаа оллоо (slow): хорвоо [v], Зүрхэнд эгшиглэсэн хонхны дуу (slow): хаврын [w]
 * Сэрчмаа:: Уяхан цэнхэр хавар (slow): хавар [v, w], Архангай аавын минь нутаг (slow): аавын [v, w], Би явлаа (lively): явлаа [v]
 * Чоно:: Айдас (above average speed): хорвоо [w], хэрвээ [w], явнаа (exactly in between), явна [w] (intervocalic: javan), хүсвэл [v]
 * Lumino:: Хүнийх (most of the time rather fast): зөвхөн [w], тайвнаар [v], давжаа ([v], but rather devoiced), лав [w], лавлан [w], зай вэ [w], од вэ [w], (ч)ааваас [w], тавин [w], харвал [v]

To summarize: in intervocalic position, our phoneme is always realized as [w], while as a part of a consonant cluster, it is sometimes [w], sometimes [v], and that may vary for a single word within one idiolect (eg Сэрчмаа). Thus, it is most reasonable to call that phoneme /w/. G Purevdorj 23:09, 9 June 2007 (UTC)


 * It is short initial [ɵ], long initial [o:] and non-initial [ɵ]. /ɵ/ would have been more intuitive than /o/, but whom do you want to follow if not Svantesson et al.? After all, it’s only a question of nomenclature for naming one phoneme with two major allophones.


 * /u/ is [u] (pretty much like in German “Hure”), so what is the matter?


 * [ɔ] as /o/ in some works on Mongolian studies is a traditionalism, reflecting Middle Mongolian. The two lines that compare Svantesson et al.’s analysis to more familiar ones probably don’t add anything but confusion. Best to delete them and to wait for someone to write the phonological part for Middle Mongolian as the only context from which they could easily be understood. G Purevdorj 01:54, 11 June 2007 (UTC)

This section was deleted by me today:
 * "The full inventory of long and short vowels can only occur in word-initial syllables. In word-internal and word-final syllables, vowels are reduced. Long vowels can only appear in initial syllables."

My justification is that the above claim is easily disproven by the following examples: улаан "red", эрээн "speckled", гэргии "wife", хороо "corral", уруул "lip", өртөө "station" and сэрүүн "cool," not to mention the many verb modifiers and verb endings that contain long vowels, witness тэгээд, явуулах, etc. -Ferronier (talk) 13:04, 2 May 2008 (UTC)
 * You cannot argue based on mere orthography. Besides, could you deign to explain that "tusalj" and not "tuslaj", while ? It's primarily the fact that this "a" is not phonemic that explains that orthographically long vowels are merely phonemic ones. Phonetically, this conclusion is supported by phonetic vowel length. First-syllable short vowels: 1. First-syllable long vowels: 2. Second-syllable long vowels: 1,4. Second-syllable short vowels: (if I remember correctly, but it definitely was somewhat to that effect) 0,5. I would have appreciated if you had uttered your comment on the talk page first and then waited a little for someone to react. Linguistic analysis usually isn't that simple. G Purevdorj 14:34, 2 May 2008 (UTC) —Preceding unsigned comment added by G Purevdorj (talk • contribs)


 * Thanks for the explanation, and sorry for precipitately deleting stuff. Point taken.


 * Given the additional information that you provided, a clearer way to word it in the article would be something like "In word-internal and word-final syllables, long vowels are reduced in length, and short vowels are reduced to the point of no longer being phonemic." This would convey the idea that the phonemic distinction between long and short vowels is preserved in word-internal syllables, which omission is the reason the original statement struck me as bizarre. Furthermore, the data you quoted itself indicates that, while second-syllable long vowels are not as long as first-syllable long vowels, they are still longer than first-syllable short vowels; which is definitely my perception on hearing the words that I cited.


 * To make it clear, my analysis was never based on orthography, but on my image of how the words sounded when spoken by the people around me. I actually had to look the spellings up in the dictionary before writing my comment. As a student of everyday modern Mongolian (with a basic knowledge of linguistics), my goal in editing this article is to help make it clearer and more useful to people like me, to native speakers, and maybe even to people who have no special background in either linguistics or Mongolian.  In other words, combining linguistic analysis with communication.  Hope we can cooperate. -Ferronier (talk) 15:38, 5 May 2008 (UTC)


 * My apologies for the sharpness of my last comment. I was totally down that afternoon and made it as biting as might still fit your contribution. Bi baga zereg hetruulchihjee. The latter change of yours is formidable, but there are two problems with it.


 * 1. Svantesson et al. don’t give any information on compounds. As compounding is to some degree productive, this could test whether long vowels in any but the first syllable GET reduced or are not there in the first place. For if this process does not take place during compounding, only suffixes are left, and for these no process of vowel length reduction could be postulated as there would be no independent evidence.
 * 2. However, non-phonemic vowels clearly are no reduced vowels. I’ll present a few examples: Written Mongolian: tusalagarai, tusalaju. Conetemporary Mongolian: туслаарай, тусалж (you can also write туслаж, but you don’t say that). WM: yabuju, yabugarai (Middle Mongolian: yabuarai) CM: явж, яваарай. WM: garqu CM: гарах. That is, non-phonemic vowels are not reduced forms of former short vowels, but get inserted during syllabification independently from the former position of short vowels.


 * I wouldn’t consider it useful to address the first point, and I wouldn’t like to present diachronic data in the article. The discussion of the syllabification rules is too complicated: it would take some effort to present Svantesson et al.s data (62-75), and as part of this is not beyond a doubt, one would have to use some other literature as well, notably Karlsson’s dissertation, a very nice work quite proficient in hiding interesting information where you’d least expect it.


 * Therefore it’s easiest only to state a few simple facts as I’ve tried to do in my last edit. Feel free to reword this as well. —Preceding unsigned comment added by G Purevdorj (talk • contribs) 18:54, 5 May 2008 (UTC)


 * Yes, it was helpful to mention consonant clusters. In that vein, putting in some examples of how the epenthetic vowels move around during syllabification might not be a bad idea.  Even without going into an explanation of the rules, surely there are a few well-attested examples (such as what you gave with forms of туслах) that could be used to give people a more concrete idea of what's going on. -Ferronier (talk) 07:29, 10 May 2008 (UTC)

Regarding my last edit to the structure of the "Sounds" section, there are some questions I have before proceeding to the next step. So, the section on vowels is now in four paragraphs instead of two. 1. Long and short vowels. 2. Epenthetic vowels. 3. Vowel groups. 4. Implementation of vowel harmony. My questions relate to the second and fourth paragraphs.

First, just how is the form of the epenthetic vowel determined? The information in the article is incomplete, so here I go with my direct observations again. Would it be correct to say the following?
 * /a/, /e/, /ɔ/, and /o/ produce centralized versions of themselves.
 * /ʊ/ produces /a/ and /u/ produces /e/.
 * /i/ produces a (centralized?) /i/ unless there is another more dominant vowel in an earlier syllable.
 * a postalveolar or palatalized consonant is followed by an epenthetic /i/ regardless of the preceding vowel.

Second, if the above is correct, then the rules for deriving an epenthetic vowel from a dominant vowel are almost the same as the rules governing the vowels in suffixes, right? In that case, couldn't we explain the rules in the vowel harmony paragraph, and briefly make reference to them in the epenthetic vowel paragraph? It seems to me that that would be placing the emphasis where it belongs. -Ferronier (talk) 07:29, 10 May 2008 (UTC)


 * /ʊ/ also produces a centralized version of itself. “by an epenthetic [i]”!


 * So to speak, a suffix like, say, /sN/ (past) doesn’t contain any vowel, while a vowel may well get inserted: явсан /jawsN/ [jawsăŋ] явсны /jawsNi/ [jawsnı]. So the rules are the same, and it should be quite possible to proceed as you suggested. Of course, this doesn’t extend to full non-initial vowels in suffixes, eg causative /Uɮ/. —Preceding unsigned comment added by G Purevdorj (talk • contribs) 09:03, 10 May 2008 (UTC)

I have noticed that the transliteration for the Cyrillic alphabet here at Omniglot differs from that given in the article, but both claim to be for Khalkha - in particular, for the Cyrillic B and P. Why is this? --Baryonic Being (talk) 12:33, 11 May 2008 (UTC)
 * There’s always been an awful lot of transcriptions around, no way to help that. The link you’ve given gives pretty uncommon transcriptions for ж and з, is imprecise for й (as is the Wikipedia transcription for it – yes, we employ a transliteration for this article and transcriptions for all others, look at Naming_conventions_%28Mongolian%29) and is extremely liberal for н (never seen that before myself). For в, „v“ is simply more common, but „w“ would possibly be more accurate. What about P, by the way? It’s identical, isn’t it? —Preceding unsigned comment added by G Purevdorj (talk • contribs) 18:03, 11 May 2008 (UTC)

Rating
I feel that this article should be in for some upgrading. It seems quite informative, concise, well-sourced and covering most relevant areas (I could think of a hundred more areas, but as this is just a lexicon many relevant subpoints just can't be entered. Think of all the dialects and would-be-dialect languages). If not, this might be a place for those doing the rating to utter their objections to the current status of this article. G Purevdorj (talk) 16:17, 22 August 2008 (UTC)

Examples needed
"While the split of ʧ into ʧ and ʦ in Outer Mongolia and its lack in Inner Mongolia is often cited as a criterion of importance, the split between -sŋ in the Central dialect vs. -ʤɛ: in the Eastern dialect..." Please support discussions like this with examples. A good model is the article Russian Language where every such case is supported by samples words. Gantuya eng (talk) 10:49, 30 August 2008 (UTC)
 * First, I will supply the additional quotations as soon as I'm home again (might take up to a week, but I'm lacking most literature at the moment). Then, the first point made in the quoted sentence can (and will) easily be supported with an example (say, zayaa and jargal vs. jayaa and jargal), but how to illustrate the second point? It is of course possible to demonstrate that a subset of sentences ending in -jee will be rejected by Khalkha but not by Khorchin speakers, and that a set of Khalkha sentences with -sn is unacceptable to Khorchin informants. But that would be original research. Inner Mongolian dialect studies don't employ negative evidence. The study by Ashimura and to a very limited degree a study by Matsuoka contain well-founded claims on restrictions on the use of -sn, Ashimura contains information on the use of -lee and -jee. Both studies could (in a very careful description that differentiates within modern Khorchin territory) be related by general (and therefore not necessarily reliable) information provided by Bayancogtu. The stochastical claims made for various Inner Mongolian dialects can be referenced (to some extent; I'm not in Inner Mongolia at the time being). But as far as I'm aware, that's all I might be able to improve on the second point. Would you consider that sufficient? G Purevdorj (talk) 11:40, 30 August 2008 (UTC)


 * By sentences ending with -jee: do you mean the Inner Mongolian phenomenon that they say "Bi irjee, Bi yavjee" instead of "Bi irsen. Bi yavsan" ? Gantuya eng (talk) 12:34, 30 August 2008 (UTC)
 * Yes (but by the way, if your example is not an over-generalization, even the analysis of Ashimura can't explain a frequent usage of -jee with the first person singular). But yes ... if that has not already become clear to you from the text, maybe there is a way to put it more clearly. G Purevdorj (talk) 12:58, 30 August 2008 (UTC)


