Talk:Consonant cluster

From Wikipedia, the free encyclopedia

I would like to kno about the noraml acquisition of consonants clusters. What are we supposed to expect at 2 years of age? (English)


--- How is rhythm a six-part consonant cluster? The y is a pronounced as a vowel. The way I hear it, there are two consonant clusters in "rhythm", "rh" and "thm".

Contents

[edit] rhythm has a vowel

rhythm in english is a 4 (or 5) phoneme, 2 syllable word [rIDm] (sorry, can't remember my SAMPA), there is most definitely a vowel.. [r] [I] [D] and a syllabic [m] (or schwa + m if you subscribe to that line of thinking).

[edit] georgian/slovak type clusters

i'd like to rewite some of this to be less misleading, because for example in slovak, r and l can be vocalic, so a word presented like štvrťžblnknutie may appear to be only 2 syllables (because there are 2 vowel nuclei), but in fact it is 4 syllables long: štvrť.žbln.knu.tie, where the r and the l are syllabic, so why is the word presented as an example of a 'formidable consonant cluster'? each syllable has at most a 3 consonant onset, which is hardly formidable.

also, the statement: the Georgian gvbrdγvnis is an example of this type, containing four syllables, but only one vowel...

what is the syllable breakdown? gv.brd.γv.nis? do not these syllables have a vocalic nucleus? if so, that means 3 of those consonants aren't parts of the clusters.... which according to the syllable breakdown i've written, there aren't any clusters at all if you look at it on a per-syllable basis.

I guess this merely means that /v/ can be vocalic in Georgian -- and, yes, this probably means in turn that there are no consonant clusters in any syllable.
George Hewitt, professor of Caucasian languages at the University of London, informs me that the consonants of Georgian cannot appear as syllable nuclei... that would seem to make the Georgian word a monosyllable, and therefore quite scary. :) I don't see how this is even possible from an articulatory perspective, but my experience of Georgian is extremely limited. I'm going to change this in the article, since Hewitt's published reference grammar ("Georgian: A Structural Reference Grammar") backs this up. thefamouseccles 02:05, 9 Sep 2005 (UTC)
What does it really mean that consonants can't appear as syllable nuclei? Surely they do so phonetically? Is Hewitt talking about some more abstract analysis? David Marjanović 16:14, 16 June 2007 (UTC)

and the importance of cross-syllable clusters? are there any references for this? i've never heard mention of this before.

Exit 3 July 2005 17:59 (UTC)

[edit] do not confuse orthography with phonetics

i removed the following lines:

The longest consonant cluster in an English word appears in latchspring. The longest consonant cluster anywhere, eight in total, appears in German Angstschweiss (agony sweat).

Certain loan words, like Nietzschean (supporting philosopher Friedrich Nietzsche's ideas), which has five, may have even more.


... latchspring may contain 5 consonants in a row, but only 4 spoken across 2 syllables.. [latS.sprIN]...

same thing with the german Angstschweiss ... sch in german is phonetically 1 consonant [S]

and if we are talking about the english pronunciation of Nietzschean, well, it would be pronounced [nitSi@n]... one consonant for all of 'tzsch' in the middle of the word

also, although czech vlk (wolf) doesnt contain any written vowels, the [l] is vocalic.

Exit 8 July 2005 03:44 (UTC)

That's like saying english word /wrd/ has no vowels.

Or that ancient Hebrew was pronounced with no vowels.

[edit] Definition

Currently the article defines a consonant cluster thus "a consonant cluster is a group of consonants which have no intervening vowel." Compare this to the definition given here. "A consonant cluster is a group or sequence of consonants that appear together in a syllable without a vowel between them." (emphasis added). Which is correct? I believe it's the latter. I'm about to edit the article accordingly.

I also noticed the following "Japanese is almost as strict, but it allows clusters of n + consonant: Honshu, the name of one of the major islands of Japan, is an example." Yeah, the major island but ... if we keep in mind the above correct definition, we notice something: these groups of /n/ + consonant are not consonant clusters at all because they cross a syllable boundary. However, this is not to say the consonant clusters cannot occur in Japanese. The language does allow consonant plus /j/ as in Tokyo.

