Talk:Consonant mutation/Archive 1

Scottish Gaelic
Jwst a note, I'm fairly certian Scottish Gaelic has no eclipses, only lenition.

If you're sure, add it. I don't know enough about the goidelic languages to be certain myself.


 * Not sure about Scottish Gaelic, but Manx has both boggaghey ("soft mutation"/"lenition") and stronnaghey ("nasalisation"). I'd want convincing that Scottish Gaelic was different.  Brief searching suggests it has the same two forms as Manx and Irish (e.g. Universität Duisberg-Essen page on Scottish Gaelic), though of course the results are different in each case. -- Shimmin Beg (talk) 21:31, 11 January 2009 (UTC)


 * The problem is that the Mutation vs. sandhi section of the article suggests that nasalization in Sc. is not a case of consonant mutation (i.e., it only occurs when the preceding word actually ends in a nasal). This is consistent with descriptions of the standard language, where nasalization is not treated as a morphosyntactically significant phenomenon, and is never indicated in the spelling (except for some instances of n- insertion before a vowel — but then that hardly qualifies as a consonant mutation). Of course you can find all kinds of interesting things in the dialects, but I don't think that this article is the right place to go into such detail. CapnPrep (talk) 08:51, 12 January 2009 (UTC)


 * Fair enough. The Manx form is indeed morphosyntactic, though. -- Shimmin Beg (talk) 11:00, 12 January 2009 (UTC)
 * Yes, though in the Late Spoken Manx of the first half of the 20th century nasalization had virtually disappeared. I don't know whether Revived Manx has a nasalization mutation or not. —Angr 11:06, 12 January 2009 (UTC)
 * I can assure you personally that it does (not that that means anything). I'm a Manx speaker and mostly work on gv.wikipedia.org.  It's got quite limited use, but it's there. -- Shimmin Beg (talk) 22:05, 12 January 2009 (UTC)

New title
I changed this article so it covers all sorts consonant mutations in all sorts of languages, not just Celtic initial consonant mutations. I moved the information on specific Celtic languages to the relevant articles. --Angr 11:59, 15 Feb 2005 (UTC)


 * Nice article! &mdash; ishwar  (S PEAK )  15:31, 2005 Mar 31 (UTC)

Merger?
Consonant mutation, Consonant gradation, Lenition, Fortition and Fortis and lenis all seem to be about the same kind of phenomenon. Perhaps they should be merged. FilipeS 21:26, 2 November 2006 (UTC)


 * I don't think so. They are different types of consonant mutation, so they aren't exactly the same thing.  The first two are pretty long articles and so combining them all seems unnecessary when we cand just interwiki between them.  Fortis and lenis isn't consonant mutation, it's a way of describing consonant pairs that differ in ways similar to voiced/voiceless but are, nonetheless, not a voiced/voiceless pair. Æµ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 22:06, 2 November 2006 (UTC)

I understand that fortis and lenis are not the names of consonant mutations. However, those two terms only seem to be applied when talking about fortition and lenition, which are types of consonant mutation, so having separate articles for them seems redundant. By the way, I've just noticed another one: spirantization.

I am not opposed to the idea of having a separate article for each (although, doesn't it make more sense to categorize by language?), but in that case I think that the examples at Consonant mutation, Consonant gradation and Lenition should be better categorized by type of mutation, and taken to a specific article. FilipeS 00:07, 3 November 2006 (UTC)

P.S. See also the discussion here. FilipeS 00:17, 3 November 2006 (UTC)


 * The only reference to Lenition and Fortitian in fortis and lenis is in the see also section. I've heard fortis and lenis applied to English consonants so that English  is fortis and English  is lenis.  The idea behind it was so that you could have one term for, say, both English and Spanish.
 * On a second look at Lenition I see that it may need a little bit of cutting because it oversteps its scope. I see why you wanted to do a merger. If you think that an example in one article would be better in another, go ahead and move it.  I've been under the impression, also, that Consonant Mutation isn't isolated to Celtic languages, it's just common in them.
 * These articles are in the phonology catagory, and consonant mutation is a phonological process. I think putting it in the consonant mutation category would be great. Æµ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 00:57, 3 November 2006 (UTC)
 * Hmm... actually I've just published an article in Lingua arguing that Celtic consonant mutation is a purely morphological process, not phonological at all. —Angr 05:22, 3 November 2006 (UTC)

With respect to the use of the terms "fortis" and "lenis" concerning English, perhaps it should be mentioned in the article.