 * Bayarlalaa. Gantuya eng (talk) 14:54, 30 August 2008 (UTC)

Footnote after full stop
Probably it's like this: if the footnote belongs to the whole sentence, it's put after the full stop. If the footnote belongs only to the last word of the sentence, then it's put after that word but in front of the full stop. Is it right? Gantuya eng (talk) 05:43, 4 November 2008 (UTC)
 * Well, I'm used to putting it before the full stop when it belongs only to this sentence, but not to the preceding sentences, while I would put it behind the full stop if it belongs to the paragraph. (But determining the exact scope of references becomes a problem with almost any scientific literature.) G Purevdorj (talk) 08:17, 4 November 2008 (UTC)


 * That's right. We don't mark the beginning of the scope. :) Gantuya eng (talk) 08:40, 4 November 2008 (UTC)

1941 or 1931
This article contradicts Mongolian Latin script and others, by saying that cyrillic was introduced in 1937, whereas others say that latin was introduced in 1941, after which cyrillic was introduced. This is a pretty serious mistake. So which one is it? --Chinneebmy talk 15:50, 19 November 2008 (UTC)
 * Svantesson et al. have 1930-1932 for Latin and 1941 for Cyrillic. They don't comment on what happened in the meantime. The formulation in this article doesn't contradict this: there would have been a phase of experimentation with the script from 1937 onwards followed by its becoming the only standard in 1941. (1941 is mentioned anywhere.) Ridiculously, I was unable to find any more detailled information now. But if you happen to be in the more or less fortunate situation to be in Mongolia, you can access all the literature you could possibly need to find the answer to this problem. Anyway, without more definite sources, I would suggest not to make changes in THIS article. G Purevdorj (talk) 20:10, 19 November 2008 (UTC)


 * Which sources say "latin was introduced in 1941"? :) Gantuya eng (talk) 03:13, 20 November 2008 (UTC)

some comments
I guess this was a lot of work. Some more thoughts: The number of speakers in China (3 million) seems unsourced. The official number of Mongols is considerably higher (around 5 million?), but not all of them soeak Mongolian. I guess it is difficult to determine how many actually do, but it would be nice to see where the 3 million number is from. Why not 2 million? Or 3.5? I also think some parts may be a bit ambiguous for the general reader. I noticed and corrected one example in the writing systems section, but maybe there are more. What is the source for Mongolian speakers in Kyrgyzstan? Another WP article points to Kalmyk speakers near Karakol, but if we include Kalmyk as a Mongolian language, we probably should mention Xinjiang and Kalmykia in the lead section too, not just Inner Mongolia and Russian Far East (btw. does Russian Far East include Buryatia?).

I have deleted most writing systems in the infobox. Cyrillic was way too far behind, and Soyombo and Phagspa just seemed not succesful enough to mention. Maybe they could be added again with some qualifier. Latin is obviously used a lot, but quite informally and usually only on computers and cell phones, so I don't know whether this fulfills the threshold for inclusion in the infobox, especially since this is not mentioned in the article. Maybe some words on how many Outer Mongolians can really read Mongol Bichig and how many Inner Mongolians can read cyrillic would also be nice, though I guess there are no studies on this. Yaan (talk) 15:37, 24 November 2008 (UTC)


 * Right, didn't review that so far. There is no proper source for the number of Mongolian speakers in China. Take the people of Mongolian nationality in China and divide them by two or multiply with 2/5. I'm not aware of any actual "educated guessses", and as we can't trust in any thematic studies published in China (there might be some studies written in Korean that might be forthcoming soon, but no other usable data at all), I can't blame anyone for not trying. Both 2 or 3.5 millions could actually be true. Then, I'm not aware of any speakers in Kyrgyzstan, and my standard sources don't mention Mongolians in Kazakhstan and Buryatia. In general, it tends to be the other way round (Kazhakhs, Oirats and Buryats on Mongolian soil), but I wouldn't be surprised if there were some considerable minorities (more than 1/1000 of the population) in these countries. Anyway, no sources no mentioning.


 * We (or better: I) (rather reluctantly) made a conscious choice not to include Buryats and Kalmyks in the wikipedia definition of "Mongolian", so these groups won't count for areal distribution.


 * Scripts ... Yes ... Not many sources on Latin, it's a pity. Highly educated Outer Mongolians can generally read and write Mongolian script, but only a subset of highly educated Inner Mongolians can read, let alone write Cyrillic. No statistical data either, I fear. G Purevdorj (talk) 18:11, 24 November 2008 (UTC)


 * I think I have actually seen some statistics on language in one part (banner? sum?) of Inner Mongolia, but I don't really remember where, and it also was some years old already. And then extrapolating from such a kind of sample is not meaningful at all. But I could try to locate that data if you need it. What's the problem with Chinese studies, though, do they make up stuff, are they just unreliable? Yaan (talk) 14:32, 25 November 2008 (UTC)


 * I wouldn't mind taking a look at the statistics you mentioned if you could locate it. As for the data from China: the local government in China has an interest in getting relatively high data for native speakers. I've seen a few studies with all extremely high (>85%) percentages for frequent Mongolian language use that do not correlate well with the general impresson or the data from Yu. (I don't read Korean, so I only remember it from a paper he presented at SIAC 18.) I didn't look into the details of these studies then, but it were usually questionnaires on usage. One fault of this method is that people claim to possess (and use) a language they identify with, but don't do so in actual fact. The next and more crucial point might be that areas that were reasonably estimated to produce favourable results were in advance chosen over other areas. I don't remember much about the sampling technique, but here are risks as well. So I would suspect that the data is either unreliable or not representative. These are just general guesses, of course, but there is no reason to assume good faith if the producer of the data has vested interests. (I would love to take a deeper look into this data myself, but that would indeed take lots of time of me and friends of mine that should better be used on other problems now.) G Purevdorj (talk) 15:06, 25 November 2008 (UTC)


 * Soyombo script might not have been successful enough to be in the infobox. But if Phagspa was used for an entire epoch of the Yuan Empire, isn't that an enough success? Gantuya eng (talk) 22:19, 24 November 2008 (UTC)


 * I strongly oppose including only one script. Both or none. But if they should be included in the info box, there must be an amount of information on them in the article that addresses more than their mere existence at some time in the past. G Purevdorj (talk) 23:16, 24 November 2008 (UTC)


 * Soyombo script--its birga has been very successful... Gantuya eng (talk) 01:23, 25 November 2008 (UTC)


 * I just gave a second thought to my comment from yesterday. Phagspa was only used for writing Middle Mongolian, which is closer to "Common Mongolic" than to "Modern Mongolian" or "Mongolian" and should not be included into an article on "Mongolian proper". It is mentioned in the article anyway, in the section on language history. That's were Soyombo might be put as well, as related to "Classical Mongolian". Neither script should be put in the info box. (I'm just in the course of making some unsourced emergency fixes to Middle Mongolian, but if this article should grow and become at least c-class, including Phagspa into ITS infobox would make sense.) G Purevdorj (talk) 08:31, 25 November 2008 (UTC)


 * I just wanted to write some arguments about modern vs. historical use, but I see that you already said it all much better than I could. In other words, I completely agree with Purevdorj's reasoning here. --Latebird (talk) 21:34, 25 November 2008 (UTC)

"History/Historical"
I accidently pressed "enter" before finishing my edit summary, so I'll explain it here instead. The preferred naming of history section in WP:LANG is "History", and since language history are mostly based on written sources, "historical" doesn't really make it clearer. However, I proposed "Historical changes" as the main section heading instead, since this also made it allowed for a shortening of the sub-headings.

Peter Isotalo 17:14, 9 December 2008 (UTC)
 * Language history is actually only partially based on literary sources as there is a great deal to be understood and described through the science of both internal and comparative reconstruction. The usage of the term "History" in a language article is appropriate only when it includes both preliterate reconstruction and postliterate description. (Taivo (talk) 18:48, 9 December 2008 (UTC))
 * Agree with Taivo. Compare the article on Mongolic languages where the term "History" IS in use. Actually, the Historical changes can't be described before this variety of Mongolic isn't first introduced. So if you read "Historical changes", you think "Eh? From where to what?". So first comes a little language history, but as this only starts with "Historical Mongolian", "History" isn't the proper term either. G Purevdorj (talk) 20:19, 9 December 2008 (UTC)
 * Other than how it's used in Mongolic languages, whare you basing the quite separate definitions of "history" and "historical". I'm a history student myself, and I've never heard of this rather sharp distinction. If you describe past phases of a language of any form, you're discussing history. "Historical Mongolian" sounds more like you're trying to describe the role Mongolian has had in history overall. And "History" is still the most common standard for languages. All language FAs use it, and it's still what's recommended at WikiProject Languages/Template.
 * Peter Isotalo 07:06, 10 December 2008 (UTC)
 * "History" sounds like undivided reference, doesn't it? All the history ... And moreover, it is worthwhile to clarify that the history of the "Mongolian language" is not itself part of a phase that could be part of this "Mongolian language", as wouldn't hold for Common Germanic and English either. "History" would tacitly suggest such an identity. That is not to say that such a suggestion couldn't be cancelled, as is done in Swedish language. Thus it would be your burden either to slightly rewrite the text or to point out why "Historical Mongolian" should be such a confusing description to use that it should be abandoned in favour of "History". (I don't see any advantage in confirming to an existing standard here, but there might be some other good reason.) G Purevdorj (talk) 09:41, 10 December 2008 (UTC)
 * Peter, the linguists here are very clear about the distinction between "History" (which includes both the documented and reconstructed history of a language) and "Historical" (which includes only the documented history of a language). We use this distinction all the time and it is not the distinction which you cite.  "Historical Mongolian" simply means that period of Mongolian that is found in the literary sources.  We could easily say "Historical English" meaning that period of the history of English which is documented after the year 700 C.E.  The "History of English", however, extends all the back to Proto-Indo-European about 6000 years ago.  Clearly the first 4500 years of the history of English is before writing was adopted and there are physical records--it is reconstructed and not "historical".  It doesn't matter what the distinction is in the field of history, this is a linguistics article and the usage of linguists is primary here. (Taivo (talk) 14:03, 10 December 2008 (UTC))

Stress, intonation
Which analysis do you have? For intonation I only know of the analysis of Svantesson et al. which is entirely based on some elicited data and thus possibly artificial. If there's any analysis based on actual spoken language, I'd be glad indeed if it could be added. Stress, then, is a very disputed area and you would need to make use of some quantitative analysis in order to add that part. G Purevdorj (talk) 23:13, 1 January 2009 (UTC)
 * Why the additional references (Bosson, Poppe etc.)? That stuff has not been cited in the article yet, has it? G Purevdorj (talk) 00:02, 2 January 2009 (UTC)


 * I know nothing of intonation. The stress I refer to is word stress not phrasal stress or contrastive focus stress. Walker's data is also elicited although the tokens are in carrier phrases (that doesn't always work in actuality but Walker doesnt mention any problem with her informants).


