Talk:Hindustani phonology

Untitled
It's looking good, Tuncrypt. I'm starting to switch from a bullet presentation to prose. kwami (talk) 01:28, 18 November 2007 (UTC)
 * Thanks, and thanks for the edits. The first part of my plan was to wipe out the script- and other non-related material. Now I'm adding in referenced bits from the sources I fortunately have available to me (Shapiro and Masica, plus Ohala on the net). Ohala's about used up and Shapiro is also nearing the end, though the Urdu article from the book Shapiro is in is quite massive, so there's more to come. It'll take a while, but yeah. Tuncrypt (talk) 04:07, 18 November 2007 (UTC)

Minimal pairs of retroflex flaps/plosives
The article says, citing Shapiro:
 * However, the adoption of English loans with alveolar stops, which are identified with Hindi/Urdu retroflex rather than dental stops (as with bat above), has led to the emergence of minimal pairs, thus conferring marginal phonemic status to the flaps.

Yet Masica, in describing the same phenomenon, specifically says that it does not create any minimal pairs. Does Shapiro provide any examples? Grover cleveland (talk) 15:19, 20 November 2007 (UTC)


 * Yeah. It gives one: "heD 'head' (of a department, organization, etc.) ~ heR 'flock' (regional)...". Can you give me the page number for Masica? Tuncrypt (talk) 16:19, 20 November 2007 (UTC)


 * Sorry, I think I misread Masica. He may only be claiming that the specific loanwords reDiyo and roD don't create minimal pairs, not that no such loanwords create minimal pairs -- it's a little unclear.  Here is the full quotation (pp. 91-92):
 * The common English loanwords reDiyo and roD, however, ..., introduce the phone D contrastively (albeit not producing minimal pairs) into the intervocalic and final postvocalic environments previously occupied exclusively by R, thus rendering the earlier allophonic distinction "phonemic", according to one criterion for deciding such matters. Grover cleveland (talk) 04:05, 21 November 2007 (UTC)
 * It might be a good idea to add the example (heD/heR) to the article. Grover cleveland (talk) 04:07, 21 November 2007 (UTC)

Status of /ʔ/
In this page /ʔ/ is not listed, but it would most certainly seem to exist in Urdu. I do remember reading that Hindi speakers, and even Urdu speakers in casual speech, drop /ʔ/ altogether, but I think it would still be good to discuss it. 67.165.249.115 (talk) 18:50, 21 November 2007 (UTC)

Also, I just noticed that /ʒ/ is not mentioned either, but it still seems present in Urdu though Hindi speakers may replace it with /dʒ/. I personally don't feel qualified enough to actually write these things into the article, but I'm hoping that I can just bring this to the attentions of those who are qualified enough. 67.165.249.115 (talk) 20:14, 21 November 2007 (UTC)
 * Do you have a reference for these claims? Cheers.  Grover cleveland (talk) 06:39, 22 November 2007 (UTC)


 * About ʒ I don't know but ʔ will be addressed. It is used in hypercorrect Urdu. Tuncrypt (talk) 15:54, 22 November 2007 (UTC)


 * Urdu does have listed as part of its alphabet. I believe I’ve seen it ‘transcribed’ in Hindi as झ़ once, but I can’t remember where, and most sources that mention  say it has no exact Devanagari correspondent anyway. Naim says it occurs in a few learned words borrowed from Persian and in the transliteration of certain European names, and gives its pronunciation as /ʒ/. The Urdu online dictionary at http://www.crulp.org/oud/ lists a handful of words with , though I myself can’t tell how common they are at all:  ‘dew, hail’;  ‘entangled’;  ‘eyelash’; and so on. Psi-Lord (talk) 21:20, 16 August 2008 (UTC)

The /zh/ sound definitely is a phoneme in Urdu though many approximate it with similar sounds. It should be included here. /?/ is not a real phoneme. (Sorry for my not being able to type IPA) Muhammad Areez (talk) 16:51, 27 January 2017 (UTC)

