Help talk:IPA/Sanskrit

ɐ अ, प a comma

Comma is not a good example. This word is pronounced diffrently in different english dialicts (American, Australian, etc.) — Preceding unsigned comment added by 66.44.30.95 (talk) 06:52, 18 November 2024 (UTC)

Dental diacritic

I don't see any reason to have a dental diacritic for Sanskrit. Spanish dental plosives, for example, are transcribed at Wikipedia without them because it would be redundant. Is there some contrast between dental and alveolar that I'm not aware of? — Ƶ§œš¹ _{[ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ]} 12:44, 15 March 2012 (UTC)

I'm not sure why I included them originally, as you are right. Note that WP:IPA for Hindi and Urdu has them, too. Actually, that's probably why—I think I copied the Hindi page as a framework for this one. — ˈzɪzɨvə (talk) 02:51, 16 March 2012 (UTC)

I have brought up the issue there, as well. If we do decide to remove the diacritic, it shouldn't be too difficult to remove them from the relevant articles. — Ƶ§œš¹ _{[ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ]} 03:12, 16 March 2012 (UTC)

Hello Ƶ§œš¹ and — ˈzɪzɨvə, keeping the dental diacritic would be a good idea for the sake of newcomers to IPA and Sanskrit, and also because:

1. IPA chart (vowels and consonants) - 2015. (pdf file) has /t/ in the alveolar; /t̪/ with dental diacritic will make it clearly in the dental.
2. Sanskrit_grammar#Consonants shows it in dental. When learning to make these specific sounds, /t̪/ clearly shows dental and in contrast with retroflex /ʈ/. (which seems not to be the case with Spanish_phonology#Consonants, just looking at the chart, Spanish has only 1 't' which is dental 't' and does not have retroflex 't')
3. The discussion about 'retroflex versus dental' in Help talk:IPA for Hindi and Urdu's talk page here & also here, talks about the difficulties for newcomers in actual pronunciation difference between dental /t̪/ and retroflex /ʈ/ . May be keeping the dental diacritic will make it a bit more easier.
4. Devanagari is from the Brahmic family of scripts. Others from the same script family are Tamil_language#Consonants, Malayalam#Consonants, which have 3 't's (dental /t̪/, alveolar /t/, retroflex /ʈ/, common usage in Tamil 2 't's: dental /t̪/,retroflex /ʈ/, voiced as /d̪/, /ɖ/ respectively) and 3 'n's (dental, alveolar, retroflex as in /n̪/,/n/,/ɳ/). For people trying to 'place' these corresponding consonant's sounds, dental diacritic will be helpful. Thanks, by contributor 2know4power (talk) 00:13, 9 February 2017 (UTC).

Thanks for your response. I'm still not convinced, though. The diacritic is redundant. A contrast between t and ʈ is already apparent, by virtue of the different letters being used. Also, our goal isn't really language instruction, which I think allows for a little bit of phonetic imprecision; this is the framework by which we omit the dental diacritic in many of the languages we transcribe that have dental stops represented with alveolar stop characters in IPA transcriptions at Wikipedia. And I don't think the origins of the writing system to be a compelling reason to maintain a redundant diacritic. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 17:49, 9 February 2017 (UTC)

Hello Ƶ§œš¹, Thanks for your comments here. And thank you very much for the great work you do for Wikipedia especially in language studies.

From an IPA newcomer's point of view, it is points 1, 2, 3, but mainly 1. For the purpose of learning, the IPA chart shows /t/ in alveolar, /t̪/ is an accurate representation of dental. Again this is an IPA newcomer's POV. Could you please say what 'redundant' means in linguistics here, (for dental diacritic in marking dental place of articulation)? Dictionary gives meaning as 'repetition, not needed'. For point 4, Dravidian languages Tamil and Malayalam, they have dental, alveolar and retroflex consonants. People speaking these languages use Tamil's Grantha script or Tamil script system with extra super/subscript notation and Malayalam script to write the Sanskrit language, if they are unfamiliar with Devanagari script. With the dental diacritic retained, the IPA notation will be comparable or corresponding between the writing systems. Thanks, by contributor 2know4power (talk) 03:00, 10 February 2017 (UTC).