 * The additional refs are just that. I renamed biblio accordingly. If someone wants to learn about Mongolian, then they can go to the primary material for real research. A single biblio that consolidates all primary sources is useful for readers. (Some people publish bibliographies for this reason, right?) I assume that you are into Mongolian and will want to give those a look. Maybe others are too. And, perhaps, I will gently suggest that they should be consulted if the article is to be of sufficient quality. If the work is not used to write the article, then just say that (or imply it by not citing it in the main text). Excluding source information because it was not used in the article is a bad idea. I realize that some folks like to omit sources (and I'd say that they are silly). But, I'm not going to fight anyone about this. peace. – ishwar  (speak)  00:53, 2 January 2009 (UTC)


 * By the way, here's a paper on harmony if youre interested: www.ai.mit.edu/projects/dm/featgeom/vdHSmith-KhalkhaBuriat-87.pdf. – ishwar  (speak)  01:06, 2 January 2009 (UTC)


 * This article originally had an additional list of sources, but the dictionary part was difficult to keep up-to-date and the grammar part was problematic without a proper discussion of their differences which requires exceedingly careful writing in order not to become original research. But an open list on literature (as you seem to suggest) seems impossible to handle: there are several thousands of works and articles on Mongolian. Maybe one could add a list of bibliographies ... Besides, eg Street's Khalkha structure is completely out of date 1. he uses a theory of grammar that according to common consence fails to represent it according to language-internal principles. 2. His interpretations are quite limited (alas, but this holds for all general grammars that are available). Bosson is pedagocial literature, thus to be excluded. (Again, an extra list for such stuff might be desirable, but it would be wrong to cite information from it or to mix it with proper linguistic literature.) But I must see to this later, I'm in a bit of a hurry right now. G Purevdorj (talk) 10:43, 2 January 2009 (UTC)


 * Having an out of date theoretical framework (what is it American structuralist?) does not mean that a description is without insight. That's why Noam Chomsky and other generative syntaticians look at Otto Jespersen's analysis (1910s) of English. Also many phonologists have looked at Stanley Newman's American structuralist analysis (1940s) of Yawelmani to mine for data to support analyses in SPE ordered-rules, underspecification theory, and Optimality Theory.


 * And as Walker mentions, Street (based on Poppe 1951) is the work that most theoretical literature uses as a source of data. That would seem to warrant mention here. And, significantly, Bosson was important because it got the stress facts correct where Street was erroneous. If Bosson is pedagogical, then that in itself would seem to warrant inclusion. If a reader was interested in learning Mongolian, they could then have a place to start presumably better than a phrasebook. – ishwar  (speak)  21:21, 2 January 2009 (UTC)


 * "And as Walker mentions, Street (based on Poppe 1951) is the work that most theoretical literature uses as a source of data." True enough, one reason why part of these writings suffer from medicore data. Take Binnick' writings on TAM in Mongolian. It is often cited because 1. it employs a transparent terminology 2. is written in English 3. is readily available. That most likely does not warrant mention. We're not talking about the history of research into Mongolian (an extraordinarily interesting, but completely apart topic), but about the contemporary knowledge that exists on this language, and this knowledge is usually either available from Mongolian language sources or from specialized research essays. But more crucially, we simply cannot include any more or less insightful work without making use of it. As for Bosson, you can easily include it by writing some stuff like "Walker 1997 (based on Bosson 1966)" which would be perfectly legitimate. A special section on pedagical works on Mongolian within this article doesn't seem desirable to me, but one might contemplate a list article that functions as an additional bibliography. Anyway, as I cannot conceive how to select the literature that should be put there, I am not sure that this would be a good idea either. You might add some reference for ATR, by the way, and this has to be related to "pharyngeal", a term that is presently still used in the article. G Purevdorj (talk) 00:22, 3 January 2009 (UTC)


 * You can cite Svantesson et al. for [ATR] or van der Hulst & Smith or John Goldsmith's phonology handbook.


 * Why would you not want to have a history of research section? That seems to be relevant to the topic. You dont need to mention everything.


 * " we simply cannot include any more or less insightful work without making use of it". As I've already said, I disagree with these practices.


 * Assuming that some people are interested in learning Mongolian, mention of pedagogical sources would benefit those people. – ishwar  (speak)  01:23, 5 January 2009 (UTC)


 * A "history of research" section might be interesting (as long as there isn't too much overlap with the material in Altaic languages). But it is not the purpose of Wikipedia to help people learn a language (see WP:HOWTO), so linking to learning resources is outside its scope as well. --Latebird (talk) 10:19, 5 January 2009 (UTC)


 * Your link does not provide the information that you assert it provides. Maybe you intended to give another link? G Purevdorj (talk) 11:11, 5 January 2009 (UTC)


 * Just as a side note: I removed the section about stress from Buryat language again, leaving the notice in place that it is the same as in Khalka. Duplicating such information verbatim is not a good idea (maintenance nightmare). It also made the article look extremely unbalanced, because no other formal information is present there yet. --Latebird (talk) 12:16, 2 January 2009 (UTC)


 * I guess it's more appropriated to respond there? – ishwar  (speak)  21:21, 2 January 2009 (UTC)

Dialects
I guess what is not clear is whether this applies to Mongolian or Chinese speakers or both: "There is officially a standard pronunciation “based” on the Chakhar dialect of the Plain Blue Banner. A standard grammar and vocabulary is from a so-called “Inner Mongolian dialect” which is contrasted to the Oirat and Barghu-Buryat dialect." Regards, &mdash; Mattisse (Talk) 01:33, 3 January 2009 (UTC)
 * As it is in a text on Mongolian and as Oirat and Buryat are Mongolic as well, I presumed that the passage is clear. But what do the others think? G Purevdorj (talk) 22:32, 3 January 2009 (UTC)

Drydock Stress
I have just considered the data presented in the article and referenced to Walker and found that 1. it doesn't confirm to the phonological analysis presented earlier in this article. With some time, this can easily be fixed, but such a contradiction should not be present in the article for several days. 2. The analysis of several words such as baiguullagaar, ah, unshsan doesn't confirm to the syllabification constraints. I'm aware that there is an old analysis of Khalkh grammarians that assumes the actual presence of CV in non-first syllables. This analysis did hold around 1900 (see Ramstedt 1903), but is obsolete by now. (Using some TV interviews from the sixties, one could see whether Bosson could actually have observed this in the 60s. But I must admit that I'm always a bit suspicious in such matters.) Anyway, it is unlikely that Walker could have observed this in the 1990s. I'll have to look into the quality of Walker's evidence later (I'll obtain Bosson 1966 from the library next week). At the least, this contradiction has to be dealt with in some way. As long as it is not, I suggest to leave the text in the drydock. G Purevdorj (talk) 14:49, 4 January 2009 (UTC)


 * Obviously, there are different analyses of Mongolian. Rather than present one you can present more than one. This is especially important if other linguists use the unmentioned analysis as a basis for theory (whether or not they should do so is not our problem). Svantesson et al. present their analysis of V length, but they also mention other analyses. Wikipedia should do the same. If you feel its too messy for the main article, you can focus on one analysis and mention the analyses here and present all analyses in detail at Mongolian phonology. One or all analyses may not be successful, but that is not for you or me to decide: just present what others have said.


 * You said above that Svantesson finds that Mongolian has no stress. We should mention that as well. (I only have access to Svantesson et al.'s first several pages & my library doesnt have the book, so I havent read what they say about it). But, Walker's acoustic analysis does find stress. Someone may be wrong or both wrong or both right. It may be that Svantesson et al. cannot predict stress within their analysis while Walker can.


 * Having no CV syllables noninitially as you mention seems to be factually wrong. In Svantesson et al.'s terms, short Vs and reduced Vs are found after the initial syllable — both are presumably CV.


 * You seem to misunderstand one aspect of phonological analysis: just because something occurs in a surface pronunciation phonetically does not mean the analysis will not posit an underlying representation that differs from the surface form. So it's not a question whether VVs occur noninitially in the 60s in speakers' utterances, it's a question of whether the analysis will posit underlying VVs or underlying Vs. The vowels could be underlying short or underlying long with subsequent shortening via phonological process.


 * Walker is not based on Bosson 1966. She elicits her own data which she recorded and analyzed acoustically and has personal communication with Bosson in the 90s. Give it another read.


 * Why are you saying that some analyses are obsolete? Is this the opinion of the sources? Is it your own personal evaluation? You are not supposed to be evaluating the quality of a researcher's evidence. You are just supposed to present the analyses and the arguments for and against them. – ishwar  (speak)  01:15, 5 January 2009 (UTC)


 * A review of Svantesson et al. does not give the analysis of lexical stress the best evaluation:


 * "Chapter 7: Prosody


 * This chapter, complete with waveform and spectrogram information, details the focal accent, boundary signaling, and word stress. It also examines the existence of a 'final prominence tone' present in many Mongolian utterance. Perhaps the most interesting aspect of the chapter is the summary discussion of prevailing and contrastive views on Mongolian word stress. The authors provide a fairly succinct account of varying analyses. Of interest however, is the authors' conclusion that no unified phenomenon of word stress exists in the language. One would like here to see concrete evidence and a clear exploration of the discrepancies which could support such a statement."


 * – ishwar  (speak)  01:55, 5 January 2009 (UTC)

1. I am in agreement with you that the information on stress provided by Svantesson et al. 2005 is unsatisfactory. I didn’t think so from the outset, but I recently read it again and also came to this conclusion. So Walker should be included. I also agree that Svantesson et al.’s differing analysis should be mentioned.

2. The phoneme set used in the article should be unified by all means. It is rather [ae], but you cannot use [ae] and [ai] (as Svantesson et al. have) in the same article. This can be done easily and I’m gonna do it, but it will take some time. So, I am only going to do it as soon as we have made up our mind on how to deal with the differing analysis of CV. That is why I took the data from the main page.

3. I have not found the time to read Walker 1997 yet, and how I understand it at the moment, she used data from a single informant (which would be somewhat unsatisfactory) that confirmed to her the analysis of Bosson 1966 which she knew beforehand. I shall read both until Thursday.