/ʂ/ and /ɳ/
It's my understanding that /ʂ/ and /ɳ/ are both found only in hypercorrect pronunciation of Sanskrit loanwords, and as allophones of /ʃ/ and /n/ when immediately preceding or following retroflex plosives. Currently /ɳ/ appears in the table of consonants as a marginal phoneme, while /ʂ/ is absent. Would there be any objection to adding /ʂ/ as another marginal phoneme? Cheers. Grover cleveland 07:06, 1 December 2007 (UTC)


 * ɳ is not an allophone of n like ʂ is of ʃ. Tuncrypt 18:15, 1 December 2007 (UTC)


 * And /ɳ/ and [ʂ] are not "hypercorrect" per se like ʔ in Urdu... they're just correct, in a standard sort of way. Tuncrypt 18:29, 1 December 2007 (UTC)

Table
Aeusoes. I guess there's no problem with condensing together Approximant and Lateral Approximant. But I don't understand your condensing of Dental and Alveolar, and your movement of Nasal to the top. Please discuss. Tuncrypt 18:24, 1 December 2007 (UTC)


 * Certainly. The official IPA charts I've seen put nasal as the first row and if I recall correctly, instructional guides say to put it as the top one.  I condensed dental and alveolar because there is no contrast between any consonant on just the dental/alveolar feature.  Also, since the dental consonants were already marked with a dental diacritic, it is clear which ones are dental and which are alveolar (if there are other dental consonants, a diacritic can go below this ones as well).  I also changed stop to plosive but I assume since you aren't asking why that that's not a controversial change.  Æµ§œš¹  [aɪm ˈfɻɛ̃ⁿdˡi]  19:40, 1 December 2007 (UTC)

/ɳ/ and /ɽ̃/
I've always wondered about ɳ. Could someone give a word where it exists (aside from nasal assimilation cases, such as ʈəɳɖ 'cold'). In all the words which have the letter ण (e.g. कारण), it seems to me that those are really cases of ɽ̃ (sort of the opposite of what is attributed to Shapiro on the actual page--in that (IMHO) the ɽ̃ is the underlying phoneme (for borrowings anyway), but the ɳ is the allomorph. But even then, not of ɽ̃, but rather just some unspecified N at the same point of articulation as the following consonant ɖ, ʈ, etc.)

I realize that ण is considered a nasal being in the 5th line of the वर्णमाला, but really, it's always a nasalized retroflex flap instead of the plain vanilla nasal (except where it is assimilation), no? Bob Eaton —Preceding unsigned comment added by 70.251.33.214 (talk) 13:34, 14 December 2007 (UTC)


 * I agree. Though I guess the way things are are the ways things are for etymological (e.g. "five corresponding nasal stops") and cross-linguistic comparison purposes. Tuncrypt (talk) 17:10, 16 January 2008 (UTC)

Hindu/Hindi
I'm not a expert in this area, but anyway, should Hindu be corrected to Hindi in the first paragraph? In the phrase "Hindu and Urdu." 71.174.112.73 (talk) 02:46, 27 December 2007 (UTC)


 * Good catch. Æµ§œš¹  [aɪm ˈfɻɛ̃ⁿdˡi]  07:00, 27 December 2007 (UTC)

Voiced postalveolar fricative?
According to Persian phonology, Persian has the voiced postalveolar fricative. This sound is not included in the list of phonemes introduced to Hindi-Urdu in Persian loan words, though. Any particular reason why not? Was it replaced by something else? —Angr If you've written a quality article... 12:33, 17 February 2008 (UTC)


 * list some words with it Tuncrypt (talk) 23:36, 17 February 2008 (UTC)
 * I don't know any; I don't know Persian. —Angr If you've written a quality article... 23:57, 17 February 2008 (UTC)
 * well it's mostly like to have been transposed to dʒ Tuncrypt (talk) 00:28, 18 February 2008 (UTC)

links to orthography
I find this rather confusing. Do we not need some mention of how these sounds tie in to specific letters?? It is difficult to work it out, especially when the sounds given at Hindustani orthography are different from those on this page. Do [e] and [ɛ] represent different letters or are they allophones or what? Widsith (talk) 09:31, 17 October 2008 (UTC)

substandard/nonstandard
Per the edit war with Tuncrypt, who points to Masica (1991) in arguing that we should say some speakers of Hindi are producing substandard speech as opposed to nonstandard, here is the actual quote in question: "Admittedly such parameters, while representing a real and growing sociolinguistic reality, are at the same time both arbitrary and somewhat vague. Speakers who fail to consistently maintain (confusing them with /ph, j, s/ are by this definition substandard, although these are 'foreign' sounds..."