Long E

eː ए, पे e between yell and Yale

Isn't the vowel in "steer" or "beer" much more accurate than the current explanation? — Preceding unsigned comment added by 94.255.182.142 (talk) 13:04, 31 March 2013 (UTC)

No, steer and beer have a vowel very close to see. I've changed the word use to approximate the sound. — Ƶ§œš¹ _{[ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ]} 14:26, 31 March 2013 (UTC)

Low vowel

Is the long low vowel really back [ɑː]? I would expect it to be central or even front. — Eru·tuon 02:22, 4 March 2015 (UTC)

@Erutuon: Any of [a, ɐ, ɑ] are equally valid ways of transcribing the open central unrounded vowel, as there is no dedicated symbol for that sound. The first symbol is the most common way of transcribing this sound, the second symbol is often used when transcribing Australian English, and the third symbol is used in Danish, sometimes also in German. That said, I think it's central, as in Hindi. Or it depends on the speaker. — Peter238 (v̥ɪˑzɪʔ mɑˑɪ̯ tˢʰoˑk̚ pʰɛˑɪ̯d̥ʒ̊) 23:33, 9 March 2015 (UTC)

@Peter238: Given that, I'll switch the symbol to ⟨a⟩, which is after all the "unmarked" Latin character. I think ⟨ɑ⟩ should only be used in vowel systems where the vowel is clearly back or differentiated from a front open vowel. I don't think that's the case in Sanskrit. — Eru·tuon 00:33, 10 March 2015 (UTC)

@Erutuon: Right, but first we need to be sure that it indeed is central (I wouldn't expect it to be front TBH.) Zieba & Stiehl (2002) cited on this page transcribe this sound as /aː/. — Peter238 (v̥ɪˑzɪʔ mɑˑɪ̯ tˢʰoˑk̚ pʰɛˑɪ̯d̥ʒ̊) 00:40, 10 March 2015 (UTC)

Remember that now that you have changed [ɑː] to [aː], you have to go through articles which have Sanskrit IPA transcriptions, and change every instance of [ɑː] to [aː]. — Peter238 (v̥ɪˑzɪʔ mɑˑɪ̯ tˢʰoˑk̚ pʰɛˑɪ̯d̥ʒ̊) 00:44, 10 March 2015 (UTC)

Hmm, sounds like a lot of work. May be able to use AutoWikiBrowser, though. But first should confirm: do other sources use /ɑː/, or what was the origin of that transcriptional choice? — Eru·tuon 00:56, 10 March 2015 (UTC)

I actually just did an AWB search-and-replace, and didn't make any replacements. Either all Sanskrit transcriptions already use /aː/, or I messed up the regular expressions... — Eru·tuon 01:06, 10 March 2015 (UTC)

I don't know, but they probably do: @Xyzzyva: @Kwamikagami: @Aeusoes1:. — Peter238 (v̥ɪˑzɪʔ mɑˑɪ̯ tˢʰoˑk̚ pʰɛˑɪ̯d̥ʒ̊) 01:00, 10 March 2015 (UTC)

AFAIK, since Sanskrit is not normally a native language, it generally inherits the phonetics of the L1 of its speakers. Even when it's been revived (if that's the right word for a language that might always have been an artificial construct), it will inherit the pronunciation of its speakers's ancestral language. So I doubt we can say what "the" pronunciation of ā is, other than generally being low. Since <a> is the default letter for a low vowel, I would support transcribing it [a:]. — kwami (talk) 01:09, 10 March 2015 (UTC)

Agreed. What is the trascription of [ɑː] based on? — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 03:45, 10 March 2015 (UTC)

Xyzzyva created the page way back in 2010, apparently based on Help:IPA for Hindi and Urdu. Strangely, the version of the Hindi-Urdu page that existed at the same time didn't have /ɑː/ in it, so I'm not sure where it came from. — Eru·tuon 07:37, 10 March 2015 (UTC)

AFAICS there's no reason to transcribe it [ɑː] then. I support the (already done) change to [aː]. — Peter238 (v̥ɪˑzɪʔ mɑˑɪ̯ tˢʰoˑk̚ pʰɛˑɪ̯d̥ʒ̊) 14:14, 10 March 2015 (UTC)

Yeah, it looks like I based it on the then-current version of the Sanskrit page, which had [ɑː]. I certainly claim no specialist knowledge of Sanskrit beyond what I know from here. Probably time to take out those dental diacritics, too. — ˈzɪzɨvə (talk) 00:30, 11 March 2015 (UTC)

Sanskrit [c][ɟ] or [tʃ][d͡ʑ] [d͡ʒ] ?