4. We have non-initial CVC(C) syllables all right; we only strive to avoid open non-initial syllables. I in particular object to the following items:


 * bai.ˈɡuːl.la.ɡaːr "by means of the organization"
 * u.laːn.ˈbaː.ta.raːs "Ulaanbaatar (ablative)"
 * ˈa.xa "brother"
 * ˈun.ʃi.san "having read"

If we follow Svantesson et al. in stating that the formerly short phonemes that were present in non-initial syllables have disappeared from the phoneme template of lexical items


 * (which I would hold is fairly evident: yabuju (Middle Mongolian) > yavaj(i) (Ramstedt 1903: 46) > yavj (today), ükülüge (MM) > *ükülee > uxxelee (Ramstedt 1903: 16) > uxlee (today), but ükülejü (MM) > uxelj (today); I have not used the exact letters (j is dsch here), but the syllable structure I’ve left intact. I would thus follow Svantesson in ruling out “underlying short [vowels] with subsequent shortening via phonological process” on that basis)

then there can’t be any underlying short V-s in non-initial syllables. So the phonemic representation of the above items would be


 * /baigullG-ar/
 * /ulanbaatr-as/ (compound from /ulan/ and /baatr/)
 * /ax/
 * /unsh-sn/

The three latter items are straightforward: We get


 * [ulanbaatras]
 * [ax]
 * [unshsən]

The first one is problematic. Svantesson et al. would indeed have it realised as


 * [baigulləGar]

(I just corrected this in the main article) but I remember an analysis on another dialect that would have it [baigulləGGar], and that’s how I would pronounce it. Anyway, as this analysis hasn’t entered our article yet (and as far as I know, it cannot be accessed in any European library), it’s [baigulləGar].

5. I am saying that some analyses are obsolete because they don’t confirm with the readily observable data, but obviously (Ramstedt 1903) did confirm with that data once. I am indeed not supposed to evaluate readily observable linguistic data on my own here, but I do hold that in the presence of evidence of superior quality presented in one work, evidence presented in other works should be discarded. That is, I may well evaluate the quality of two sources against each other. Here, I conclude that Svantesson et al. 2005 will most likely have presented better evidence than Walker 1997 as far as syllabification is concerned. I agree that both approaches should be presented in an article on Mongolian phonology, but I don’t intend to start that article. (Mind you, such an article could not be restricted to Khalkh Mongolian phonology.)

G Purevdorj (talk) 11:36, 5 January 2009 (UTC)


 * I'm back. I still dont think that you should be evaluating anything. That's usually called "original research" in wikipedia. If there's a discrepancy, then the discrepancy can be noted. The evaluation should be left to a published comparison.


 * I dont know the details the general Western analysis. (I trust you will learn it if you dont already.) The major problem I see with the above words is the extra vowels vs missing vowels:




 * ulaan-'baataraas || ulan-baatras (< ulan-baatr-as)
 * 'axa || ax
 * 'unʃisan || unʃsən (< unʃsn)
 * }
 * 'unʃisan || unʃsən (< unʃsn)
 * }


 * Basically the eastern data lacks short Vs that occur in the western data. If the western analysis wants to derive the eastern data, it could posit a V deletion process. I dont know why Walker has this discrepancy in her data. As you suggest, maybe it comes from the older Mongolian recorded 100 yrs ago. Or maybe it's a different dialect. At any rate, we dont how Walker will account for the missing Vs in the data you cite. Presumably, stress assignment will be a problem with this data, and I guess that Walker will need to assume an underlying phonological representation that is more abstract than Svantesson et al's more surface-ish analysis.


 * This issue with long vs short seems trivial. If I understand this, there is the following correspondence:


 * {| class="wikitable"

! !! Western !! Eastern ! rowspan="2" | Initial ! rowspan="2" | Post-initial
 * VV || VV
 * V || V
 * V || V
 * VV || V
 * V || ə
 * }
 * }


 * Basically, this is a difference in theoretical perspective. They both account for the data. The western analysis has to have some account for why the initial syllable is lengthened for both VV & V (or alternately shortened post-initially); the eastern analysis has to have an account for why the VV-V contrast is restricted to initial syllables.


 * The problem for the eastern analysis is deriving the correct stress from [baigulləGar]. This is presumably CVV.CVC.CV.CVC. Walker's stress algarithm would give initial stress for this word, which is incorrect (according to Walker, of course). If it's CVV.CVVC.CV.CVVC, then the correct stress is derived.


 * Our question is what do we present in the article? – ishwar  (speak)  23:51, 7 January 2009 (UTC)

Yep, the relation of both analyses is trivial enough and can be translated in the way you have suggested. But the reformulation of Walker 1997 in the model of Svantesson et al. 2005 would be:


 * Primary stress falls on the penultimate phonemic vowel.
 * If the first syllable contains a short vowel and there is no other phonemic syllable apart from the word-final syllable, the word-final syllable gets the primary stress.
 * If there is only one syllable containing a phonemic vowel, this syllable gets the primary stress.
 * All syllables with full or long vowels and possibly every word-initial syllable get a secondary stress.

I originally intended this to be a simple translation, but then I noticed that it of course fails to reflect her generalization over long vowels. Yet, one might indeed separately define syllables containing full (non-initial phonemic) and long vowels as heavy and then proceed. Light syllables would then get a sub-classification into phonemic and non-phonemic ones in order to clarify her rule (27 iii). I think one could present it in this fashion.

Second, the translation into Svantesson et al.’s terminology basically identifies her secondary stress with vowel length. Remember that she defined stress “in terms of pitch, duration, and intensity”? Here, duration alone would be sufficient. Thus, I would tentatively hold consider this (minor) part of her analysis insufficient and refrain from presenting it.

An additional footnote would specify the type of her evidence.

Would that be a way to go?

(By the way, if you’re in the US, you could maybe easily take a look into


 * Walker 1995: Mongolian Stress: Typological Implications for non-finality in unbounded Systems. In: Phonology at Santa Cruz 4: 85-102.

It might be that this paper contains some data that provides “pitch, duration, and intensity” in the syllables in question in a quantitative way.)

G Purevdorj (talk) 00:38, 8 January 2009 (UTC)

PS: 1. Original research is the overt input of non-trivial information that hasn't been published so far. I've not been proposing to do such a thing. You might argue that selection of literature is POV (a label that I wouldn't accept either), but never "Original research". 2. I've added some information on phonetic vowel length in order to justify the proposed line of argument. G Purevdorj (talk) 10:12, 8 January 2009 (UTC)

In Khalkha, heavy (H) syllables have either a long vowel or a diphthong. CV and CVC syllables are light.

Stress
Primary stress. In the following Khalkha Mongolian words, primary stress generally falls on the rightmost heavy syllable (Walker 1997 using data from Bosson 1966):




 * HˈHLL || || "to be organized"
 * LHˈHL || || "to separate (modal)"
 * LHHˈHL || || "the residents of Ulaanbaatar"
 * }
 * LHHˈHL || || "the residents of Ulaanbaatar"
 * }

However, there is a restriction that the stress cannot fall on the word-final heavy syllable. Thus, stress falls on the rightmost nonfinal heavy syllable.




 * ˈHH || || "dry cheese curds"
 * HˈHH || || "angrily"
 * ˈHLH || || "sad"
 * LˈHH || || "by means of his own horse"
 * LˈHLH || || "seventh"
 * LHˈHH || || "by one's own sea"
 * HˈHLH || || "by means of the organization"
 * LHˈHLH || || "Ulaanbaatar (ablative)"
 * }
 * LHˈHH || || "by one's own sea"
 * HˈHLH || || "by means of the organization"
 * LHˈHLH || || "Ulaanbaatar (ablative)"
 * }
 * LHˈHLH || || "Ulaanbaatar (ablative)"
 * }
 * }

The nonfinal restriction does not apply when the only heavy syllable is word-final:




 * LˈH || || "goose"
 * }

In words without heavy syllables, Khalkha shows a "default-to-opposite" effect. Here, the stress falls on the initial syllable:




 * ˈLL || || "brother"
 * ˈLLL || || "having read"
 * }
 * }

Secondary stress. Heavy syllables without primary stress receive secondary stress:




 * ˌHˈHLL || || "to be organized"
 * ˈHˌH || || "dry cheese curds"
 * ˌHˈHˌH || || "angrily"
 * ˈHLˌH || || "sad"
 * ˌHˈHLˌH || || "by means of the organization"
 * }
 * ˈHLˌH || || "sad"
 * ˌHˈHLˌH || || "by means of the organization"
 * }
 * }

There is also variability in secondary stress on initial syllables.




 * LˌHˈHL || or || ˌLˌHˈHL
 * LˌHˌHˈHL || or || ˌLˌHˌHˈHL
 * LˈHˌH || or || ˌLˈHˌH
 * LˈHLˌH || or || ˌLˈHLˌH
 * LˌHˈHˌH || or || ˌLˌHˈHˌH
 * LˌHˈHLˌH || or || ˌLˌHˈHLˌH
 * LˈH || or || ˌLˈH
 * }
 * LˌHˈHˌH || or || ˌLˌHˈHˌH
 * LˌHˈHLˌH || or || ˌLˌHˈHLˌH
 * LˈH || or || ˌLˈH
 * }
 * LˈH || or || ˌLˈH
 * }

The initial syllable effect is generally a fall in fundamental frequency (F0) with an increase in vowel length. However, some speakers show a rise in fundamental frequency when the initial syllable is heavy. Thus, this initial syllable effect may not be secondary stress but rather some type of word edge effect.

Buryat has the same stress pattern as Khalkha.


 * Bosson, James E. (1964). Modern Mongolian. Uralic and Altaic series (No. 38). Bloomington: Indiana University.
 * Walker, Rachel. (1997). Mongolian stress, licensing, and factorial typology. (Online on the Rutgers Optimality Archive website: roa.rutgers.edu/view.php3?id=184.)

Language regulation
Will the Law on State Official Language of 2005 (Төрийн албан ёсны хэлний тухай хууль) qualify as regulation for the language in the Infobox? Gantuya eng (talk) 14:53, 4 January 2009 (UTC)
 * Certainly. The sections 5.16, 5.17, 5.18, 6.6 (with the possible exception of 6.6.4 and nonwithstanding 6.6.5), 7., 8.1.2 and 8.1.3 constitute devices to actively influence the development of Official Halh Mongolian. There's the prescription of terminology, obligatory use of this terminology and literary language by officials (I guess this includes school teachers), and some penal law to enforce it. There is also a part that pertains to normal citizens (7., 8.1.3), but I don't understand under what circumstances normal citizens could violate 7. If you have a clear idea of it, could you explain it? Anyway, the existence of 5.16, 5.17, 5.18, 6.6 alone would constitute language regulation. So now how to enter this information into our article on Mongolian language? G Purevdorj (talk) 15:43, 4 January 2009 (UTC)


 * Sorry, I don't know too, how Provision 7 could be violated by citizens. There's penalty for citizens for violation in 8.1.3 though :)
 * The Infobox states: "No official regulation". I wanted to insert the name of the Law instead of it, but couldn't understand how to edit the Infobox. Gantuya eng (talk) 16:07, 4 January 2009 (UTC)


 * I've reformatted the infobox code so that it is easier to maintain (there's no reason to cramp it together to save space). The relevant parameter for the "Regulated by" entry would be "agency=". However, it seems to be targeted at a regulatory body or an academy who might have the task to provide such regulation. Strictly speaking, that would probably be the Great Khural in this case, but I guess we could still write the name of the law (in English) there. Do we have any information on the regulatory situation in China? --Latebird (talk) 17:51, 4 January 2009 (UTC)