Jimp 04:39, 18 January 2006 (UTC)

I think there is some basis for the first definition, because there are some languages which allow particular sounds word-finally, but not before another consonant. I think the native vocabulary of Finnish is an example. —Felix the Cassowary 05:07, 18 January 2006 (UTC)

Yeah, I hadn't read the bit where the article mentions that there is disagreement over which definition is better. This info had been placed way down towards the bottom, though. I've moved it up into the intro where it belongs. Even so, that Japanese example is still a little off considering that the [n] in Honshu is a syllabic consonant and therefore phonologically a vowel. Jimp 06:12, 18 January 2006 (UTC)

No, it isn't. The [n] in Honshu isn't syllabic; Honshu has only two syllables in Japanese. It has four moras, and moras are arguably more important in Japanese phonology (and certainly poetry) than syllables, but still only two syllables. I think the definition of consonant cluster as any string of consonants regardless of syllabification is best to start off with, since there are languages (Bella Coola) that have been argued not to have syllables, and since syllabification is usually a matter of theoretical opinion rather than objectively observable fact, meaning we tread into the realms of non-NPOV and OR when we start talking about syllable boundaries. And finally, there are indeed many languages that allow more complicated clusters at word edges than at word-internal syllable edges. --Angr (tɔk) 06:20, 18 January 2006 (UTC)
I'd say the kana of "tsu" and "chi" has consonant clusters, but that might be me... 惑乱 分からん 16:04, 5 August 2006 (UTC)
No, they're affricates, which are single segments, not consonant clusters. The famous minimal pair in English is why choose with an affricate versus white shoes with a cluster. In Polish they constrast even within a word: czy has the affricate [t͡ʃ], while trzy has the cluster [tʃ]. User:Angr 16:35, 5 August 2006 (UTC)
OK, so that's it, thank you! 惑乱 分からん 20:21, 5 August 2006 (UTC)

[edit] The Georgian word

I might be wrong here, but I'm almost entirely sure that "he is plucking us" is actually gvprcʼkvnis in Georgian. I'll try to find out which word is correct and then maybe correct the entry if no one objects or can certify the word in the entry. I think I'm going to include Georgian spelling for this word as well... —N-true 18:12, 1 May 2006 (UTC)

Actually, gvprckvnis (no ejective) is "he's peeling us". The root to pluck is given by brdɣvn-: gvbrdɣvnis he's plucking us. See George Hewitt's 1995 Georgian: a structural reference grammar, page 20. Thefamouseccles 22:51, 31 July 2006 (UTC)

[edit] Polish consonant clusters

In Polish 4-consonant clusters are also not uncommon (e.g. Strwiąż, wstręt, pstry, wstrzymać, wzgląd, wstrzelić, oszczerstw), 5-consonant clusters also sometimes occur (e.g. źdźbłko). Also, you only show examples of clusters within a single word (although I believe Angstschweiß should be counted as 2 words, which in German can be in many cases joined together). If you counted inter-word clusters, you would find even 9-consonant examples in Polish (e.g. słuchać "oszczerstw z wstrętem", wszedł "Herbst z pstrągami")

[edit] Angstschweiß

"The German word Angstschweiß (cold sweat) is another good example."

Uh... and how would that be? 203.154.48.179 13:53, 1 July 2007 (UTC)

[edit] Why Choose White Shoes?