Every single page about Sanskrit and Hindi in wiki and also different scientific resources use different sound transcriptions.
E.g. https://en.wiktionary.org/wiki/%E0%A4%9A
http://en.wiki.x.io/wiki/Ca_%28Indic%29#Devanagari_script
http://en.wiki.x.io/wiki/Devanagari#Consonants
What are the actual sound values of these letters? Is there more reliable resource on Sanskrit IPA than http://www.sanskritweb.net/deutsch/ipa_sans.pdf — Preceding unsigned comment added by 95.54.226.38 (talk) 02:55, 17 August 2015 (UTC)

As kwami wrote above, the way people pronounce Sanskrit depends on their native language. I bet these are most commonly alveolo-palatal, either stops [c̳, ɟ̳], affricates [t͡ɕ, d͡ʑ], or something in between (i.e. affricated stops, like Slovak /c, ɟ/). That's what I'm often hearing in Indian English. Unfortunately, I'm not aware of any sources that describe this (probably because I've never looked that much into it). Peter238 (talk) 22:27, 17 August 2015 (UTC)

Audio file for pronunciation with ISO 15919 for Hindi

Audio file-Consonants with ISO 15919 for Hindi

Found this helpful audio file at Hindustani phonology (diff's version at retrieval) for pronunciation, with IAST & IPA. It does not contain audio for all the consonants, but only for plosives and affricatives.

These audio files seem relevant for discussion about 'retroflex Vs dental' in Help talk:IPA for Hindi and Urdu's talk page here & also here.

I think, a person with Hindi as L1 might use the same pronunciation for these consonants in Sanskrit too, so this audio file may be helpful for the other non-Hindi speakers to learn Hindi & Sanskrit pronunciation. Thanks, by User 2know4power (talk) 02:14, 12 January 2017 (UTC).

Pronunciation for Hindi versus Sanskrit

May be people who know Sanskrit could, please take a look at this: audio files for 'stop consonants' in Hindi pronunciation, & say if they can be used for the pronunciation of 'stop consonants' in Sanskrit too?. Thanks in advance, by User 2know4power (talk) 19:07, 16 January 2017 (UTC).

The IPA chart at Sanskrit_grammar#Consonants is almost the same as at Hindustani_phonology#Consonants. Thanks, by User 2know4power (talk) 12:15, 22 January 2017 (UTC).

It seems like all the pronunciations are the same here so phonetically I think it's fine. The only problem would be the translations of the sample words are still all hindi words which might be mildly confusing (thinking the examples are all sanskrit words instead). Iṣṭa Devatā (talk) 20:37, 6 February 2017 (UTC)

Thanks Iṣṭa Devatā, for your helpful suggestions. By contributor 2know4power (talk) 22:54, 8 February 2017 (UTC).

Move discussion in progress

There is a move discussion in progress on Help talk:IPA which affects this page. Please participate on that page and not in this talk page section. Thank you. —RMCD bot 16:18, 15 July 2017 (UTC)

Consistency with Source?

There is one source cited here, which does seem to be reliable. Why then are does the page not consistently use the IPA provided in the source? These should either be made consistent, or other sources provided. — Preceding unsigned comment added by Shaav (talk • contribs) 07:16, 18 July 2017 (UTC)

Related to your comment on my talk page, the use of the retroflex tap symbol ⟨ɽ⟩ in the source seems like a mistake: usually syllabic consonants are continuants. But okay, it's in the source, and I don't care to argue about it. — Eru·tuon 08:49, 18 July 2017 (UTC)

Actually, what's especially odd is ⟨ɽ̩ː⟩: a lengthened syllabic tap? That seems a contradiction in terms. Does it mean a trill? — Eru·tuon 08:51, 18 July 2017 (UTC)

Yes, [ɽ̍ː] (the diacritic should be placed above by the way) doesn't seem like a correct transcription. Let's just pick one symbol and stick to it, writing ⟨ɹ, ɽ⟩ etc. is an overkill. Mr KEBAB (talk) 10:40, 18 July 2017 (UTC)

Not impossible because it wasn't actually syllabic by Classical times, but a (retroflex?) flap followed by /i/. But people transcribe it as syllabic 'r' etc. regardless. — kwami (talk) 01:30, 9 March 2019 (UTC)