What Masica is saying is that if you consider this feature to be intrinsic to standard speech, then you are the kind of person to see differences from the standard as distance from quality. However, he is not endorsing this view and in the overall context he's saying that the measure of conserving some sounds and not others is arbitrary. Given the context, it seems pretty clear that the word "substandard" is ironic and because we're taking information from him out of context that keeping "substandard" violates WP:NPOV. — Æµ§œš¹  [aɪm ˈfɻɛ̃ⁿdˡi] 22:23, 22 March 2009 (UTC)


 * Okay, true. I concede. Go ahead with your edit. Tuncrypt (talk) 22:35, 22 March 2009 (UTC)

grapheme
Most phonology articles of other languages also list the basic correspondence between grapheme/alphabet and the phoneme, despite the fact that there is usually a separate article about the script. Here we should also have one for Hindi-Urdu in devanagari and Perso-Arabic. I am planning to add vowel and consonant table in a few days. Please consider.Cygnus_hansa (talk) 21:55, 28 May 2009 (UTC)

Stress
What does "the stress falls automatically on the penultimate heaviest syllable" mean? If the word is medium-light-heavy, I assume that it will not have penultimate stress, but have no idea if it would be antepen. or ult. kwami (talk) 07:24, 17 July 2009 (UTC)


 * I found some info, a summary of Hussein, but it wasn't very well worded. The original would be better. Other stuff I've found doesn't appear to agree, but all are problematic. kwami (talk) 09:10, 17 July 2009 (UTC)

Sound files for unique contrasts
This article would benefit from sound files of minimal pairs. Some phonological contrasts, like dental and retroflex consonants, and voiceless, voiced, breathy-voiced, and aspirated stops, aren't found in English, and it would help English-speaking readers if they could hear examples of them. — Eru•tuon 19:55, 13 January 2015 (UTC)

I'm looking for good Hindi examples to add to this page. I'm going to list them here, but I'm not ready to have them recorded yet. Let me know if I've written the Devanagari wrong; also, if anyone can type the Urdu for me, that would be wonderful. — Eru•tuon 19:27, 8 March 2015 (UTC)


 * मील / mīl  "mile"
 * मिल / mil  "meet"
 * मेल / mel  "harmony"
 * मैल / mail  "dirt"
 * माल / māl  "goods"
 * मल / mal  "rub"


 * बैटरी baiṭri
 * बैंक baiṃka


 * कूल / kūl  "shore"
 * कुल / kul  "lineage"
 * बोल / bol  "speak"
 * कौल / kaul  "Kol" (?)


 * है / hai  "is"
 * हैं / haiṃ  "are"
 * कहा / kahā  "said"
 * कहाँ / kahāṃ  "where?"
 * वही / vahī  "that very one"
 * वहीं / vahīṃ  "right there"
 * हो / ho  "(you) are"
 * हों / hoṃ  "(they) might be"


 * बड़ा / baṛā  "big"
 * बढ़ा / baṛhā  "increased"


