Talk:Retroflex consonant

Headline text

Whilst I am sure this article is meaningful to some, could it not be explained in simpler language. It's rather too technical for my understanding. Should it not instead contain a statement like: pronounced with the tongue touching the palate. Rellis1067 22:00, 28 July 2005 (UTC)

I agree. The language of this article assumes the reader already knows a lot of detail about oral anatomy and how it applies to articulatory phonetics. These people already know what a retroflex consonant is, so this article isn't much use to anybody in its current state. I will try to tone down the technical language. Nohat 23:22, 28 July 2005 (UTC)

That's great! The other thing I don't understand is the caption: Sublaminal retroflex plosive. It would be clearer - to me at least - if someone could reword it, making it less technical. Rellis1067 19:03, 31 July 2005 (UTC)

I changed it from 'sublaminal' to 'subapical' to match the description of the article. If you still have trouble, then we need to work on the article some more, not just the caption. kwami 23:07, 2005 August 15 (UTC)

I think it's pretty understandable if you have some knowledge of phonetics. The article was very very useful to me when I found it. I was just wondering if someone could maybe put up an image of the articulation of the retroflex fricative comparing it to the alveolo-palatal and palato-alveolar fricatives'. Radek

My edit

I removed the Mandarin example for [ʐ] and replaced it by Russian and Polish. Firstly, Pīnyīn r isn't [ʐ] but something more complicated (usually it's simultaneous [ɻ] and [ʐ]*, though some pronounce it as [ɻ] alone); secondly, the transcription was quite broad anyway; thirdly, it increases the diversity of the examples (there is already a Mandarin example, but no Slavic one); fourthly, it shows people at a glance that retroflex consonants are also common in less exotic** languages.

(*) Analogous to Czech ř, just with "American" [ɻ] rather than "Spanish" [r].

(**) From a western European POV – which most readers of the English-language Wikipedia will inevitably have.

David Marjanović | david.marjanovic[at]gmx.at | 2005/11/19 | 16:34 CET

That's fine, though I wonder about your description of Mandarin r, since simultaneous [ɻ] and [ʐ] (an approximant and fricative at the same place of articulation) isn't physically possible. Do you mean it's [ʐ] with some sort of secondary articulation? I can pronounce it, but I'm not quite sure what I'm doing! kwami 15:53, 19 November 2005 (UTC)

I thought "simultaneous" meant "at the same time". You could describe the sound as [ʐ] with a secondary articulation or as [ɻ] with a secondary articulation. I don't think it makes much sense here to tell apart which articulation, if any, is primary and which secondary. (Of course you're right that two places of articulation are involved.)

It's impossible to pronounce [ʐ] and [ɻ] at the same time unless you have two tongues: You can't have the tongue close enough to the roof of the mouth to produce turbulence, yet at the same time have it too far away to produce turbulence at the same location. You're claiming coarticulation with a single place of articulation - a contradiction in terms. kwami 23:18, 5 January 2006 (UTC)

It is, however, true that the thing is pronounced [ɻ] or [ɹ] or both either by some people or at some occasions (I haven't listened enough, but probably Peter Isotalo is right in saying it is the word-final allophone) or both.

David Marjanović | david.marjanovic[at]gmx.at | 2006/1/5 | 23:50 CET

According to San Duanmu in The Phonology of Standard Chinese, the initial /r/ is not retroflex, but rather an [ɹ] of some sort. He bases this partially on very well-founded skepticism concerning the existance of [ʐ], a voiced fricative, as an allophone in a phonology that has doesn't have any voiced/voiceless phoneme pairs. I think the only allophonic exceptions are in words like bàba ("father") where the second consonant is voiced because its between vowels. Whatever the Mandarin initial /r/ is, it is not the same as the [ɻ]-suffix which forms the rhotic final in words like háir ("child") or ménr' ("door").

I will reserve myself for perhaps missing some essential explanation of the nature of the sound, since the book won't be available to me again until late January.

Peter ^Isotalo 17:10, 1 January 2006 (UTC)

Best to leave it out then. And you're right, it's not like the [ɻ]-suffix. It's definitely voiced, though, and is quite fricated. Maybe not as fricated as sh, but then voiced fricatives tend to be less fricated than voiceless ones. Think of English thy vs. thigh - that's the difference in quality I hear between Mandarin r and sh. I never spoke the language well, though, so my opinion can't count for much.