I agree that it should be consistent. If the source is inaccurate, we should find another more accurate source. Currently, the transcription shows र as [r], which is a trill. This seems counter to any source or anecdotal evidence (which would suggest it should be [ɾ] or an alveolar flap). Getsnoopy (talk) 21:38, 4 November 2019 (UTC)

I don't have a problem with deviating from a cited source if it's justified (and I don't know if that's the case here). If we do change our transcription conventions identified here, we would also want to change our actual transcriptions throughout Wikipedia accordingly. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 15:00, 5 November 2019 (UTC)

That we would. My original intention was to change the transcription here, but since the IPA key was as it was, I wanted to change it here before I started editing articles. It seems like there are quite a few discrepancies between the cited source and the key, as seen in my edit that was reverted. Getsnoopy (talk) 17:37, 19 November 2019 (UTC)

Transcription of palatals

I went through all the Sanskrit transcriptions to remove the dental diacritics and to conform transcriptions to what's at this chart. While I changed all instances of [ʃ] to [ɕ], and instances of [tɕ] and [dʐ] to [ʃ] and [dʒ] respectively, I wonder if we shouldn't be more consistent. What do people think? — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 17:49, 1 January 2019 (UTC)

@Aeusoes1:. I came here to ask just that. Whether we use <ɕ> or <ʃ>, we should be consistent, unless sources say the articulation was different. — kwami (talk) 01:38, 9 March 2019 (UTC)

Redid table

Someone asked me to check the table after an editor gave it an eastern (Oriya/Bangla) pronunciation. I restored it to Classical pronunciation as best I could. But that meant changing some longstanding transcriptions, such as the 'syllabic' r and l being transcribed as actual syllabic consonants. According to our Vedic Sanskrit article, these had broken by Classical times. (Unfortunately, it's a bit difficult to tell when that article is describing Vedic and when Classical pronunciation.) Also made <r> retroflex, the diphthongs <ai, au> > [ai, au] (unless I'm misreading that and they should be [əi, əu]), removed the retroflex el, etc.

We could have two columns, Vedic and Classical, if people like. But otherwise, given that we use stress rather than tone, IMO we should stick to Classical. — kwami (talk) 01:38, 9 March 2019 (UTC)

@Erutuon:, @Mr KEBAB:, @Peter238:, @Aeusoes1: If y'all think the table is okay, I'll use AWB change the transclusions to match. We should decide about the preceding question too, & also whether I messed up and the <a> in [ai, au] in our Vedic article was supposed to be the same as plain <a> = [ə]. — kwami (talk) 02:59, 9 March 2019 (UTC)

As for our source, Zieba & Stiehl, they don't say what they mean by 'original'. Sanskrit has a long history. The transcription of ऐ as [aːi] would be Vedic, but then ए should be [ai], right? They also have [x] for visarga. Was that also Vedic? I understand that there were Vedic allophones of visarga such as [ɸ].

I do think centralized [ɐ] for the inherent vowel seems more likely than mid [ə], though. — kwami (talk) 18:14, 9 March 2019 (UTC)

In this edit I replaced the existing English approximants by the examples in Goldman's Devavāṇīpraveśikā. I have self-reverted for now because while these examples are well-sourced, not all of them are necessarily improvements over what was there previously and they necessarily reflect (a dialect) of North American English (for example, the 'o' in mom as an approximant for ā). Also note that the source/examples are certainly for classical/contemporary Sanskrit.

Welcome feedback on if/which of the Goldman's examples are worth retaining. Abecedare (talk) 18:25, 9 March 2019 (UTC)

I don't see how 'tint' is better for retroflex tee. But s.t. like skin, kin would be nice to have the same base letter as the IPA. 'Cinch' is probably better than 'canyon' though. Although 'canyon' is closer to the IPA letter, 'cinch' is presumably closer to the sanskrit. But I'd prefer to delete the separate entry for that altogether. (Below.)