 * BTW, shouldn’t the vowels in mail, kaul, hai, and hain&#x331; be long vowels ([ɛː] and [ɔː])? —ШαмıQ ✍ @ 21:54, 8 March 2015 (UTC)
 * Yes, it appears so. I've changed them accordingly. — Eru·tuon 22:01, 8 March 2015 (UTC)
 * BTW, for back vowels, these examples may be more appropriate: luṭ / lūṭ / loṭ / lauṭ. —ШαмıQ ✍ @ 15:47, 9 March 2015 (UTC)
 * लुट / luṭ  "be looted"
 * लूट / lūṭ  "loot"
 * लोट / loṭ  "roll/wallow"
 * लौट / lauṭ  "return"
 * Thanks, that's much better since it has the same consonants in all of them. If you can think of an example of the vowel in a loanword from English that's similar to the front vowel examples, etc., that would be great. — Eru·tuon 19:03, 9 March 2015 (UTC)
 * बैंक, बैटरी — isoham (talk) 17:52, 10 March 2015 (UTC)
 * Added these. Not totally sure about the transliteration and IPA transcription. Can you verify them, ? — Eru·tuon 18:02, 10 March 2015 (UTC)


 * I know how to read Devanagari, but I am not an expert in Indic transliteration... I do not know how to transliterate "bank". I can transliterate Urdu using ALA-LC per WP:MOSUR. It should be like this:
 * बैटरी / baiṭrī  (battery) -- [r] is trilled (should it be changed?), i is long.
 * बैंक / baink  (bank) -- there is an [ŋ] sound.
 * These are the best examples I could think of. But, in one example here, bel does not have the same consonant in the end.
 * बीन / bīn  "lute"
 * बिन / bin  "without"
 * बेल / bel  "creeper"
 * बैन / bain  "between"
 * बैन / bain  "ban"
 * बान / bān  "habit"
 * बन / ban  "bun"
 * Since this isn't any better than the original, I would just suggest adding the word "map" or "mat" to the mīl examples. The long /æː/ sound in map/mat is different from the long /ɛː/ sound in mail (dirt). —ШαмıQ ✍ @ 18:56, 10 March 2015 (UTC)


 * Devanagari is all correct. I've never heard of कूल and feel कौल should be replaced by something that's not a proper noun — isoham (talk) 11:43, 9 March 2015 (UTC)
 * Somehow I didn't notice your comment. I can't verify the words myself, of course, but I found कूल in an online dictionary. It would be ideal to replace कौल with a common noun, but in order to have a series of words that have the same consonants, we may have to use it.


 * Oh, never mind. We don't have to use the proper name, because found another set of examples. — Eru·tuon 18:44, 12 March 2015 (UTC)

ph
In the box where all the stops are listed it gives IPA [fal] for phāl. That can't be correct, surely? Nothingbutmeat (talk) 13:09, 9 March 2015 (UTC)
 * Yes, [fal] is incorrect. —ШαмıQ ✍ @ 14:38, 9 March 2015 (UTC)
 * Yeah, the word is supposed to have, but the audio recording currently has . — Eru·tuon 19:04, 9 March 2015 (UTC)

External links modified
Hello fellow Wikipedians,

I have just modified one external link on Hindustani phonology. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:
 * Added archive https://web.archive.org/web/20071025082631/http://www.speech.sri.com/people/rao/papers/icslp96_wbhyp.pdf to http://www.speech.sri.com/people/rao/papers/icslp96_wbhyp.pdf
 * Added tag to http://www.rosettaproject.org/archive/hin/phon-3?page_view=image_view

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

Cheers.— InternetArchiveBot  (Report bug) 05:28, 4 November 2017 (UTC)

[w] or [β]?
I suspect that the bilabial pronunciation of the phoneme /v/ is actually the bilabial approximant (approximant, not fricative!) -- rather than the labialized velar approximant. The two sounds are barely distinguishable in practice, but the thing is: Why should the sound undergo velarization in one position and not in another? It's unlikely. German is another language where [v]~[ʋ]~[β] are allophonous and the pronunciation there is definitely [β] (used post-consonantally by some speakers). Something very similar can be said about Dutch, too. — Preceding unsigned comment added by 84.188.186.58 (talk) 14:53, 24 February 2018 (UTC)

ʂ
Isn’t that sound only present in Sanskrit and pronounced like a ‘Sh’ in Hindi? Taimoorahmed11 (talk) 18:29, 3 May 2020 (UTC)