Philosophic arguments based on an ideally symmetrical phonemic inventory aren't very convincing. There are just too many irregularities in the world's languages, so we need laboratory measurements. Mandarin r seems to have historically been some sort of nasalized fricative; perhaps a originally a palatal nasal where palatalization became fricated. An odd history for a phoneme like that could easily make the inventory asymmetric. kwami 17:44, 1 January 2006 (UTC)

Gotta revise my statement after reading San more carefully. He uses the [ɹ] for narrow transcription ([r] elsewhere), but when he analyzes the sound, he defines it as retroflex and doesn't make a distinction between the suffix, the /i/-allophone and the initial. He notes that there are those who transcribe it as a [ʐ], but that they are few and then goes on to cite studies that show that the sound has little or no friction. So actually transcribing it as a fricative seems rather misleading to me, especially considering that the suffix-/r/ never has friction. So it's more than just a philosophic argument.

Peter ^Isotalo 20:16, 4 February 2006 (UTC)

Isn't Mandarin initial "r" a voiced apical flat sibilant? That would make it different from the Polish/Mandarin laminal postalveolar s, and perhaps similar to the Hindi sound. A voiced apical flat sibilant would be a lot like [ɻ] though more sibilant, and like [ʐ] though apical rather than sub-apical. That sounds like David Marjanović's description as "simultaneous [ɻ] and [ʐ]". And, by the way, Taiwan Mandarin coronal fricatives sound quite different from Beijing ones, which leads me to guess that there are a lot of variations across the Mandarin-speaking area.

American English

Shouldn't it be mentioned that there are allophones of some consonants preceeding retroflex approximant :

• retroflex nasal: corn, earn, morning
• voiceless retroflex plosive: part, sort
• voiced retroflex plosive: sword, third
• retroflex approximant: road, red
• lateral retroflex approximant: girl?
• retroflex flap: party, thirty, forty

Is it true that lateral approximant becomes lateral retroflex approximant under the influence of retroflex approximant? --Dennis Valeev 14:48, 11 December 2005 (UTC)

Yes, that's a good illustration. It may be pretty opaque to native speakers, but the more ways to get an idea across the better. I doubt it's a universal feature though: not all speakers of rhotic dialects have retroflex ars. People in the same family may differ, as there are at least three articulations that produce the same acoustic effect. I don't know if they all have the same articulatory effect on following consonants, even if the acoustic effects are the same. It could be very confusing to someone to tell them that GA /r/ is retroflex if their articulation is otherwise.

As for the el, it seems slightly retracted to me, but not retroflex like the others. Girl is sesquisyllabic, as all syllables with a diphthong followed by el are, and the slight break between the r and the l is enough to allow rearticulation. (That's one argument for /i/ and /u/ being diphthongs in GA English: peel is a syllable and a half, while pill is a single syllable.)

kwami 19:56, 11 December 2005 (UTC)

Yeah, I agree that there are at least three kinds of Rs out there: alveolar, retroflex, and "bunched" as it were. I was surprised by the fact that I couldn't quite force myself into articulating "lateral retroflex approximant" (I wish I had an auto-replacing soft running on this computer; do you have any idea of good resident substitute for a MS Word with its autoreplace feature?); because I heard that distinctive "click", which can be easily reproduced by saying "tl" together. Do you have an IM? --Dennis Valeev 20:07, 11 December 2005 (UTC)

Are 'd' and 't' in "drink" and "tree" retroflex for you? As far as I am concerned, I start from alveolar though a bit retracted "d" and "t"; I find it very hard to start from retroflex coronal consonants in such preconditions as "in the" (where 'n' is dental), as in "in the truck", because there's such a huge gap between dental position of your tongue and retroflex one, though I have no difficulty in pronouncing "through", for example. --Dennis Valeev 20:43, 11 December 2005 (UTC)

Sorry, I don't know what 'IM' is.

I pronounce tree and drink as you describe them, with my tongue retracting to pronounce the r, but with some anticipatory retraction in the plosive.