I took the stress off काल, as it seemed dubious it would be on the 2nd syllable. — kwami (talk)

• Proposal: Given Goldman's equivalence of ञ ñ as in "cinch", what's the point in having a separate IPA letter? I find it misleading. IMO we no more need IPA ɲ for ञ than we'd need IPA ɱ for a nasal before व. The assimilation is automatic for English speakers, just as the assimilation of /m/ before /f,v/ is, so I propose removing IPA ɲ from the chart and transcribing ञ with IPA n (though keeping IAST ñ as that's standard). If we were Spanish WP, I'd also propose removing IPA ŋ, but since that's phonemic in English I think it's best we keep it. — kwami (talk) 19:57, 9 March 2019 (UTC)
  • ñ doesn't only occur before an obstruent; see for instance ञकार. It could be replaced with ⟨n⟩, but only in some cases. — Eru·tuon 20:02, 9 March 2019 (UTC)

    I didn't think it was ever phonemic. Is a word like that ever likely to be in WP? Do we even need it in the key? If we used the dental diacritic we'd mark the difference that way, of course, but that's easy enough to automate if we ever decide to go back to it. — kwami (talk) 20:08, 9 March 2019 (UTC)

    I don't know enough about Sanskrit to answer either question. I did find terms with ñ following j, derivatives of the root ज्ञा (jñā), when searching with the query : incategory:"Sanskrit lemmas" intitle:/ञ/ on Wiktionary. That could be predictable since it's after a palatal obstruent, but I don't know. — Eru·tuon 20:17, 9 March 2019 (UTC)

    I was just going to ask about ज्ञा. That has its own pronunciation in the modern languages. It's not just the [ɲ], but also the [dʑ] that should be confirmed. (e.g., perhaps the affrication of *ɟ didn't happen in this context.) I've found it in Paadal Petra Sthalam and Aathara Stalam and it may be in anja. — kwami (talk) 20:22, 9 March 2019 (UTC)

(edit conflict)I don't think phonemicity is a good rationale to remove phonetic accuracy here. That's not what we do for other languages. We've got the IPA characters at our disposal and it's not too much of a burden on readers or editors to use them. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 20:27, 9 March 2019 (UTC)

I agree in theory, but in this case it isn't phonetically accurate. <ñ> isn't palatal [ɲ], but post-alveolar [n̠]. It isn't even [n̠ʲ] like in many languages we transcribe with <ɲ>. If we transcribed dental <n> with [n̪], then we'd certainly want to maintain the distinction, but since we've already abandoned the dental diacritic, our attempt to align the IPA with IAST is making us less accurate. (I'd support restoring the dental diacritic, BTW.) — kwami (talk) 20:42, 9 March 2019 (UTC)

That's quite a bit of phonetic precision that you're gleaning from Goldman saying Sanskrit ⟨ñ⟩ is like the ⟨n⟩ in English cinch. Or do you have something more in-depth that talks about this? — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 21:41, 9 March 2019 (UTC)

Yes. All sources agree that the sounds are homorganic. [ɲtɕ] (or [ɲtʃ]) are not. A homorganic sequence would be written [n̠t̠ɕ] (or [n̠t̠ʃ]) if precision were necessary, though in the Sanskrit article we use [n̠] for the retroflex.

On the other hand, modern reflexes of jñ such as gny [ɡnj] and gy suggest that in the reverse order the nasal truly was palatal and the stop didn't become an affricate. That's supposition, though, since Sanskrit as a construct is not the same as the ancestor of the modern languages. — kwami (talk) 05:49, 10 March 2019 (UTC)

This all sounds like supposition. It could very well be that ⟨ñ⟩ is [n̠] before postalveolar affricates and [ɲ] elsewhere. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 06:06, 10 March 2019 (UTC)

I think alveolo-palatal nasals are sometimes written with ⟨ɲ⟩, as in the article for alveolo-palatal nasal. — Eru·tuon 07:14, 10 March 2019 (UTC)

True, but AFAIK only written that way when standing on its own. ⟨ɲtɕ⟩ would be quite odd -- makes it look as though they're not homorganic. I've seen ⟨n̠t̠ɕ⟩ and ⟨ȵȶɕ⟩, though, when people want to be precise about the place. I suppose we could have ⟨nʲtʲɕ⟩ as well. Considering that we don't even mark the dentals as such, that would seem to be overkill for us. — kwami (talk) 00:36, 17 March 2019 (UTC)

I also wonder about tatsama. Is that t+s, or a modern pronunciation of c? — kwami (talk)

• Could we please not replace all transcriptions across the whole of wikipedia unless we know what we're doing? I'm sorry I can't contribute in a more positive way as it's been a while since I've last delved into Sanskrit, but some things strike me as distinctly odd: like treating the rhotic as a flap, or transcribing ṛ as ɽɪ, which only seems to reflect the vernacular pronunciation now current in parts of northern India; also, you can't dispense with ɲ altogether as it doesn't only appear before palatals (it's also an alophone of j before n), and the ts of tatsama strides a morpheme boundary, etc, etc. Sanskrit phonology is a pretty well researched area, and there's plenty of literature out there for anyone willing to look for it. – Uanfala (talk) 20:54, 9 March 2019 (UTC)