 * Correct عُثمان (talk) 17:23, 16 March 2023 (UTC)

e Vowel in Phonology
The Hindustani languages also has some words and or nouns with e (not eː) vowel like مراکش/मराकॆश/Marākeś and پشاور/पॆशावर/Peshawar but apparently it's not mentioned in Hindustani Phonology. Am I wrong or right in suggesting that the e vowel exists in Hindustani? Taimoorahmed11 (talk) 19:57, 14 August 2020 (UTC)


 * @Taimoorahmed11 You may be correct, however, I think its better to mention it here first (with a source if available) and explain here whether the short [e] occurs, word initially, medially or finally in contrast with the long [e:] (just like how this page and a footnote in the IPA Help page mention that the short [i] and [u] occur word-finally in contrast with the long [i:] and [u:]) before mentioning it in the Help:IPA/Hindi and Urdu because mentioning it there without any supporting mention in Hindustani phonology does create inconsistency with this page and confusion for readers which I assume is why Aeusoes1 has reverted your changes in Help:IPA/Hindi and Urdu. Broman178 (talk) 18:03, 12 September 2020 (UTC)
 * @Broman178 Point noted. I'll try and find some references but I doubt I will find any suitable references since I believe most of the studies are based on the Hindi/Urdu alphabet however there is no native diacritic (in either scripts) for non-prolonged e vowel neither are there many words with that sound. Nevertheless, I will try and find some references. — Preceding unsigned comment added by Taimoorahmed11 (talk • contribs) 23:46, 12 September 2020 (UTC)

"Shortened" tense vowels
Notification:

I have opened a discussion about the notion of "shortened" tense vowels in Help talk:IPA/Hindi and Urdu which is also relevant for this article. –Austronesier (talk) 09:12, 10 November 2020 (UTC)

Diaphonemic analysis of consonants
Hiya! I have a suggestion regarding the discussion of Hindustani consonants on this page.

First, the concern. Hindustani is a pluricentric language and all of its speakers don't share a single phonemic inventory (at least as far as the consonants are concerned; there's significantly more uniformity when it comes to the vowels). I think this pluricentrism is what leads to the contentions mentioned above regarding the phonemic status of certain consonants (/ʂ/, /ɳ/, /ʒ/, ...). Presumably /ʂ/ is roughly as marginal a phoneme as /ʒ/, but the latter is listed as a marginal phoneme in the chart while the former isn't. On the flip side, the retroflex nasal /ɳ/ is accorded full phonemic status, even though it's probably roughly as marginal as something like /q/. The velar nasal /ŋ/ is also accorded full phonemic status in the chart (even though it only occurs in homorganic clusers and could legitimately be called an allophone of /n/), while the palatal nasal /ɲ/ isn't even given marginal status (even though it too only occurs in homorganic clusters).

Anyway, my point is this: there are some tricky questions to answer regarding Hindustani's phonemic inventory, and the reason they're tricky is because there's genuine variation in the phonemic inventories of varieties of Hindustani because we're dealing with a pluricentric language. I presume we'd all agree that it would be good to answer each of these questions accurately and without privileging one variety of Hindustani over another.

My proposal for accomplishing this is that we give a diaphonemic analysis for Hindustani consonants. For example, we might say that #q and #k are two distinct Hindustani diaphonemes. Speakers of some varieties maintain these two diaphonemes as two distinct phonemes (/q/ and /k/, respectively), while speakers of other varieties collapse both of these diaphonemes into a single phoneme (namely, /k/). Similarly, we might say that the retroflex nasal #ɳ and the dental nasal #n are two distinct diaphonemes; speakers of some varieties maintain these as two distinct phonemes, while speakers other varieties collapse them both into a single phoneme /n/.

