Talk:Ubykh phonology

Notation for /w/ and /r/
The /w/ and /r/ are described as bilabial and alveolar approximants. If this is the case, please use the appropriate IPA symbols. Using a convenient transcription and calling it IPA only confuses people. It's not a problem in a grammar, but is in an encyclopedia, especially when it's covert. kwami 08:11, 2005 July 28 (UTC)


 * I don't know who did this table originally, but strictly, the /r/ is trilled, and the /w/ is a labiovelar approximant; thus, the symbols were both correct, although their positions in the chart aren't. I've now changed the hyperlinks appropriately, but I'm afraid I don't know how to edit the table to give an extra row for the alveolar trill. thefamouseccles 01:36, 15 Nov 2005 (UTC)


 * Kwami, please note that I said the labial approximant of Ubykh is labiovelar (like the English w), not labiodental (as found in Dutch). The transcription Labiodental approximant is therefore inappropriate. Also, opinions are divided on the status of the Ubykh postalveolars, which is why I had preserved the more traditional IPA symbols for the postalveolars and not used J. C. Catford's sibilant-plus-caret symbols for this series, which transcription is entirely unused outside of Catford's publications. thefamouseccles 12:15, 18 Jan 2006 (UTC)


 * Sorry, I must've just misread the note on the /w/. I'll also restore the frics with a comment in the note that the s^ notation is not widespread. kwami 19:59, 18 January 2006 (UTC)


 * No problem; with a language as phonetically complex as this, there's bound to be some miscommunication eventually. :) thefamouseccles 13:44, 21 Jan 2006 (UTC)


 * I've also altered the hyperlinks from the alveolar stop and fricative series, which for some weird reason were pointing to the dental consonants rather than the alveolar ones. User:thefamouseccles 01:45, 15 Nov 2005 (UTC)


 * I've relabeled the laterals as specifically fricative or approximant. Please correct me if I got it wrong. kwami 20:21, 18 January 2006 (UTC)


 * No worries - you have it right, from a purely phonetic point of view at least. Interestingly, despite its clear approximated phonetic nature, ['] is probably phonologically a voiced lateral fricative rather than an approximant, since (a) it participates in voicing assimilation in consonant clusters, where other approximants do not - compare ' seven, ' red and ' four, and (b) it seems to be at least partially cognate with Adyghe and Kabardian [], the only voiced lateral in those languages (which also participates in cluster assimilation). Nevertheless, that is mostly conjecture and partly original research, so doesn't belong on Wikipedia... thefamouseccles 13:52, 21 Jan 2006 (UTC)

Phonetic or Phonemic?
The allophony rules are given as, e.g.,

I would think that this should be

since it's simply describing the realization of the two phonemes, but I'm not sure enough of the original intent to make the change. Would someone correct the entry or explain why it should stay as it is (so no one else makes that mistake)? --CRGreathouse 22:20, 23 April 2006 (UTC)

How reliable is this analysis?
This analysis of Ubykh phonology has bothered me ever since I first encountered it. 84 consonant phonemes and 2 vowels looks to me like an extreme misanalysis. I don't speak Ubykh, and so I don't know if there's internal evidence in the language that supports the conventional analysis, but based on what I do know, I would guess that it would be better to consider the various vowel sounds as phonemic (giving a 3- or 5-vowel system) and some of the consonant variations (e.g. labialization) as allophonic. I hesitate to put this in the article, at least with my current state of knowledge, but maybe someone who knows more can say more... --Marnen Laibow-Koser (talk) (desk) 15:26, 13 March 2007 (UTC)
 * Speaking as one who has studied Ubykh in detail, there is a general consensus among the academic community about the magnitude of the Northwest Caucasian languages' phonemic inventories. I know of no phonemic analyses since at least 1960 that differ in more than a few details (for instance, John Colarusso actually postulates one more consonant - // - and the status of the velar triad // is still debated, as is the precise place and manner of articulation of the postalveolar affricates and fricatives). There is absolutely no doubt that the analysis is essentially correct as presented, and that the number of consonants is approximately correct (that is to say, this chart includes the plain velar stops, but not //, whose putative existence isn't accepted by many scholars anyway). This is proven by several facts:
 * setting up a vowel inventory // would necessitate the postulation of far too many rules prohibiting the appearance of certain vowels in certain environments, which would need to include: (a) that // and // - the two most common vowels in the language - cannot appear after velar stops, (b) that // cannot appear after any alveolar consonant, any retroflex consonant or any pharyngealised consonant, and (c) that cannot appear after non-stop alveolars (but can after alveolar stops), postalveolar affricates (but can after postalveolar fricatives), velar fricatives (but can after velar stops), and all retroflex consonants;
 * consonants retain labialisation or palatalisation whether they are word-initial (e.g. // wagon, pronounced [] or []) or final; the definite form, // the wagon (which has dropped its final vowel), is pronounced as written, and it is unreasonable to believe that there is a whispered vowel that remains in this case, as the same thing happens with voiced consonants, e.g. // child vs. // the child;
 * the underlying vowel can be (and very often is) generally substituted for one of its allophones without a change in meaning (for instance, // tomorrow may be pronounced either as [] or as [], although substituting, say // for // in a word like // heart would probably be interpreted rather as being phonemically //, as phonetic [] and [] are in free variation with [] and [];
 * more than one vowel in a word may be allophonically conditioned by a single consonant (for instance, // in truth, as it happened (a polymorphemic verbal participle in origin) appears phonetically as [] or []);
 * because labialisation has several different realisations, not all instances of labialisation condition an allophonic variation in vowels; // white is normally pronounced as written, as is labialisation on all alveolopalatal consonants.
 * I understand your concern, because it was something I grappled with myself when I first began to study it, but this is a well-documented phonological phenomenon that occurs not only in every member of Northwest Caucasian, but also in Arrernte (in Australia) and Marshallese (in the Pacific). See vertical vowel system for a brief overview. The phenomenon has also been reported for the short vowels of Irish, although the evidence for it in that language is the subject of much debate. Thefamouseccles 03:40, 23 March 2007 (UTC)

I have not studied Ubykh myself, and this seemingly wasteful analysis bothered me at first, too. However, I have since seen a reason that it seems to be a more accurate analysis than 6 vowel phonemes. the combination of palatalization and the schwa, for example, makes [i], so it would make sense to just call it an /i/ and not worry with all the superscript j's. However, we must remember that the palatalization goes with the consonant, not the vowel, and therefore must be tagged with it. I have seen in my wiki-browsing that such caucasian languages will not shy from single consonant verb roots and such. And a root, if a vowel, will be a or schwa. However, a consonant itself may be palatalized or labialized. therefore, we should tag them as such. —Preceding unsigned comment added by 67.212.106.138 (talk) 01:49, 16 November 2009 (UTC)

Invitation to a discussion that is relevant to this page
I've reverted some recent edits to this page, and explained myself in here. The discussion concerns many articles, so I thought it would be a good idea to keep it central. I hereby invite everyone who is interested to join the discussion on Talk:Northwest Caucasian languages. Thank you

Vito Genovese 13:42, 13 April 2017 (UTC)