It seems clear enough that spirantization is a type of lenition (or is it a type of consonant mutation?...) If "consonant mutation", "consonant gradation", "lenition", "fortis and lenis", "fortition" and "lenition" each refer to different things, then their respective articles should explain better what is the difference between them, and the examples currently given definitely need to be redistributed by the articles.

While Wikipedia multiplies the number of articles about phenomena which are, as far as I can tell, very analogous, another thing is still sorely laking in each of them: a more in-depth discussion of diachronic sound change. The vast majority of the examples in these articles concern only synchronic (grammatical) sound changes. FilipeS 10:47, 3 November 2006 (UTC)


 * Spirantization is a type of lenition, and in some cases can also be a consonant mutation. But by definition, consonant mutations have morphosyntactic triggers, not phonological ones, while spirantization can have a purely phonological trigger. Spirantization in Modern Hebrew is morphosyntactically triggered and so is a mutation. Spirantization in Biblical Hebrew was probably purely phonologically triggered, and in modern Spanish certainly is, so in those cases it isn't a mutation. I don't think Spirantization needs a separate article though; it can be beneficially merged with Lenition. Fortition can also be either phonologically or morphosyntactically triggered and so isn't necessarily a mutation; it should have its own article (and the current article can be greatly expanded). Fortis and lenis should also be a separate article so that these terms can be defined independently of the fortition and lenition processes. (For example, it's linked to from the discussion of "fortis and lenis sonorants" at Irish phonology.) Both Fortition and Lenition should discuss both synchronic and diachronic examples, and both phonologically and morphosyntactically triggered examples. —Angr 11:09, 3 November 2006 (UTC)


 * I appeal to Angr's expertise. Although, even if Celtic consonant mutation is a purely morphological process, it must have arisen from an at least partially phonological one. But then, maybe it would've been called something else, like assimilation.  Æµ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 22:29, 3 November 2006 (UTC)
 * As a cross-word phenomenon it would have been a form of external sandhi. See the discussion at the bottom of the article. —Angr 22:54, 3 November 2006 (UTC)

Removal of Finnish
The consonant alternations in Finnish standardly known as consonant gradations are conditioned by mostly syllable type; in the case of /k/, the vowel environment additionally plays a role. The stronger "grade" occurs at the beginning of a open syllable (syllable ending in a vowel) and the weaker grade at the beginning of a closed syllable (syllable ending in a consonant). As for k > j, k > v, k > null, the choice is, as mentioned above, also phonologically conditioned.

In the early years of the generative linguistics era, certain researchers, imbued with the generativist penchant for overemphasizing the theory construction criterion of *parsimony*, broached the feasibility abolishing the syllable from linguistic theory, and specifically from Finnish linguistics. It was readily acknowledged that this would be a big nuisance for Finnish phonology -- the "gradations" are so much easier to explain by invoking syllables -- but it was argued that it would be at least possible to contrive an explanation without recourse to the syllable. But the big discoveries in African linguistics (with William Leben leading off) killed off that suggestion. Or would be expected to have done so.

Comrie in his classic LULT compared the gradations in Finnish and Estonian and showed how developments in the latter have disrupted the old phonology and thus now complicate the phonological explanation of certain consonant alternations. But discussing Estonian morphophonemics/morphonology in this article would be nothing but a distraction.

By the way, there needs to be a fuller explanation of the difference -- or lack thereof -- between "mutation" and (morphophonemic) "alternation". Hurmata 21:32, 9 January 2007 (UTC)