@Uanfala: Well, currently it's consistent -- or at least more than it had been, I'm sure I didn't catch everything -- and once it is consistent it's easy enough to change our minds w AWB. I've seen several sources that the rhotic was a flap and that ṛ had broken by classical times. Those claims may be wrong, of course. I think we need a consensus source to base this on. We don't seem to have ever had one, and even in the Sanskrit article people would revert each other without much evidence presented. In that article I moved the IPA out of the consonant tables and into a 'pronunciation' section, where hopefully we can have a table comparing the claims of different sources. As for ɲ, I suppose it's as likely to be in a link to this key as syllabic el. In our links to this key, we've just got ज्ञ in the two articles I mentioned, and a couple cases of ñc, ñj. I don't think I've ever seen a source that ñ was palatal unless analyzing c, j as palatal as well. — kwami (talk) 21:23, 9 March 2019 (UTC)

Correction: Zieba et al. have ñ palatal but the c, j it's assimilated to alveolar. But then they also claim ŗ is a syllabic flap, so I'm not giving them much credence.

Given the pronunciation of ज्ञा jñ in modern languages, it would seem this didn't affricate in Sanskrit. So I would expect it to still be [ɟɲ] in the Classical language, which would require another row in this key. (One source I found even had [ʒ̃]!) — kwami (talk) 22:37, 9 March 2019 (UTC)

Pronunciation of Syllabic r in sanskrit.

Syllabic r of Sanskrit is here given 'ri' as its IPA value, but it is known very well that it is true only for Hindi and other modern Indo-aryan languages. For Sanskrit it is syllabic, as its name suggests. Not in a single wikipedia page, it is 'ri'. It seems the page to be a copy from Hindi IPA page with several edits. I am replacing 'ri' with syllabic r and I don't think that such a well known fact should be more ignored anymore, as my last edit was reverted and I don't expect this to be repeated. — Preceding unsigned comment added by Ashutosh Jha (talk • contribs) 15:29, 13 May 2020 (UTC)

You make a compelling case, but please wait and give the community time to respond to your proposed changes. We don't want to make changes without consensus for a help page like this that is linked to by hundreds of articles (including, for example, Krishna, which does indeed transcribe this syllabic r in the way that the Help page lays out).

What do people think of this? I'm particularly interested to hear from @Kwamikagami:, the editor who initially made the edit to the current form. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 16:03, 13 May 2020 (UTC)

(edit conflict)Yeah, that's really odd. Aeusoes1, do you have any sources at all that describe Classical (or pre-Classical) Sanskrit ṛ as /ɽi/? As far as I know, this pronunciation is an artefact of a later pronunciation of Sanskrit, which was used long after the language had died as a medium of spoken discourse. From what I remember of Allen's Phonetics in Ancient India, the extant descriptions of the sound in the Pratishakhyas vary quite a lot, which could indicate either dialectal divergence or conflicting interpretations, but I don't think it was ever pronounced as an r + i. – Uanfala (talk) 16:11, 13 May 2020 (UTC)

I was going off our articles, so you'd need to check the sources there. One is Tiwari ([1955] 2005), which describes the changes from Vedic to Classical. Remember, this IPA key is for Classical, not Vedic. As of two minutes ago, our Vedic article stated the following. How much of this is Tiwari and how much might be OR I dont't know.

The pronunciations of syllabic [r̩] (ऋ) and [l̩] (ऌ) and their long counterparts [r̩ː] (ॠ) and [l̩ː] (ॡ) no longer retained their pure pronunciations, but had started to be pronounced as short and long [ri] (रि) and [li] (ल्रि) during the decline of Sanskrit. It partly varies by region like in southern India, including Maharashtra, people will pronounce ऋ as ru (रु)and ऌ as lru (ल्रु), this is pronounced according to Taittiriya Pratishakya of the Taittiriya Shakha or branch of the Yajur Veda. Virtually no region pronounces as would in vedic Sanskrit generally. Only while reciting Vedas is it normally heard. But if the Pratishakhya, i.e. the pronunciation manual, mandates it otherwise, the latter is followed.