 * The institution in question is the Төрийн хэлний зөвлөл (see 6.1), maybe "State Language Council". Its decisions have to be confirmed not by parliament, but by the government (6.3), so the language regulation is conducted by these two institutions in cooperation. G Purevdorj (talk) 21:30, 4 January 2009 (UTC)
 * In South Mongolia, we have the Kele bicig-ün ajil-ün jöblel, but there are other institutions and I am presently not sure about their hierarchy. G Purevdorj (talk) 21:34, 4 January 2009 (UTC)

References to supply

 * The sentence "Үү and Өө are sometimes written as Vv or Її and Єє, mainly when using Russian software or keyboards that don't support them." is still unreferenced. It's a very useful sentence, so it would be a pity to delete it in the course. Maybe someone can check if http://www.prokg.com/promail/mail_ru1.htm provides some information that is independent from Wikipedia and could be used as a reference here. Yag zun huv' hurehgui ch bolno, undsen bodol sanaag ilerhiilbel hangalttai. But indeed, any wiki-independent source would be welcome. G Purevdorj (talk) 01:01, 9 January 2009 (UTC)


 * The Russian reference supplied here is about the ways of typing in the Kirghiz language. Although the Turkic languages feature the same feminine vowels (they call it "soft") o and u, this study cannot be automatically applied to the Mongolian language. Some Mongolian language websites seemed to be using Latin Vv instead of Mongolian Cyrillic Її a couple of years ago. Gantuya eng (talk) 13:28, 21 January 2009 (UTC)


 * Here's the evidence
 * 2009-01-21 08:48 | Цагаан cap бол билгийн улирлын тооллоор шинэ хуучин он солигдох шинийн нэгний єдєр. Энэ єдрийг монголчууд хахир євлийг давж, урин хавартай золгож, нэг нас нэмсний баяр болгон тэмдэглэж ирсэн. Мєн цагаан cap нь угтаа хувхай хоосон єдєр гэгддэг хар єдєр юм байж. Vvнийг элбэг дэлбэг цагаан єдєр болгохын тулд цагаан cap хэмээн тэмдэглэдэг
 * www.olloo.mn
 * Gantuya eng (talk) 13:31, 21 January 2009 (UTC)


 * Yep, I know. Vv looks far better than Її. But still, original research. G Purevdorj (talk) 13:41, 21 January 2009 (UTC)


 * I guess an alphabet quoted from an official text is evidence enough for Її and Єє, but of course the conditions of usage where not specified in the source and accordingly had to be deleted in the article. The same holds, at the time being, for Vv. G Purevdorj (talk) 13:51, 21 January 2009 (UTC)

[l] or no [l]?
The article says, "Mongolian lacks a true phoneme /l/; instead, it has a voiced alveolar lateral fricative,, which is often realized as ." However, example words in the article include "the residents of Ulaanbaatar" and  "goose". Is [l] an allophone of too? If so, that should mentioned. Or are these examples wrong; should they actually be and ? +Angr 06:22, 2 July 2009 (UTC)
 * Both examples were wrong, and there were several other mistakes in the Stress section as well, partly due to different notational conventions. (When Poppe writes /b/, he just means a lenis bilabial plosive, be it due to possibly contextually induced voicedness or lack of aspiration, but I'm not sure whether Walker was aware of that.) Anyway, I've fixed it, thanks for pointing it out! G Purevdorj (talk) 07:48, 2 July 2009 (UTC)
 * Thanks for the explanation and correction. Typologically, it is very weird that Mongolian has /p/ but no /b/ and /g ɢ/ but no /k q/. +Angr 09:34, 2 July 2009 (UTC)
 * "/g ɢ/ but no /k q/" is weird, but historically conditioned, as > /x/ [x ~ χ]. "/p/ but no /b/" seems rather innocent to me, cp. German where weak voicing is usually combined with unaspiratedness and voicelessness with aspiration, so that you could indeed argue for aspiration instead of voicedness as distinctive. [p] should be less marked than [b], so the absence of /b/ in the presence of /p/ shouldn't be that remarkable. By the way, while the phonetics of German and Mongolian plosives and affricates do differ slightly, labial to velar stop pairs are exchangable, eg German /d/ [d] maps onto Mongolian /t/ [t] and German /t/  maps onto Mongolian /t/ . (That might also be due to intervocalic voicing of Mongolian /t/ that I personally perceive as quite regular, but I don't have any scientific data on this interesting question and am usually not inclined to trust my own ears on a matter like this.) G Purevdorj (talk) 10:31, 2 July 2009 (UTC)
 * /p/ but no /b/ is weird in a system that otherwise has voiced stops. Although in general voiced stops are more marked than voiceless ones, with the labials it's the other way around: because of the comparatively long distance between the glottis and the place of articulation, it's actually easier to produce a voiced labial than a voiceless one. With dorsals it's the other way around: the short distance between glottis and POA makes it easier to produce a voiceless dorsal than a voiced one. The Paradebeispiel here is Classical Arabic, which has lost /p/ (by changing it to /f/) and /g/ (by changing it to /dZ/), but has kept /b/ and /k q/. It would be interesting to know exactly how strong the voicing of Mongolian /g G/ is; if they're only weakly voiced except in intervocalic position, an argument could be made that they're really /k q/ after all. It would certainly be a more symmetrical system that way. +Angr 11:10, 2 July 2009 (UTC)
 * Thanks for the info on the typological markedness of [p]; that was new to me. I fear I cannot tell anything about the exact strength of voicedness of /g G/, Svantesson et al. 2003 don’t provide any specific info. I’d personally rather perceive [b] than [p], while I clearly perceive the dentals and palatals to be voiceless at least in some positions. But that is a question that you might answer for me: I’ll email you a wave-form from Svantesson et al. 2003: 18. I cannot read such stuff myself (I don’t have any special training in phonetics), but I guess you can tell the difference between [p] and [b] from it. If you’re really interested about /g G/, I could probably find some internet news and provide you with a transcription of some relevant words, so that you could listen for yourself. You might also want to compare for  with an amazing frequency of the word 'sneakers'.  has a pretty interesting /g/ at 1.00-1.02 "... chi gertee ochij ...", at the end of this very interview you have "za, bayarlalaa, bayartai", thus two /b/-s to have a look at. G Purevdorj (talk) 12:09, 2 July 2009 (UTC)
 * Also, addressing the first issue, /ɮ/ is an affricate consonant and technically not /l/ alone. This can be seen in languages that like a /tʃ/ sound but the language may not have /t/ alone (though this example is rather unlikely, it was meant to prove a point). User:Lawfulreasoning (User talk:Lawfulreasoning) 10:50, 5 June 2012

Miscellaneous questions 31 August 2009
1. Lead. number of voices. Is there consensus or majority agreement on the number of grammatical voices, and which ones they are?

2. Classification. "In Inner Mongolia . . . . it is assumed that they jointly provide a standard grammar despite their internal grammatical differences." This literally doesn't make sense to me, that a combination of dialects "provides" a standard grammar.

3. Morphology. which case marking obtains when an indefinite noun is specific?

4. “a number of postpositions exist that usually govern genitive, ablative, or comitative case or an oblique form”. Oblique typically is means “any case other than nominative and accusative”. You seem to be using it to denote certain derivatives of a word stem. Apparently this is what you meant in mentioning “the always oblique agent” in a passive sentence. In any case, to say that a preposition “governs” an oblique form is an unexpected collocation, although it seems to be terminologically consistent when I ponder it. Do you mean that these prepositions require that the stem be augmented?

5. Syntax. There are two versions given for ‘it.’: ‘’ ünijg’’ and ‘’ü ünijg’’. Dale Chock (talk) 01:43, 1 September 2009 (UTC)

6. Use of 'q' in transliteration, e.g., Kökeqota. 'q' is not included in the alphabet table. How is it pronounced? Dale Chock (talk) 02:59, 1 September 2009 (UTC)


 * I'll see to the recent changes later, now only a list of short answers to your questions:


 * 1. Usually 5, sometimes 4. I don't remember why I put it as "includes" back then, but that formulation appears inappropriate to me now. Changed to "consists of".
 * 2. The exact wording in the Mongolian text is that "South Mongolian" (the standard language) is "based" on the dialects in question, while its pronunciation is "based" on Chakhar. "provide" made sense to me, but you might find a more appropriate word.
 * 3. Anmicay and textual context plays a role then, but that hasn't been worked out in detail yet. Guntsetseg's dissertation (probably forthcoming in 2010) will show.
 * 4. Yes, there is a term "oblique case" in Mongolian philological terminology that marks 1. the stem including the floating -/n 2. any non-subject use of the nominative. I've replaced it.
 * 5. Corrected.
 * 6. Difficult issue. None of Written Mongolian transcription is included in the transcription table. That is the same as "kündü, reconstructed as "; modern Inner Mongolian standard would be /xunt/ ~ [xunt]. That is, the question is also when to pronounce q. In Modern Mongolian (that is, Khalkha or Chakhar) Kökeqota is, . For Middle Mongolian, Svantesson et al. reconstruct , . G Purevdorj (talk) 08:31, 1 September 2009 (UTC)

Stress

 * Köke doesn't allege that his data cannot be accounted for by Walker's analysis as he doesn't quote her, but it's trivial to conclude this as both analyses consist of rather precise descriptions of generalized stress patterns. The other point with POV was Köke's "conflicting" data as he disperses with possible ideolectical characteristics. Anyway, given that Köke is better able to deal with this data in chapter 6 than in chapter 4, it needn't really be mentioned after all. As for the quality of the data: Köke is the only one to use data from three informants and on F0, intensity, vowel length and (partly) vowel quality, thus providing transparent and non-arbitrary data. Walker on the other hand just relies on one informant who might produce unrepresentative data for various reasons. Thus, Köke's evidence is really better and would have precedence of that of Walker if it was not for possible dialectal differences.
 * Note on my last edit: I did the first edit after finishing chapter 4, assuming that sentence intonation in chapter 5 won't affect the preliminary conclusions. It did, as the words discussed previously were a part of a carrier sentence with continuously falling F0, thus a syntactic-phonetic level F0 for two adjacent syllables might indicate a morpho-phonetic LOW-HIGH. I changed that accordingly, including the fact that Köke (in the final chapter) is (unexpectedly) careful in drawing any conclusions. G Purevdorj (talk) 00:08, 8 September 2009 (UTC)

Due to lack the knowledge of the language or the linguistic research on the language, it is hard for me to agree or disagree with specific claims made for/about Mongolian grammar. The motivation of my most recent edit was to adhere to the POV and OR guidelines. (Apology: when I wrote "POV", I meant "OR", but they can cooccur.) If a Wikipedia article claims that "there is no way that data set X can be accounted for by the argument of researcher Y", that claim has to come, not from a Wikipedia editor, but from some other researcher who published in a "Reliable Source". In my edit, I therefore improvised a solution whereby there would be a stronger implication that the claim about accounting for the data came from an RS. This was just a tactic: I expected that if the edit were erroneous, it would soon be corrected. If there's a "next time", I'll probably change tactic and raise the question here, on the Talk page. :) Minor spelling note: idiolect. Dale Chock (talk) 17:53, 8 September 2009 (UTC)
 * I did disagree with "POV", but you're right, a claim as strong as that did amount to OR, especially as there is no need to use such a blatant formulation. By the way, thanks a lot for your copyediting so far, the article did really improve! G Purevdorj (talk) 20:30, 8 September 2009 (UTC)