 * Hi. As you say that consonant gradation is a case of sandhi, I have moved the material you deleted to the article on Sandhi, which was still a little empty. Thanks for your contribution, and regards. FilipeS 21:48, 9 January 2007 (UTC)
 * If the distribution is "strong in the onset of an open syllable, weak in the onset of a closed syllable", it isn't sandhi either, and doesn't belong there. I must admit I'm skeptical of this analysis, though, as I can think of no phonological reason why the onset of a closed syllable should be more sonorous than the onset of an open syllable, or indeed any reason why the onset of a syllable should know or care if the syllable is open or closed. —Angr 03:53, 10 January 2007 (UTC)
 * Yes, Finno-Ugric consonant gradation does not fit the definition of "sandhi" given in the opening sentence of the sandhi article. Replies to Angr's objections. A linguist Finn, Karttunen, has written on the Web that the evolution of Finnish in the last 500 years or so has spoiled the original simple explanation of the phenomenon. A grad student in Berkeley ling has claimed that some occurrences are nonphonologically driven. Nevertheless, the following are true: (1) in modern Finnish, *almost* 100 percent of the distribution can be explained by openness of syllable; (2) the phenomenon is a fundamental part of the historical phonology of the Finnic languages that occurs in verbs as well as nouns; (3) once upon a time in Finnish (Karttunen says 500 years ago), gradation was strictly an automatic alternation due to syllable openness/closedness. There may be other Finnic languages for which (3) is still the case. Hurmata 19:19, 10 January 2007 (UTC)
 * I believe Finnish is not a pure case any more, as Hurmata points out, but it nearly is. Have to add info on Estonian here: it's almost the same thing, but with apocope and such wreaking havoc on the structural description, and hence in modern Estonian you can't avoid stating it in terms of morphological categories.
 * Another point: is the term consonant mutation really used in English to describe the Russian phenomena? I do not believe I have ever seen this use. And generally the article appears to need a thorough rewrite (what's English doing here?). Nivkh also has lexically conditioned mutations that override the sandhi rules given, and Bantu stem-final consonant mutation isn't mentioned, as well as Mande, all Atlantic languages minus Fula that have the alternations & so on & so on. I will try it when I have more time (in a week or so, hopefully) Edricson 17:09, 14 January 2007 (UTC)
 * "Consonant mutation" is a general term that can be used for any of these phenomena, though of course scholars of each language have tended to use different terms as traditional in their fields. As for what English is doing here, just read the section. I don't know if it's ever been called by the name "mutation", but that's what it is nevertheless. —Angr 17:33, 14 January 2007 (UTC)
 * Whoops, the comment about English was totally supefluous: on re-reading the bit I though I had deleted it, but it turns out I hadn't. Sorry about that. I agree that that's what it is, though a quote from an authoritative source (or a textbook) could come in handy. Still, a comment on these traditions is clearly in order: for instance, in Africanist parlance, the term is more often "consonant alternation", with "consonant mutation" reserved for stem-final alternations in Bantu before the reflexes of super-high vowels (you will find that Cheryl Zoll (Consonant mutation in Bantu. LI 26, 1995), for instance, uses "consonant mutations" to refer to the latter without any qualification). Edricson 18:14, 14 January 2007 (UTC)

New book about mutations
In my opinion the term mutations should be limited to i n i t i a l  consonant alternations. That's what's usual. They also occur in some Otomanguean languages (in Mesoamerica), in Tibetan, and in many more languages. I'm about to publish a whole book about the topic called "Reconstructing the mutation system of Atlantic". Anyone who wishes to be informed when the book is out, please send me an e-mail and you will be taken up on a list: . Jan Henrik Holst. —Preceding unsigned comment added by 134.100.172.24 (talk) 17:32, 30 September 2008 (UTC)
 * The article was originally only about initial consonant mutations, but I saw no reason for it to be limited in scope when there are languages that have mutation in other positions in the word. —Angr 18:49, 30 September 2008 (UTC)

Well, the point is that linguistics has its terminology, which is fixed here. The term "mutations" simply implies that it's at the beginning of words. The term for such alternations anywhere in the word is "consonant alternations". One could write a Wikipedia entry about that, of course. Jan Henrik Holst —Preceding unsigned comment added by 77.186.222.151 (talk) 18:43, 10 November 2008 (UTC)
 * Some linguists may assume this narrow definition, but the usage is not fixed. For example, Slavicists often use the term mutation for alternations involving stem-final consonants. CapnPrep (talk) 05:45, 12 November 2008 (UTC)

Husas, houses
d [s], while hūsas 'houses' and


 * afaik, the plural of hus in Old English was hus, not *husas. An error met-with occasionally. 131.220.251.26 (talk) 09:40, 3 December 2008 (UTC) Wojciech Żełaniec
 * I'll double-check this when I get a chance. —Angr 10:41, 3 December 2008 (UTC)