I think this would allow us to describe Hindustani's consonant system, and the variation in the consonant system, without privileging any one variety of Hindustani. I'd be happy to write this up, but I'll wait a few days to see if anyone has any suggestions/objections about this proposed change. EmptyStardust (talk) 05:00, 29 January 2021 (UTC)


 * @EmptyStardust It's a not so open secret that all of the phonemes purported to be specific to Urdu or Hindi are not used at all by native speakers. Those who pronounce sounds such as /q/ are doing it as a learned affectation and reliably revert to /k/ when not conscious of their speech. This is evidenced particularly by the presence of a very large number of Perso-Arabic loan words which were introduced solely through speech and not writing in which the spellings reflect their true pronunciation. The verb یرکنا "yarakna" ultimately comes from Arabic یرق "yaraq" yet we never see /q/ in this word because to the ear of native Hindustani speakers the actual sound of ق in spoken Arabic sounds indistinct from /k/.
 * Technically it is true that a certain number of speakers use some of these sounds, but they tend to insert them in places where they aren't spelled as such due to these being sounds native to other languages of the region which carry over in their accent speaking Hindustani as a second language. Native speakers of Hindi/Urdu pronounce وقت as "vakat" but native speakers of Pashto or Hindko are more likely to pronounce the same word as وخت "vaxat" as the sound represented by خ exists in those languages and has a tendency to replace /k/. Likewise, most native speakers of Punjabi will pronounce words like ہونا "honā" as "hoṇā" with the retroflex nasal even when speaking Hindi/Urdu because it is a native sound in Punjabi which is suffixed to most verbs. The sound represented by /ʂ/ is likewise probably present in the Hindi/Urdu speech of many Bengali speakers, but as a spontaneous substitute for other silibants. These tendencies are only effects of the phonologies of these other languages and say nothing of the native phonology of Hindustani.
 * The only conceivable way sounds from Sanskrit or Arabic could have entered Hindustani is if there were a large number of speakers who were bilingual with these languages, and this has never been the case. Speech always comes before literacy and the presence of certain letters in languages which have always been literary languages in the context of the recent history of the region has no bearing at all on pronunciation. You would not be able to find a child fluent in Hindi/Urdu who knows to pronounce وقت as "vaqat," and adults who pronounce it this way would have to adjust their pronunciation to be understood by their less pretentious peers. عُثمان (talk) 17:19, 16 March 2023 (UTC)

Diphthongs in Hindi & Urdu?
Hi everyone, does anyone know if there are any diphthongs which occur in both Hindi & Urdu because I feel this article currently seems to be lacking on information as to whether or not diphthongs occur in Hindi & Urdu and if they are present, what are those diphthongs? I would be grateful if anyone who knows the Hindustani phonemes quite well, or who speaks Hindi or Urdu or even both languages can comment on this because if diphthongs are present in Hindi & Urdu, it needs to be mentioned in this article (with sources if available) or at least needs more clarification, many thanks. Broman178 (talk) 09:21, 2 February 2021 (UTC)
 * The answer is simple: no and yes. No, if you only consider Urdu amd the standardized Hindi of Dehli. Yes, if you consider dialect interference and the way Standard Hindi is pronounced in the wider Hindi Belt. Corresponding to the standard open mid-vowels and, you will often hear diphthongs  and  (Shapiro 2003, Kelkar 1968), which is the regular realization of these sounds especially in Eastern Hindi "dialects" such as Awadhi. So speakers from variants with diphthongs will take them over even when speaking Standard Hindi. So we could mention the diphthongs as often heard non-standard variants of  and  in Hindi (but not in Urdu).–Austronesier (talk) 11:46, 2 February 2021 (UTC)
 * Thanks for your reply, it makes a lot more sense to me now. Broman178 (talk) 09:54, 3 February 2021 (UTC)

Allophony of /ʋ/
The article's IPA chart specifies /ʋ/ as the labiodental consonant. Its postconsonantal allophony with [w] is explained well enough in the subsection "Allophony of [v] and [w]". I'm questioning how this subsection gives only [v] as the labiodental. You see, I came to this article following a link from the [v] article giving the word vrat̪ as an example (the same example given here). I was looking if fricative [v] could be found in any Indian languages. The example of vrat̪ appeared to be another positional allophone, this one triggered by word- or syllable-initial preconsonantal position. I was already aware that /ʋ/ is the Hindustani labiodental, so I would not expect to find [v] at all, unless provided for by allophonic rules. If /ʋ/ has two positional allophones, the article needs to make that clearer. Johanna-Hypatia (talk) 09:13, 5 January 2022 (UTC)