Here's a new question about the article's Stress section. With regard to Köke 2003, "most extensive phonetic data collected so far", we need to specify whether this applies to the Chakhar dialect alone or to all Mongolian. Dale Chock (talk) 00:02, 9 September 2009 (UTC)
 * All of Mongolian. I am not sufficiently sure what is happening in Japan, though. It would be possible (though improbable) that a more extensive study has been published there since 2004. Mongolian studies love to be disconnected. G Purevdorj (talk) 08:38, 9 September 2009 (UTC)

1. As far as I can see (and contrary to the way I originally cited it), it would have to be "Harnud, Huhe", not "Huhe, Harnud". As this would be misleading, it is probably better to cite him using his actual Mongolian name "Köke". 2. You've inserted a review but we not usually do this. This would only be worthwhile if it contains significant corrections or if it would be the foundation for an assessment mentioned in the article. Does it? (If you have an electronic version, you might send it to me so that I can see for myself. This journal is usually easy enough to access, but I'm in the countryside.) If it remained in the article, it would have to be integrated into the bibliography. G Purevdorj (talk) 09:20, 10 September 2009 (UTC)


 * 1. Yes, my mistake. "Harnud" is the family name; when I looked up "Harnud" in Google Scholar, I found many papers coauthored by Huhe Harnud. I do not understand why your insistence on referring to him/her as Köke, especially by that solely in any spot (you use solely that name in the note, you use it together with the alternative version in the bibliography). So that people can find an author in the literature, the author needs to be cited by the name they go by in the literature. I am editing the article accordingly; but of course I am open minded if someone wants to explain that I am error. Pardon my ignorance of the language, but are you sure "Köke" corresponds to "Harnud" and not to "Huhe"? Now that I searched Google Scholar for  (i.e., not as a phrase), there was only one result — due to an editing error in the result. The author inadvertently referred to Huhe Harnud as "Köke" just once in her dissertation (and there's no corresponding entry in her bibliography). Check this out, it's a brand new dissertation on word stress in Oirat by a native speaker, in public access PDF. It contains many remarks about Mongolic. 2. As for the review, I inserted it because it is much more widely accessible than the actual work. This is an innovation of mine inspired by this very use of the Harnud study. I don't really know how to best cite it. I still think what I did is the best solution, but I'm open to discussion. —Preceding unsigned comment added by Dale Chock (talk • contribs) 17:43, 10 September 2009 (UTC)
 * 1. His name is Köke ~ Huhe ~, of course. If you're South Mongolian and write in a western language, you just have to make up some surname. You can take your father's name, your tribal name etc. Thus, "Harnud" is a kind of artefact. In Chinese language publications he uses "Huhe", in Mongolian publications "Köke". As this author published much more in these languages than in English, it gets all mixed up as no one can identify Harnud with Köke. Therefore, this article does. For another quotation of Köke on Wikipedia, see Chakhar dialect. 2. Yep, it's real difficult to get a hand at this publication, thus your idea also to quote a review is possibly justified. There are some other publications quoted in this article that are even less easy to access, though. G Purevdorj (talk) 18:47, 10 September 2009 (UTC)


 * As for whether to direct readers to published reviews, maybe it should be left to the reader's professional savvy: "when doing literature searches, don't just look for monographs, look for reviews of those monographs". Dale Chock (talk) 00:38, 11 September 2009 (UTC)


 * In the dissertation on Oirat, several letters are not displayed (eg on page 21). Did you have this same problem when reading or is it the fault of my Acrobat reader? G Purevdorj (talk) 15:40, 11 September 2009 (UTC)

Cleanup of the bibliography citations
The citations were misformatted, at least by Western practice. You don't set off publication year with a colon; you don't run the serial title together with the volume number. ¡Even if you italicize volume numbers, you still separate them from the serial title with a comma! Some publications still enclose publication year in parentheses, but I think that's a total waste of time, although I wouldn't fight persistently to keep those parentheses out. (Enclosing the year in parentheses in citations in the text is another matter. I noticed that just last month Purevdorj asked on a different talk page whether that is passé. My first reaction was to agree. But I checked very recent issues of several linguistics journals and was surprised to find that the choice of practice varies from journal to journal.) When you cite a chapter from an anthology, this is how to format it: Smith, John. 2003. Politics in Kansas. In Martin, John, and John Brown (eds.), Rightwing devastation of America. Sometown: Publisher X. Series: University of XX studies in American politics; 17. pp. 236-283. Details vary, of course. The sequence of Author, year, article title is key. Dale Chock (talk) 00:53, 11 September 2009 (UTC)

When citing journal articles, you don't introduce the journal title with anything, including "In:". That's very wrong. I hadn't yet noticed that when I wrote the preceding. Dale Chock (talk) 04:48, 11 September 2009 (UTC)


 * The "In" before journal titles is rare, I know. The rest appears to be quite in variation, and I would somewhat prefer "(1999):" to ". 1999.". Doesn't make much of a difference, though. The conventions I usually follow do away with the series. ", pp. 177-179" instead of ": 177-179" is a waste of ink. By the way, how conventional is "177-9"? G Purevdorj (talk) 08:08, 11 September 2009 (UTC)


 * Again, I can tolerate parenthesizing the year (and Wikipedia's template does so). As for following the year with a colon, I have never seen that elsewhere. Series are perhaps not included by everyone, but at the least, it's a common practice in linguistics. As for pagination in the style, "177-9", my sense is that it has plunged in popularity in the last 30 years, maybe even 40. Up until at least 1970, it was common, although maybe not in linguistics. Dale Chock (talk) 18:16, 11 September 2009 (UTC)

What about the letter, ė?
I didn't notice this character in the orthography table. What is its IPA? Should it be applied in the Mongolian transliterations? (I already respelled helnij with it.) I simply cut and pasted it from a Web page, specifically the WorldCat site. Dale Chock (talk) 04:56, 11 September 2009 (UTC)
 * "ė" [e] is used to transcribe Cyrillic "э" in Russian, while Cyrillic "e" is transcribed as "e" [e] (but signalling that the preceding consonant is palatalized). This is not practical for Mongolian as "e" is rare and "э" omnipresent. Therefore, a quite conventional practice for Mongolian Cyrillic is to transcribe "э" as "e" [e] and "e" in another way, eg as "je" (in words of Mongolian origin). This is what this article does. "helnij" is preferable. G Purevdorj (talk) 08:20, 11 September 2009 (UTC)


 * OK, I've reverted 'ė' to 'e'. Dale Chock (talk) 18:16, 11 September 2009 (UTC)

Mongolian Academy of Sciences (MAS)
In a edit summary, the question has just been put: why insert the acronym MAS alongside ŠUA? I had found out the ŠUA stood for Academy of Sciences and that its full name was, of course Mongolian Academy of Sciences. I wanted to put a link to their Web site, www.mas.ac.mn/en/. Now that the objection to "MAS" has been raised, let's have some discussion about it. Dale Chock (talk) 01:56, 19 September 2009 (UTC)
 * I wasn't aware that they are using MAS themselves. As they do, I have no objections and just undid my last edit. G Purevdorj (talk) 07:26, 19 September 2009 (UTC)

Language usage edits 21-22 September 2009
1. 'Comprise' or 'constitute'? Comprise is a permanent nuisance in English usage. The usage of these two words ought to be: the parts constitute the whole, the whole comprises the parts ('comprise' comes from 'comprehend'). (Also, the whole consists of the parts.) Although many people do use 'comprise' as a synonym for 'constitute' (see discussion here), in normative use this is to be avoided. In an academic essay such as this article, ambiguity of vocabulary is a bad thing. And since 'consist of' and 'constitute' are never "flipped" in meaning, they are the reasonable choice of point of departure; 'comprise' should be assigned a meaning that does not conflict with their meanings. Another alternative would be to eschew 'comprise' in order to avoid confusion.

2. The use of "as well" in the following edit is inappropriate. Some might argue that this might be more a logic issue than an issue of idiomatic English usage, but I think it is a vocabulary issue, at least in part. "Mongolian has been written in a variety of alphabets. . . . Between 1930 and 1932, a short-lived attempt was made to introduce the Latin script in the Mongolian state, and after a preparatory phase, the Cyrillic script was declared as mandatory by government decree. From 1991 to 1994, an attempt at reintroducing the traditional alphabet failed in the face of popular resistance.[86] In informal contexts of electronic text production, the use of Latin is common as well." In my opinion, the "as well" would only be apt if the immediately preceding sentence made explicit reference to either the use of some other alphabet or to some purpose other than electronic text production. Apparently, the intended interpretation is "in informal contexts of electronic text production, the Cyrillic alphabet is of course usual, and the Latin alphabet is common as well". But as I just said, the first part of this thought is not expressed in close enough proximity to the clause about the Latin alphabet. In any case, it would be ridiculously superfluous to mention that the Cyrillic alphabet is used in informal contexts of electronic text production for Mongolian — either often or most often — given that that's the alphabet Mongolian is written in otherwise. (Unless it were in fact the case that Mongolians tend to avoid their own established alphabet when producing electronic texts. But that would be big news deserving of more conspicuous announcement.) Since we are thus enjoined from stating that Cyrillic is often used for that purpose, it would not be apt to state that a different alphabet is "also common" or "common as well". If, despite this argument, one has a yearning to employ the phrase "as well", here might be one way to legitimize it (I think this wording would be strained, still not quite apt). ". . . . [in the 1930s] the Cyrillic script was declared as mandatory by government decree. But with the advent of electronic text production, the use of the Latin alphabet has, in informal contexts of this activity, become common as well. From 1991 to 1994, an attempt at reintroducing the traditional alphabet failed in the face of popular resistance.[86]." Note that I think the old sequencing of the three sentences created a disruption of the train of thought. Dale Chock (talk) 01:45, 22 September 2009 (UTC)

Үү written as Її
In the source depicting presence of the Үү replacements with Її this phenomenon occurs when Internet Explorer/FireFox determines encoding table for the site presentstion as Cyrillic (Windows-1251) but if we change this encoding table to the West European (ISO-8859-1) all Її will be visible as Үү (as these symbols were written in the original text). So dealing with the Mongolian Cyrillic web-sites you need sometimes exchange encoding tables, or simply mention that Її = Үү. But in the original source was not written Її, this symbols are erroneously presented at your computer as encoding table was misdefined. Bogomolov.PL (talk) 07:47, 9 November 2009 (UTC)

'Uigur' or 'Uighur'?
This is in response to an edit 11 Nov 2009 by G Purevdorj. After all, two of the article's four tokens of "Uigur" occur in a single sentence, yet in that sentence you only added the 'h' to the first token. Back 2-3 months ago when I saw that the spelling without 'h' is used by none other than the article's leading source, Svantesson et al., that induced me to refrain from adding the 'h'. Since in English language scholarship the 'h' spelling is conventional, it was scholastically appropriate for me to insert the footnote. Dale Chock (talk) 22:50, 19 November 2009 (UTC)

ү replaced by v
Well, one possible reference for this:. Іt's not the strongest statement; just a computer scientist describing his own transcription practise, not a linguist pointing out a widespread phenomenon in the language.