Dubious sandhi
Remoov'd section:
 * Nivkh: stops become fricatives after vowels, and fricatives become stops after other fricatives. Examples:
 * 'neck', 'neck of a reindeer'
 * 'shoot', 'to shoot a bear'

See main article. Apparently the current understanding is that 1) there is no fricative → stop change, and 2) the process is not entirely phonologically conditioned. It may be better to take this away until we figure out if we can still call this sandhi, or not. -- Trɔpʏliʊm • blah 17:04, 1 September 2009 (UTC)
 * I believe the "remooved" is very particular. Speling12345 (talk) 6:59, 27 December 2013 (UTC)

Merge with Sound change

 * Oppose. These are separate topics. CapnPrep (talk) 13:06, 18 December 2009 (UTC)


 * Sorry, I didn't even mean merge in the classic sense, just a reorganization. See the rationale below (it took a while to express my doubts) No such user (talk) 13:30, 18 December 2009 (UTC)
 * Oppose, since as the proposer of this said, this isn't even supposed to be a merge. Kennercat (talk) 19:20, 16 September 2010 (UTC)

Article reorganization
Background: this morning I started enthusiastically broadening the concept of Devoicing to Consonant voicing and devoicing, but I found that the organization of articles (and partly of concepts) outlined in Template:Sound change is rather messy. The problem is that the whole concept of sound changes and alterations is pretty multi-dimensional. So I contributed to the entire mess, I suppose.

For the start, I don't like the approach which is half-heartedly taken in this article. Apparently, it started describing stem-initial mutations as in Celtic, but changed direction in the course, trying to encompass all alternations, but ended up non-systematically. For example, Russian alternations (first palatalization, second palatalization, iotation) are not usually described as "mutations".

I feel that the whole thing needs some restructuring, but I'm not competent enough to tackle it. I propose something like:
 * Sound change should focus on diachronic sound changes
 * Alternation (linguistics) should focus on synchronic sound changes, with the list of most widespread ones.
 * Consonant mutation should be renamed and rescoped back to include just initial mutation, (as oposed to Sandhi).

If someone has another proposal how to organize the articles so that they have relatively clean scope and organization... No such user (talk) 13:28, 18 December 2009 (UTC)


 * Well, let's see. The Dholuo facts can go into Consonant voicing and devoicing. (Initial consonant voicing is a rare enough phenomenon that it should probably be merged there too rather than being a separate article.) Fula can stay in an article about initial consonant mutations. Rendaku can be mentioned at CV&D, though it has its own article. But I don't know where to put the Hebrew, Russian, Indonesian/Malay, and Ute facts if this article is to be restricted back down to initial consonant mutations. +Angr 16:26, 18 December 2009 (UTC)

the Finnish question again
So I just put back a section on consonant gradation, and now I see the discussion here on the old removal. I'm no expert on whether the process can be swept into the closet of phonology in every case, but it seems a mention of it certainly belongs here; if in Estonian it is morphologically conditioned (even if only under shallow analyses) then that's all the justification one needs; even if not, it would be worth pointing out how it's phonologically conditioned and why that makes it a non-example, if that's the tack we're taking. In any case, Consonant gradation opens with the assertion that
 * Consonant gradation is a type of consonant mutation [!]

Does it really really not belong? (And FWIW, in my impression these sort of phenomena seem prone to getting different names in different linguistic traditions, and it would be totally misguided not to mention a process here 'cause its traditional name is something else, if that's anyone's thinking.) 4pq1injbok (talk) 01:37, 6 September 2010 (UTC)
 * Generally speaking, it certainly is an applicable case even if Finnish itself may still be fairly straightforward about the conditioning. Samic languages have generally obscured the conditioning just as bad (if not worse). Gradation might be workable into an example of how consonant gradation may develop from phonetical into phonological into morphophonological. -- Trɔpʏliʊm • blah 00:13, 7 September 2010 (UTC)

Hebrew example
The "v" in "tov" is a bheth; it's spelled טוב. It just happens not to mutate in the form given. Maybe a better example would be קו ("qav", line). phma (talk) 05:36, 11 February 2011 (UTC)
 * It doesn't matter whether it's spelled with vet or vav; the point is just that there's no alteration between /v/ and /b/ in /tov ~ tovim/, while there is alteration in /dov ~ dubim/. —Angr (talk) 17:43, 12 February 2011 (UTC)