I don't follow the other question. I have no problem making the /t/ of truck retroflex in the phrase in the truck, though of course I would never pronounce it that way. kwami 21:06, 11 December 2005 (UTC)

Oh, I see, I can't pronounce "in the truck" with retroflex "T" either, although I read in a book on linguistics that it is, in fact, retroflex before "r". If you have a google account you can search for this book: "Essential Introductory Linguistics" by Grover Hudson, or access this page and loading the aforementioned book on page 48: http://print.google.com/print?q=consonant%20retroflex%20american

It says that "coronal consonant /d, t, s/ is retroflex before [r]: ʈruck, ɖrink". What? --Dennis Valeev 23:05, 11 December 2005 (UTC)

Maybe in his dialect it is. Not in mine. kwami 05:49, 12 December 2005 (UTC)

Perhaps he just means 'non-palatal postalveolar'? I could see the /t/ of truck being postalveolar. kwami 06:29, 21 December 2005 (UTC)

Exactly, this is the way I used to pronounce "truck" but he deliberately used retroflex T down there, so what was that for? And he's definitely an expert in his field, but all the same I don't see how it's possible to easily articulate this, no matter how many people I enquired they answered something along the follwing lines: "no! that's absolutely impossible, the pal must be nuts". Yet, I have another question pertaining "retroflex taps". Is there any tap here: "ask her to be here" or "it's similar to the..."(there is the retroflex tap in "where did you go?", as far as I'm concerned). Thanks! --Dennis Valeev 22:07, 23 December 2005 (UTC)

Definitely a flap in "ask her to be here": I retract my tongue for the /r/, then release it as a flap for the /t/. The retraction is pretty retroflex, but I don't think the flap is: it is released from a retroflex position, but the contact appears to be alveolar. It's a little hard to be sure what's going on with such a brief sound, though. kwami 22:16, 23 December 2005 (UTC)

Yup, that's how I was saying this, too; I think it has something to do with "her" word, because there are so many people who don't care to say both 'h' and 'r' carefully in it, so they end up saying [æskɚɾəbihiɻ] or even [æsəkɾəbihiɻ], but when it comes to saying something more complicated as "it's similar to..." or "where did you go?" the tap seems to be completely of retroflex nature.

Both /r/s are the same for me, though it's harder to tell what's going on with similar to because the word's so much longer. Doesn't seem any more retroflex to me than her to did. And where did you go? is palatalized [wεɹʤuɡo]. So no, I don't appear to retroflex any of them. kwami 23:35, 23 December 2005 (UTC)

As for the "where did you go?" example, it's [ʍɝɻɽəʤuɡɔʊ] to my ear, though I've got a tin ear problem, and need some help myself *smiles*. --Dennis Valeev 23:57, 23 December 2005 (UTC)

I thought it was [weɜɻɾɪdʒəgoʊ], with [eɜ] being a diphthong. ionas68224|talk|contribs|email 15:39, 30 July 2007 (UTC)

Some clarification required regarding Retroflex consonant in English & Hindi

In article page, Hindi `t' in `tapu' is called as Retroflex consonant. But it's pronunciation is same as `T' found in pronouncing Tea , Train , Total etc. So, can we call that even English contains Retroflex consonant ?

I am an Indian and knows Hindi like mothertongue. WIN 12:04, 13 December 2006 (UTC)

No. Maybe in Indian English, but certainly not in other dialects. Ƶ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 12:14, 13 December 2006 (UTC)

Words like "tea", "train", "total" are pronounced quite differently by Indians and Americans. The Indian "t" is retroflex and unaspirated; the American one is alveolar and aspirated. —Preceding unsigned comment added by 129.49.74.185 (talk) 15:24, 12 February 2008 (UTC)

Retroflex, postalv, alv-pal. et cetera

What is the difference between a "postalveolar consonant", a "retroflex", a "palato-alveolar consonant", and an "alvoelo-palatal consonant"? They seem all the same, further back than the alveolar ridge, and more forward than the palatal region. That is, to say, why is there a differenced between [ʃ], [ɕ], and [ʂ], or their voiced equivalents? Also, is the word palatal pronounced [ˈpæː.lə.təɫ] or [ˌpə.ˈleɪt.l]? bibliotheque (Talk) 09:05, 9 July 2007 (UTC)