I checked out the sources in our Sanskrit article that I could access, and Goldman & Goldman (2002) says that the <ṛ> is pronounced like the ri in "rig", which supports our Vedic article and what I had changed this key to last year.

In any case, it would be foolish for our IPA guide to contradict our articles on Sanskrit. Start a discussion on changing the phonology/phonetics sections of those articles per best sources, and if people agree on a change, then change this key to match. But don't do it backwards.

IMO, as I said in the previous thread, we really should have a consensus source to go off of, rather than reverting each other based on whatever source we happened to read yesterday. For a living language we can look for expert phonetic descriptions, but that's not going to work for liturgical and artificial ones. — kwami (talk) 05:52, 14 May 2020 (UTC)

BTW, the Whitney grammar Ashutosh used as a ref to support the retention of Vedic pronunciation in Classical (which AFAICT it never addresses) was published in 1879 and is scarcely intelligible. For one thing, it says the 'syllabic' ar is "untrilled". Does that mean it's [ɹ] as in General American English? Regardless, it contradicts Ashutosh's change to a syllabic trill.

I removed the 'syllabic' els. The short one only occurs in the single verb klp 'to fit', and so is extremely unlikely to ever be linked to this key (we can always add it if it is), while the long one doesn't occur at all. — kwami (talk) 08:32, 14 May 2020 (UTC)

This text in Vedic Sanskrit appears to be a recent addition (and so does the one in Sanskrit#Pronunciation, which agrees with it) - they certainly weren't around when I gave up watching those pages a few years ago. As far as I can see, they're confusing the pronunciation recommended in Sanskrit instruction manuals for students (which is based on the learned pronunciation of northern India of recent centuries) with the pronunciation of actual Classical Sanskrit. Unfortunately, I don't have access to the sources now (they're all offline, so they'll remain off bounds for quite some time), but we can't keep the status quo. It's as ridiculous as, say, giving Caesar's Latin pronunciation as [tʃezar]. – Uanfala (talk) 12:56, 14 May 2020 (UTC)

I'll remove the IPA values for ṛ: better have no information than information that is wrong (and besides, the sound doesn't appear to be much used in transcriptions, and the one major article that uses it - Sanskrit - has mercifully ignored the values in this page). – Uanfala (talk) 13:02, 14 May 2020 (UTC)

Please don't do that, Uanfala. There are, indeed, only a handful of articles that transcribe words with the relevant syllabic rs, but removing content from this help page is not the right choice because it doesn't help readers understand our transcriptions.

It sounds like the choice to do ɽɪ vs r̩ is a stylistic one. I like the latter personally, but I'm by no means an expert on the topic. Kwami, the editor who introduced ɽɪ seems to fine with whatever as long as we are consistent with Wikipedia articles that talk about Sanskrit. That seems reasonable to me. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 15:03, 14 May 2020 (UTC)

Well, so many reverts have piled up again – this endless reverting and restoring of dubious material is what made editing here so frustrating before. Sorry, I give up again, sort out your mess yourselves. – Uanfala (talk) 15:16, 14 May 2020 (UTC)

That's a bit dramatic, Uanfala. As I said from the very beginning, it is important that the help page match what is in our transcriptions. We've been reverting changes that undermine this all while trying to generate an explicit consensus on the best choice. I suggest next time that you try sticking to talk pages during disputes to avoid unnecessary and disruptive editing wars. The status quo version is over a year old. A few days more of being wrong isn't going to do any harm. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 18:21, 14 May 2020 (UTC)

Sorry I got annoyed. I did try to change the text to reflect the uncertainty, but got baldly reverted – and as visible in the history everybody else who's tried to make any changes has similarly been reverted. I hope you can appreciate how frustrating it is to see the great ease with which some very hasty changes have been applied across the encyclopedia, and the great difficulty everybody has had in trying to question them. And it's not that the the transcriptions are consistent now – as far as I can see, last year's change to /ɽi/ has been reverted, by various editors, on at least half the pages that use it. It's not perfect that the transcriptions are inconsistent, but it's definitely better than if they were consistently incorrect. – Uanfala (talk) 18:46, 14 May 2020 (UTC)