Above, User:Angr mentioned the difference between two distinct consonants put together into a cluster (even if the cluster is bridging two syllables, rather than being contained within a single syllable -- I'll leave that a separate issue!) and an affricate, which is generally percieved and pronounced as a single consonant. Contrary to IPA renderings, the English "CH" or "TCH" is (almost?) never pronounced as a "T" followed by an "SH": The word "chew", for example, isn't pronounced in a way that could be equally well rendered as "tshew"; "why choose" and "white shoose" are not homophones, nor are "catch it" and "cat shit" even though "catch up" and "ketchup" are. While most examples of consonant clusters counting affricates as more than one consonant have been removed from the article, I think it would benefit the article to directly address the subject. One thing in particular that I'd like to see addressed is the fact that if one draws out the "CH" sound, it has a sharp, almost tinnient sound compared to a drawn out "SH" sound, as the "CH" is made with the tip of the tongue (with the supposed "T" sound really being incidental to its pronunciation) and the "SH" is made with a much broader area of the tongue contacting the roof of the mouth further back. To be honest, I can't see why IPA stopped using ligatures for the English "CH" and "J" in favor of two seperate symbols for each, when even the ligatures conveyed the (completely incorrect) idea that the English "CH" and "J" each contain more than a single consonantal sound quality, thus calling for exactly the opposite sort of replacement: losing each of the two ligatures with its own, single, distinct symbol. Why a single character is used in IPA to distinguish a combination of two totally different sounds in the case of [ɲ] (made up of an [n] followed by a very distinct [j]), but two characters are needed for a single sound in both the case of the single sound misleadingly rendered [tʃ] and the single sound misleadingly rendered [dʒ] is absolutely inconcievable. The only reason I could possibly think of for this is that the usual pronunciation of these consonants throughout almost the entire Anglosphere radically differs from mine own native dialect of Cascadian English, in which case there would be unique consonants in Cascadian English not represented at all in the IPA. --Þórrstejn [ˡθoɝ.staɪʲn]: Hammer of Thor talk 09:59, 19 August 2007 (UTC)