• Reply: No theory is flawless. One of the weaknesses of Place-and-Manner theory is that 'retroflex' is neither a place nor a manner of articulation. It is, rather an aspect of articulation. Thus, in the IPA chart we can have a series of plosives, nasals, fricatives etc in the same part of the roof of the mouth where we have the label "retroflex". If we use the blade or the front of the tongue and make a complete closure and a sudden release, this is a normal plosive (e.g. [c] or [ɟ]). However, if we use the area beneath the blade of the tongue, i.e. we curl the tongue, then it is a retroflex (e.g. [ʈ] or [ɖ]). So if we want to redefine a retroflex it is, then: "a palatal sound with a sublaminal active articulation" 9abdulla —Preceding unsigned comment added by 85.154.153.131 (talk) 10:09, 29 March 2008 (UTC)

Swedish/Norwegian example

I fixed the bungled explanation of Sveco-Norwegian retroflexes: it's rC sequences that give rise to them, not Cr seqs, as you indeed can tell from the example already given: Martin.

"used in some languages"

Do we have to use the line "used in some languages" in every article about a consonantic or vocalic sound? If that particular sound weren't "used in some languages" we wouldn't have an article about it... --Taraborn (talk) 13:16, 31 January 2008 (UTC)

Phonemic vs. allophonic

Wikipedia is written for a fairly general audience. Much of what has been written in the article is getting quite technical. I think it would be more helpful to the general audience if there was more about how consonants pronounced with retroflexed articulation contrast with non-retroflex consonants in a number of languages. The details of allophonic retroflexion are true, but may not be what most readers want to know. Pete unseth (talk) 21:24, 20 April 2008 (UTC)

I disagree. I (and possibly most people) read articles such as these in hopes of improving our pronunciations of foreign names (and maybe learn a foreign language). In this case allophones are everything. A precise, concise, and jargon-free statement of the "articulatory gesture" (in Ladefoged's sense) is what such readers seek. Examples in several languages are interesting and helpful (it would be nice to have an .ogg of a careful native speaker for each example, but that might be asking too much). — Solo Owl (talk) 16:46, 13 January 2011 (UTC)

Help on denoting retroflex consonants

Is there any way to specify secondary and tertiary aspects of a retroflex consonant? For example, I need to denote a sub-apical retroflex nasal that is pronounced in the palatal region, as opposed to the postalveolar region. As far as I can figure out, the best I can do is use the IPA symbol for retroflex nasal, and perhaps add the IPA symbol for palatalization, although I don't think that is quite right. Thanks for any help. —Preceding unsigned comment added by Jclu (talk • contribs) 22:25, 29 January 2009 (UTC)

See Talk:International Phonetic Alphabet#Help on IPA way of denoting a consonant for discussion of this question. —Angr 15:31, 30 January 2009 (UTC)

The Polish and Chinese consonants

Polish consonants sz, cz, ż and dż are NOT retroflex, while the Chinese zh, sh and r ARE. Please correct this section in the article. By the way, no Slavic language has retroflex consonants. —Preceding unsigned comment added by 193.71.38.142 (talk) 07:18, 3 February 2010 (UTC)

Polish sz and friends are laminal retroflex consonants, despite being often inaccurately labelled as [ʃ]. See Polish phonology. Moreover, ш, ж in Russian and other East Slavic languages are also retroflex. — Emil J. 11:48, 3 February 2010 (UTC)

Actually guys polish sz, cz and rz are subapical postalveolars not retroflex consonants — Preceding unsigned comment added by 24.47.6.147 (talk) 22:01, 27 July 2012 (UTC)

Subapical postalveolars are the prototypical retroflexes and don't occur in Polish. So you got everything wrong I'm afraid. Mr KEBAB (talk) 11:07, 31 October 2017 (UTC)

Strange characters

When I read the article I get ideographs beside some of the glosses -  and  - which are apparently from the private use area, which to me look like a dog's head and, I don't know, a pair of antlers? respecitvely. They seem to be from the private use area, so I'm guessing some font somewhere displays something else - what are these supposed to be? Moszczynski (talk) 17:33, 29 March 2011 (UTC)

What would help a layman...