 * 我々はＪＩＳコードに含まれるキリル文字を用いてモンゴル語を表記することとする. しかし、モンゴル新字にはＪＩＳコードのキリル文字セットには含まれていない２種の文字が存在する. ӨとҮである（小文字はөとү）. 我々は、前者に対してギリシャ文字のΘを用い、後者に対しては英文字のＶを用いることで対処した（小文字はθとｖ）. こうすることでモンゴル新字をＪＩＳコードの範囲内で扱うことができる.

If you don't read Japanese, that roughly says:


 * (In this paper) we use only Cyrillic letters which are included in the JIS encoding to write down Mongolian. However, there exist two letters in the new Mongolian alphabet which are not contained in the JIS encoding's Cyrillic character set. These are Ө and Ү (in lower case, ө and ү). Here, we use Θ (theta) for the former and V for the latter (in lower case, θ and v). By doing this, we can handle the Mongolian alphabet while staying within the limits of JIS encoding.

It may be rather obscure, but hopefully it qualifies as WP:RS: EHARA Terumasa, HAYATA Suzushi, and KIMURA Nobuyuki, "茶筌を用いたモンゴル語の形態素解析", in Proceedings of the 10th Conference of the Association for Natural Language Processing, pp. 709-712, March 2004 (link to association's website). cab (talk) 06:55, 9 November 2009 (UTC)


 * Or, (google cache) makes direct reference to this phenomenon too:


 * “ө”, ”ү” гэж хоёр онцгой үсгийг өөрийнхөө компьютер дээр ариал мон-оороо бичихэд зүв зүгээр мөртлөө Интернетэд хэрэглэж болдоггүй “ө” үсгийг буруу харсан ”э” буюу ”ү” үсгийг латиний ”v” үсгээр орлуулж бичнэ.


 * Don't ask me for a translation, I only understood the 2 crucial parts Интернетэд and орлуулж бичнэ =). And I don't quite know what's the provenance of that document, it just happened to be sitting on some Mongolian government department's website. cab (talk) 07:23, 9 November 2009 (UTC)
 * The first source is not sufficient, as it is only an instance (which might be insignificant or unrepresentative), but the second source will do nicely. It deals with the phenomenon as such and indeed appears to be authored by some Mongolian bureaucrat from a ministry, which is just perfect. I'm afraid I cannot deal with the matter today, but your source will enable us to settle this matter once and for all! G Purevdorj (talk) 12:55, 9 November 2009 (UTC)
 * Alas, the Mongolian text is gone! Has anybody downloaded it in time? G Purevdorj (talk) 23:44, 24 December 2009 (UTC)


 * Mirrored here (google cache). Cheers, cab (talk) 08:21, 28 December 2009 (UTC)

New video
Obviously, we got a new video in the phonology section. The sound quality is not that perfect, and it should be completely unintelligible to beginners anyway, but that's probably okay. The video seems to be private, but all of the speechactions are commonplace, so that it is not that likely that somebody will consider her or his privatesphere violated by this. (Unless, of course, one would altogether object to being displayed at a prominent place on the internet.) If the video is indeed to be considered unproblematic, I have a suggestion to make. If somebody (maybe Ganaa) would provide a transcription into Cyrillic or Latin Mongolian, I would be willing to provide a phonemic transcription, morpheme analysis and translation. If we would have some opportunity to embed this in wikipedia as well, we would have a wonderful. By the way, do you prefer "translates it as GOOD as he can" or "as WELL as he can"? G Purevdorj (talk) 10:20, 2 January 2010 (UTC)


 * In formal English, "good" is an adjective and "well" is an adverb. Don't get confused by colloquial use of "good" for both. --Latebird (talk) 19:57, 5 January 2010 (UTC)


 * Always vodka. Vodka everywhere. Sea of vodka. Disgusting. Gantuya eng (talk) 11:55, 2 January 2010 (UTC)
 * Well, pretty typical, but I guess that is an answer to my suggestion. G Purevdorj (talk) 12:26, 2 January 2010 (UTC)


 * OK, I'll tell you what I know about it. I took it in a public place, when I was in Mongolia, so it certainly should fall within fair use.  The image quality is not the best, but that might be good; it makes it pretty hard to recognize the individuals, which is a good thing for fair use.  My Mongolian is only good for a few phrases, so I certainly can not translate it.  Even without a translation, I think it is interesting to have this, just so a novice can hear what the language sounds like.  I do have some more videos.  I have one (which is very dark) of a clip of the whole group singing a folk song (that is unrecognized to me) that I may add to the Music of Mongolia page soon.  Maybe someone can identify the song, but it might be nice even without this since there is no music on the page now.  Thanks for your comments, let me know if you have more questions or comments.--Qfl247 (talk) 17:53, 2 January 2010 (UTC) P.S. Yes, there is ALWAYS Vodka...

The phonemes /k, kʰ, x/ in Middle Mongolian
My latest edit in this regard is as follows. "Middle Mongolian documents show only two velar plosives, /k/ and /kʰ/ (each having one allophone), but /k/ disappeared in some instances and not in others. Evidence as to the identity of the conditioning factors for the instances of disappearance is elusive, although it has been hypothesized based on evidence from Phags-pa that they have to do with distinctive vowel length or stress.[98] However, all four orthographies for Middle Mongolian retain a letter for h, restricted to word-initial position, which might mark the same phoneme /h/ (possibly [x]) as some of the instances of the letter for k. This would then imply that x → h → Ø." A minor point is that the previous version, was punctuated in such a way as to give misleading prominence to the clause that /h/ was restricted to word initial position. While this may be important as evidence for the speculation whose statement follows, the speculation is the actual point of chief interest: "retain a letter . . . which might mark the same phoneme /h/ as . . . ."

My major point is to raise a question. I don't understand what you're trying to say with the sentence I have put in boldface, because it's not made clear what the reflexes of /k/ are. If /k/ merely did not disappear in some instances, then its reflex in those instances would be [k]. How would indications for /h/ "be the same" as indications for /k/? Dale Chock (talk) 01:57, 22 September 2009 (UTC)


 * Well, ok, we’re at the most important single problem of Mongolian historical linguistics. The phonemes are /k/ and //. These sounds are restricted to front-vocalic words, though, while back-vocalic words have [q] and [], respectively. While the aspirated phoneme was fully retained, some instances of /k/ (both [k] and [q], marked by the letters  and &lt;q&gt;, respectively, in pre-classical Mongolian script) disappeared. This disappearance almost always took place in inter-vocalic position, but some instances the letters  and &lt;q&gt; were also retained as /k/. No clear evidence has been found that would point to the reason why some instances of , &lt;q&gt; disappeared and others did not. There was the speculation that it disappeared before or after primary long vowels that are alleged to have been indicated by one obscure letter in Phags-pa script only, but this matter is doubtful. Another thesis was not primary long vowels, but stress. (Note: I say primary long vowels, because the so-called secondary long vowels arose in most Mongolian languages after the vowel written by /k/ was lost in those instances where it was lost.) Now a thesis has been put forward, I think by Doerfer, but at least it was promoted in Svantesson et al. 2005, that another explanation is more feasible. There is no doubt that there is a separate letter  in Arabic, Phags-pa and Chinese Mongolian script, which is not present in Pre-Classical Mongolian. It is a mystery to scientists why it is absent in Pre-Classical Mongolian which reflects a stage of Mongolian phonology that is at least 50 years older than the evidence conveyed by the other scripts. One thesis speculates about dialectal differences, but again (and in this case this is not even disputed), there is no actual evidence for that. On the other hand, Pre-Classical Mongolian is the only script where the letters , &lt;q&gt; is still retained in those instances where it was to disappear, while the other scripts already have zero in those instances. Now it is a bit strange to have a phoneme that occurs in world-initial (instead of, say, syllable-initial) position only.  disappeared, and some instances of , &lt;q&gt; disappeared, without it being possible to find convincing conditions why. If, however, the letters  and &lt;q&gt; represented two phonemes each, and one disappeared and the other not, but the phoneme that disappeared was for some phonetic reason differentiated in word-initial position, a neat explanation would have been found. Did you understand what I wrote, or was it still too unsystematic? (Tables help a great deal in matters like this.) G Purevdorj (talk) 13:54, 22 September 2009 (UTC)


 * I think I understand it, but I'm not sure I do. As far as I do, your exposition seems to have some loose ends. I can read the exposition in Svantesson 2005, and if I still have questions after doing that, I can put them. Dale Chock (talk) 06:27, 24 September 2009 (UTC)


 * I don't think that the revised passage starting from "Another prominent disagreement ..." is much of an improvement. It is necessary to spell out this matter naming the correspondences, and to do so within the text. I'm afraid I don't have time to make more detailed suggestions at the time being because I'm in an awful hurry. Bye. G Purevdorj (talk) 09:09, 29 September 2009 (UTC)


 * You didn't spell out specifics either. Anyway, I disagree with your objection because what you wish to report is in fact a piece of pure speculation on an obscure, minor point. It would be just as well not to mention it at all, but I have deferred to you in your wish to report about it. But the original explanation was confusing. By checking your source, I came to understand how your explanation was patchy. There are three correspondence sets, not two. You did not quote their rationale for the medial *h reconstruction, yet without it the reader cannot make sense of the further claim that those reflexes should be united with another, word initial reflex. My revision is quite clear, it just leaves out the specifics, which is good because this is all speculation from a single published work, Svantesson et al. 2005, and the specifics are lengthy. See WP:WEIGHT.