  • Um, [ɲ] is not "a combination of two totally different sounds ... (made up of an [n] followed by a very distinct [j])"; it's a single sound with a single, palatal, place of articulation. It doesn't occur in English, so don't be thrown off by statements seen in introductory language books that claim that [ɲ] is like the ny in canyon; it isn't. Are you aware of any published literature discussing the inadequacy or otherwise of transcribing affricates as sequences of stop + fricative? —Angr 10:24, 19 August 2007 (UTC)
I appreciate the correction on [ɲ]; when I referred back to the IPA chart and articulated the sound myself, I found it was very different from an [n] followed by a [j], though, and I'm willing to admit this could very well be an error in pronunciation on my part, I could not make the sound without noticing something very similar to the English affricates: I couldn't make the sound without another consonantal sound (in this case an approximant) following, however faintly, in the wake of the nasal voicing (for want of a better term). Unfortunately I'm not aware of any literature describing the inadequacy/inaccuracy of transcribing affricates as sequences of stop + fricative, or else I'd have been bold and added the information myself, as it seems relevant to the subject of consonant clusters. The relevancy, as I see it, would lie in the affricate being an essentially condensed consonant cluster itself. I remember at one time, when the ligatures were used, it wasn't at all uncommon to see [that bridge-like diacritic, the name of which momentarily escapes me], which usually denotes a consonant sound articulated in a spot between the characters that make up the ligature, which (being the amateur that I am) seemed to me to indicate that those who used this convention considered the English affricates to be singular consonants and not stop + fricative sequences. Is this a misinterpretation on my part? If not, surely not every phonologisthas condoned the change in convention from single consonant represented by ligature and diacritic to a sequence of two consonants. Though I'll concede that the English affricates do contain two audible sound qualities, it doesn't seem to me this is any truer of the affricates than of some other consonants represented by single characters (not sequences nor even ligatures) on the IPA chart. Like diphthongs, many consonants sound very differently at the beginning of the articulation from how they sound at the end; were this untrue, all consonants would sound essentially the same when recorded and played backward, and not many do. The main problems I see are, firstly, the inconsistency when using ligatures/sequences for some consonants and not for others (not so much in the fact of using them for some and not for others, but the seemingly arbituary manner in which this is applied), secondly, and probably least relevantly, that in the case of the English affricates, the "stop" portion does not come to a full stop, but blends seemlessly into the "fricative" portion, and thirdly, being to me most importantly of all, that, even were a stop + fricative sequence to be used to represent the English affricates, the both the "stop" and "fricative" of each affricate are pronounced differently, with, at least in some dialects (or perhaps just a noticeable enough number of idiolects to call attention to itself) a different enough place of articulation to merit their own place on the IPA charts, or, in what seems to me the majority of cases, not quite a different enough place of articulation to merit such treatment but to be mentioned in the phone's name (a good comparison to illustrate what I mean here is how denti-alveolars don't merit their own position on IPA charts, but merit a difference in name), and a very different manner of articulation, from the characters used in sequence to denote them; the "[t]" part of the English "CH" isn't articulated in place or manner quite the same as the [t] of the English "T", nor the "ʃ" part of the English "CH" at all in manner or place like the "ʃ" in the English "SH". I doubt this is unique to Cascadian English, as I've never seen it mentioned in reference to the dialect. I can't imagine that no phonologist has ever observed these differences in pronunciation; that the ligature with the bridge-diacritic was thrown around so loosely and arbitrarily that those phonologists who once used it all now unanymously agree that each English affricate is just a sequence of other phones heard in the language. Mainly, I was hoping someone else, with more knowledge on the subject of the history of the IPA representations of English affricates than I, might have something to add on the matter, or at least be able to point me in the direction of published literature on the subject if said someone didn't feel like doing it himself. --Þórrstejn [ˡθoɝ.staɪʲn]: Hammer of Thor talk 12:08, 19 August 2007 (UTC)
P.S. Please forgive my relative ignorance on the subject of phonology, my main interest in it comes from a personal conlanguage-family project and an obsession with trying to figure out how on earth the Celts managed to turn a C/K/Q into a P (akin to my parallel obsession with trying to imagine how the hummingbird evolved from an insect-eater; both of which are easy to describe in a stepwise process on paper, but extremely difficult to envisage occuring practically in the real world). --Þórrstejn [ˡθoɝ.staɪʲn]: Hammer of Thor talk 12:26, 19 August 2007 (UTC)
By "the bridge diacritic" I assume you mean the symbol often used thus: [t͡ʃ]. That symbol hasn't been deprecated in the IPA; only ligatures like [ʧ] have been. The "tie bar" symbol isn't used much for English, because [tʃ] is mostly unambiguous without it (the difference between white shoes and why choose can be shown in transcription by separating the syllables with a period, or by separating the words with a space); and it's not used much here at Wikipedia because most people see IPA transcriptions in Arial Unicode, which has a bug in the "tie bar" symbol, making it appear in the wrong place. However, in other languages, using the "tie bar" is important: Polish, for example, has a contrast between the cluster [tʂ] in trzy and the affricate [t͡ʂ] in czy. You're right that the stop portion of the English [tʃ] affricate is not made in exactly the same place as the stop [t] in isolation, but it does fall within the range of the /t/ phoneme's allophonic variation. I think when I say white shoes the [t] of white is in pretty much exactly the same place as the stop portion of the affricate when I say why choose: the [t] assimilates in place to the following [ʃ]. As for Celtic, it wasn't a straightforward [k] that became [p], it was a labialized [kʷ]. Since that sound already has a labial component, it's not so far-fetched that it could change to [p] over time, especially since the Celtic languages had no native [p] at the time. Languages' phonologies like symmetry, and an asymmetric system like that of Proto-Celtic, which had a [b] and a [kʷ] but no [p], was inherently unstable. —Angr 13:16, 19 August 2007 (UTC)

[edit] Tendency for some consonant clusters to appear more frequently?

I've been wondering this for a long time: why do "tr", "dr" appear a lot more often than "tl", "dl" in European languages that I've seen (Germanic and Romance mainly)? "tl" and "dl" appear mostly between syllable boundaries, and are rarely ever present in initial and final positions (excluding the syllable "l" right now), while "tr" and "dr" are common enough regardless of whether the "r" is an alveolar approximant, trill, or even an uvular fricative. Similarly, "sl" is more often than "sr", and "vr", "vl", "zr", "zl" are all quite rare (except for "vr" in French)? What quality governs the probability for a certain cluster to appear and to remain in a language?

Keith Galveston (talk) 07:09, 16 April 2008 (UTC)