A little background. I have always loved languages. I am a native speaker of Arabic and English. I have been dying to learn Hindi. Thanks to Wikipedia, I learned that Urdu would be a better place for me to start since I can already read the script.

However, I landed on the Retroflex consonant page when I could not read a few of the Urdu punctuation marks.

For me, the most glaring omission on the Retroflex consonant page is a small audio file so I can hear each sound. Telling me what it sounds like in a language I do not know (like Swedish or Polish) does not help. I was able to copy some words to Google Translate and listen to them there but that was of limited use.

I see that other articles have sound, video and animation so I do not think this would be a technical problem.

Just trying to help. — Preceding unsigned comment added by Ndajani (talk • contribs) 23:44, 1 January 2012 (UTC)

This article is most likely wrong

The article strikes me as simply inaccurate or at the very least mixing up phonetics and phonology.

Ladefoged & Maddieson (1996: 25–7) mention only subapical palatals and apical postalveolars as variants of sounds considered retroflex. Hamann (2004) says: "Unifying articulatory criteria for all retroflex stops are their apicality (the tongue tip is the active articulator), their place of articulation behind the alveolar ridge (i.e. postalveolar), and a sub-lingual cavity".

So I seriously doubt laminal post-alveolars and apical alveolars can ever be considered retroflex phonetically. Of course those sounds may be heard in place of phonologically retroflex sounds or sounds traditionally described as "retroflex", but I think this article should be written primarily from a phonetic point of view. Nardog (talk) 14:45, 26 December 2017 (UTC)

@Nardog: Apical alveolar sounds are definitely not retroflex. For flat postalveolars, see Hamann (2004) and Zygis (2003). Personally, I'd never consider them retroflex, but that's just me. Mr KEBAB (talk) 15:17, 26 December 2017 (UTC)

@Mr KEBAB: Thanks for the references. So am I correct in understanding that laminal postalveolar fricatives are sometimes described as retroflex to emphasize and distinguish the lack of the domed tongue shape present in the palato-alveolar fricatives, which are usually described/transcribed as postalveolar? Nardog (talk) 16:19, 26 December 2017 (UTC)

@Nardog: Yes (not trying to be picky, but it's not quite enough to call them laminal postalveolar. Palato-alveolar and alveolo-palatals are also laminal postalveolar. The 'narrow' term Ladefoged and Maddieson use is 'flat postalveolar'). But it's even more complicated: some scholars consider flat postalveolar [tʃ, dʒ, ʃ, ʒ] to be apical, but apical alveolar, not apical retroflex, and it's not the same apical alveolar articulation as in Spanish or Basque. The latter are more like [ts, dz, s, z], the former more like [tʂ, dʐ, ʂ, ʐ], but not genuinely retroflex, only auditorily retroflex, if you know what I mean.

I don't know whether they're usually described as postalveolar, but I do know that the term 'postalveolar' as well as the symbols [tʃ, dʒ, ʃ, ʒ] and [tʂ, dʐ, ʂ, ʐ] are ambiguous, but the ambiguity is usually solved by detailed descriptions of sounds or sagittal sections, which probably are the best way of showing the shape of the tongue. Mr KEBAB (talk) 17:04, 26 December 2017 (UTC)

@Mr KEBAB: Thank you for the summary. That raises a few questions:

• Can one call a flat postalveolar sound "retroflex" from a purely phonetic point of view? In other words, do those who treat flat postalveolars as retroflex do so merely for logistical reasons that depend on the language and its phonology, or would they call a flat postalveolar sound "retroflex" in any language?
• Is labeling flat postalveolars as retroflex limited to fricatives? If so, giving them an equal status with subapical and apical postalveolars doesn't seem fitting.
• Isn't treating apical postalveolars as retroflex something that came later? Perhaps they were, say, initially thought of as truly retroflex but turned out to be merely apical? And even if that was not the case, isn't the subapical articulation at least the canonical version of what a retroflex sound is?

In any case, the current version of the article seems to be doing an utterly poor job of explaining what retroflex consonants are. Especially the opening sentence doesn't tell you a damn thing about how a retroflex consonant is different from any other postalveolar sound, and definitely needs some revision. Nardog (talk) 08:55, 31 December 2017 (UTC)