 * The evidence underlying the received assumption that there did exist a medial phoneme /k/ prior to "Middle Mongolian" is from Uigur Mongolian, which your source (Svantesson, et al. 2005) implies ought to be included in "Middle Mongolian". According to this source's speculation, the proposals about conditioning factors — which in turn are pure speculations — are pointless because there was in fact no split in reflexes of a single reconstructed medial phoneme (rather, our reconstruction ought to have multiple medial phonemes). Therefore, you are spending too much time presenting the speculations about conditioning factors because the issue of whether there was a split is so obscure and is poorly evidenced. Dale Chock (talk) 15:36, 5 October 2009 (UTC)


 * 1. Svantesson et al. is sometimes a little intransparent as they don’t quote their sources properly, but just gives a list of literature for every section. Janhunen 2003c: 5-7 holds the same view (though both make conflicting suggestions about Late Pre-Proto-Mongolian), and the whole idea goes back to Doerfer. It is thus fairly widespread, probably the most common view for Europeans nowadays. As far as I know, it has not been adapted by any Mongolian scientists, so admittedly quoting both views is more appropriate.
 * 2. Just noted that Svantesson et al. do not misquote Poppe 1955, see 225 where they quote him verbatim.
 * 3. I don’t see “three correspondence sets”. UG has k, q, the rest has k, q, 0. That means:


 * four sets
 * UG : remainder


 * [k] : [k]
 * [q] : [q]
 * [k] : 0
 * [q] : 0


 * two sets
 * UG : remainder


 * /k/ : /k/
 * /k/ : 0


 * Including /kh/ (third set: /kh/ : /kh/) just leads to confusion. I was also greatly confused by your text that for me seemed to implicate that correspondence sets (instead of single phonemes that are indicated in the source texts) straightforwardly lead to the reconstruction of extra phonemes. I really needed quite some time to figure out what you mean. I hope that my new version is more comprehensible than both my previous version and yours. (By the way, we really cannot leave out this matter. For anybody interested in the history of Mongolic/Mongolian, it is absolutely crucial.) G Purevdorj (talk) 20:21, 5 October 2009 (UTC)


 * Apologies for replying weeks later to the immediately preceding comment of 20:21, 5 October 2009 (UTC). For the benefit of those unfamiliar with this discussion, it refers to pages 121-122 of §8.7.2, "The fricative *h" in Svantesson et al. 2005. (§8.7 is Old Mongolian consonants.) These authors (not me) reconstruct three particular phonemes: *kh, *k, *h.


 * The authors in fact present three pairs of correspondences as examples from their data (The authors draw from five dialects (or five corpuses, if one prefers): "Uighur M", Halh, "Sino-M", "Arabic M", and 'Phags-pa. Halh is a modern dialect. The other four are historical dialects named by the alphabets used to write them.) Hence I reported "three correspondence sets". (In historical linguistics a "correspondence" consists of cognates, or more precisely it consists of the sounds that occur in corresponding positions in a set of the cognates of a single notion. G Purevdorj seems to assume that a "correspondence" can only compare two things, e.g., two dialects. On the contrary, on page 121 the authors present five way correspondences.) On page 121, each of the three pairs consists of a uvular correspondence and a velar correspondence. The "correspondences" presented are (in same order as my listing of the dialects).


 * 'dog' -q-, -x-, --, -q-, -q- (Halh reportedly has a velar, -x- instead of a uvular)
 * 'to die' -k-, -x, --, -k-, -- (Halh underwent apocope subsequently, rendering the medial consonant a final consonant)


 * 'snake' -q-, --, --, --, -q-
 * 'word' -k-, -g, -k-, -k-, -g- (Halh underwent apocope subsequently, rendering the medial consonant a final consonant)


 * 'mountain' -q-, the other four null
 * 'above' -k-, the other four null


 * (Actually, the authors give whole words, but I have abstracted the sounds in question, namely the back consonants in intervocalic position.) From these three pairs of correspondences, the authors reconstruct the phonemes *kh, *k, *h, respectively, with each phoneme having uvular and velar allophones. (For *h, the phonetic content, i.e., the identity of the allophones, is deemed unrecoverable.)


 * Now to explain what the issue is in Mongolian historical linguistics that the above analysis pertains to (please refer to two previous comments under this heading, by G Purevdorj on 13:54, 22 September 2009, and by me on 15:36, 5 October 2009 (UTC)). According to G Purevdorj on 22 Sept., the issue has long been defined as: medial &lt;q&gt; and  from "Uighur M" sometimes have zero reflexes in other dialects, sometimes nonzero reflexes. Given linguistic theory, this definition of the data leads to a hunt for conditioning factors that would produce the divergent outcomes (divergent reflexes). Svantesson et al. 2005: 121-122 reject this binary conception of the problem, replacing it with a ternary conception: "The root of the problem is the fact that the sounds represented by the Uigur [sic] Mongolian letters ___ &lt;q&gt; and ___  in intervocalic position may develop in three different ways in the modern languages, for example Halh". These authors counter that the nonzero reflexes are not a homogeneous group, but a heterogeneous group. Therefore, if we insist on a single protophoneme for all the correspondences involving intervocalic back consonants in Mongolian dialects, the challenge to us has been intensified massively: we now have to explain a three way split, not a two way split. The authors argue that there were three protophonemes, hence there never was a conditioned split. That is why I previously argued (on 5 Oct.) that it was a bad idea to go into detail about the speculations on the conditioning factors. The proper course for writing this article is to explain that there accounting for the various reflexes depends on how many protophonemes there were (three, two, or one), and leave it at that. Dale Chock (talk) 18:59, 24 November 2009 (UTC)
 * I'm sorry I won't be read your text for at least this week. But I will address it. G Purevdorj (talk) 15:50, 25 November 2009 (UTC)


 * Due to an error in text editing, my concluding sentence is incoherent. It should read: "The proper course for writing this article is to confine ourselves to explaining that how one accounts for the various reflexes of intervocalic back consonants depends on how many back consonant protophonemes (three, two, or one) one posits there to have been in intervocalic position." Dale Chock (talk) 23:33, 3 December 2009 (UTC)


 * Except for leaving out WM (the sixth “dialect”) which does not exactly correspond to UM in the case of aγula ‘mountain’, there is little that can be added to your recapitulation of Svantesson et al. (I might add that the word for ‘dog’ in Khalkha has a uvular fricative, but it is only an allophone of /x/ which is provided by Svantesson et al.)


 * When I approached presenting this problem, I wanted to do away with kh which is not clearly distinct from k in UM (it is subject to dispute whether there is already a tendency to indicate it (Weiers) or rather a different use of the letter γ only to facilitate reading in certain positions (Tserensodnom and Taube)), while it is clearly distinct in WM. Therefore, while Svantesson et al. (following Tserensodnom and Taube) assume a problem of differentiating between k and kh in UM, this was seen as totally unproblematic in the old research works, and this minor problem is immediately resolved by Svantesson et al. by following WM with respect to kh. I had considered it wiser to follow the established view in this respect, but Tserensodnom and Taube do have their merits and should therefore probably not be glossed over like this.


 * I just repeat your first suggestion:


 * “A prominent long running disagreement concerns certain correspondences of word medial consonants among the four major scripts (UM, SM, AM, and Ph, which were discussed in the preceding section). Between Uigur Mongolian words containing back voiceless plosives (again, whether uvular or velar) in word medial position and cognates in SM, AM, and Ph, there are three distinct sets of correspondences. Traditional scholarship has given two of the sets a common reconstruction (which means there is a split in the reflexes of a single reconstructed sound), which begs the question of what the conditioning factors were that led to the split. Svantesson, et al. propose instead that there was no split, rather that the three correspondence sets derive from not two, but three different Old Mongolian medials.”


 * Traditional scholarship wouldn’t have established these three sets in the same way as Svantesson et al., as they tended to give precedence to the phoneme analysis given in UM (as “clarified” in WM, in accordance with the opinion of Weiers) and just tried to explain later developments that are reflected in the other scripts. Given this different approach, Svantesson et al.’s conclusion is less obviously arrived at than implied in your version. Finally, I think, the sounds under discussion should be specified. I’m not sure whether elaborating on your last version or on mine is the better idea. Any suggestions, and who of us is to make the next try? Best wishes, G Purevdorj (talk) 00:16, 25 December 2009 (UTC)


 * Yesterday evening I only commented on this discussion, but the current form of the article still seems okay to me. It leaves out kh, granted, but that is just possible because kh in WM or Ph (which contains the richest information of all scripts) does not correspond to k or zero in any modern languages, but has very unproblematic reflexes. "k", by the way, is quite different: atukusu 'horses' > Khalkha, Oirat . Janhunen addresses this problem, Svantesson et al. don't, but they should have done so. Maybe some discussion that incorporates even that data would be helpful. But let me ask from you: what information that is currently present in the article would you like to drop or reformulate? G Purevdorj (talk) 10:34, 25 December 2009 (UTC)


 * Hi, I cannot understand your latest contribution. I also can't be sure what the abbreviation "WM" means (I'm guessing "Western Mongolian"). I will ponder some more. Happy New Year.Dale Chock (talk) 01:44, 12 January 2010 (UTC)
 * Sorry for writing incomprehensible. I'm not sure whether this discussion is worth your continued efforts, though. Anyway, WM is "Written Mongolian" which denotes the "Classical" form of the Mongolian script, in contrast to preclassical Middle Mongolian in Uighur letters which is not called "Written Mongolian" by Svantesson et al. Happy New Year! G Purevdorj (talk) 22:45, 12 January 2010 (UTC)
 * Hättest du lust, die stellungnahmen auf meiner User_talk-seite auf deutsch vorzulegen? Dale Chock (talk) 06:18, 13 January 2010 (UTC)

links
should http://www.mongoliacourses.org/ added as another external link? Seems the course materials for two of the three courses are free with registration. Yaan (talk) 15:55, 5 March 2010 (UTC)
 * What is more: you can access the course material without registration! We definitely need to add that link! Good contrib! G Purevdorj (talk) 23:43, 5 March 2010 (UTC)
 * Any idea what happened to http:\\tbg.torama.ru Sonsov uu? Is the site down or is the server just too slow? Yaan (talk) 10:54, 9 March 2010 (UTC)

A claim about the vowels that needs clarification
"Short non-initial vowels have been reduced to 71% the length of short word-initial vowels and become centralized, in the course of which losing their status as phonemes and becoming non-phonemic."

This sounds very strange and should be clarified. How can a vowel phoneme be said to have "become non-phonemic"? If the non-initial short vowels are "not phonemic", what are they? Allophones of something? Of what then? Or is the intended meaning that they are neutralized (i.e. that all are realized the same way, so that several distinct phonemes share the same allophone)? If so, that should be stated explicitly.--91.148.159.4 (talk) 13:48, 20 July 2010 (UTC)

Thanks, it's clearer now.--91.148.159.4 (talk) 14:26, 20 July 2010 (UTC)

ATR harmony description mystifying
"Some suffixes can occur with /a, ɔ, e, o/, following the last phonemic vowel in the word stem, in which case underlying /ʊ, u/ are realized as [a, e] respectively."

The example given in the text does not show /ʊ, u/ being realized as [a, e]. Nor is it clear where those underlying /ʊ, u/ are supposed to be located. In the suffixes? No, they never change to anything but /ʊ, u/ according to the table in Svantesson et al. (p.48). In the stems? But I don't see them changing at all. Svantesson et al. are cited, yet the description there is very different and, in particular, doesn't mention "underlying /ʊ, u/ being realized as [a, e]". --91.148.159.4 (talk) 14:25, 20 July 2010 (UTC)


 * The idea of that formulation presumably was that harʊl-ar should have an assimilation of ‚a‘ in ‚-ar‘ to the preceding ‚ʊ‘, but that’s a far cry from “underlying”. There is nothing like an underlying archi-phoneme /U/ in the suffixes in question, and flexional suffixes don’t affect the stem in any such meaningful way in Mongolian. Just deleted that strange statement. G Purevdorj (talk) 04:07, 24 July 2010 (UTC)