@Uanfala: I wouldn't care if we transcribed it as a syllabic trill, except that all the sources I'm seeing say it's not a trill. We've had claims that it's a syllabic flap, but on the face of it such a thing would be impossible. I suppose ⟨ɾ̩⟩ (to indicate a flap bordered with short indeterminate vowels that make it syllabic) would work, but we'd need to explain what such an odd transcription meant. Or ⟨ᵊɾᵊ⟩. Yeah, I know sources can be a problem right now, but there's a lot of stuff online that wouldn't normally be accessible. — kwami (talk) 02:51, 15 May 2020 (UTC)

Use ⟨r̩⟩ if the actual pronunciation is unknown, just like we use ⟨r⟩ for English [ɹ]. Even if the actual pronunciation wasn't syllabic (but rather a sequence of a consonant and a vowel, or the other way around), it wouldn't be unprecedented to transcribe it as such. For instance, some sources use ⟨r̩⟩ for what surfaces as [əɾ] in Slovene (and it's actually a better transcription, AFAICS - at least as far as phonemics are concerned. But this isn't Help:IPA/Slovene). Kbb2 (ex. Mr KEBAB) (talk) 15:25, 14 May 2020 (UTC)

The difference from English is that English WP IPA is a phonemic notation, so it doesn't matter which symbols are used. But in actual IPA, [r] is a trill, and AFAICT Sanskrit /r̥/ is not a trill. If our Slovene transcription is inaccurate, that's not a good argument to make Sanskrit inaccurate too.

Some of those sources don't even distinguish Vedic from Classical. Cardona says that Vedic was Northwestern Indic and didn't have /l/, while Classical was Central Indic and did distinguish /l/ from /r/, as well as having some influence from the northeast where they only had /l/.

r̥ and r̥̄ are metrically short and long and bear different pitches, but they are phonetically combinations of vowel and consonant segments: r flanked by very short segments which have different vowel colors in different dialects. In the dialect Śaunaka represents, r̥ consists of a consonantal segment r – produced either at the roots of the teeth or the alveolae – flanked by segments ă (one-fourth of a mora) which are produced further forward than a due to the tongue positioning for r: ără. This accords with the statement that r̥ has the same place of production as velar (jihvāmūlīya) stops k and so on – in contrast to a, said to be produced further back – and with the R̥gvedaprātiśākhya order of vowels: r̥ follows a ā. There is also evidence that r̥ was pronounced ĕrĕ and ĭrĭ. In descriptions where the place of the consonantal segment is assigned to r̥, both this vowel and r are said to be retroflex elements. Modern pronunciations of r̥ vary; e.g. ri in the midlands, ru in the southwest.

(Cardona 2003 "Sanskrit", in Jain & Cardona The Indo-Aryan Languages, p. 108-109) — kwami (talk) 00:49, 15 May 2020 (UTC)

References

I can't understand why the edits are reverted again and again. All the grammar books, whether traditional and modern, states that it is a syllabic vowel. But it is again and again being reverted to 'ɽi'. If it is pronounced so today, that doesn't mean it is it's original pronunciation. As श is pronounced nowadays as ʃ, but it's original pronunciation is ɕ. Even if in classical Sanskrit, it was 'ri', we must be nuetral. In English, the most common pronunciation of 'r' is 'ɹ', but it is transcribed as 'r'. To stay consistent with the script, we must follow 'ɹ̩' or 'ɹ', and even r̩ and r are acceptable. But noway, it could be ri. It not confuses the reader, but also makes it hazardous for the Sanskrit literature, mostly in verse, as they are set in prosodic meter which is very sensitive to the length of the vowel, as 'ri' will make previous consonant conjunct, making that syllable heavy Ashutosh Jha (talk) 19:34, 26 June 2020 (UTC)

Yeah, this conversation kind of fizzled without getting a real consensus. Ashutosh Jha, you were good to start the discussion (and come here again when asked to), but it doesn't seem like you're engaging with what people are saying.

I'll say this again. We don't want to create a disparity between the transcriptions that we have across Wikipedia and our help key. It will confuse readers. When we make changes, we want to make sure to keep things consistent.

As can be seen, it isn't true that all relevant texts say the same thing, as Kwamikagami has identified (and quoted) sources that say otherwise. How do you think we should reconcile what you've seen and what Kwami has? — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 02:24, 27 June 2020 (UTC)

r̥ and r̥̄ are metrically short and long and bear different pitches, but they are phonetically combinations of vowel and consonant segments: r flanked by very short segments which have different vowel colors in different dialects. In the dialect Śaunaka represents, r̥ consists of a consonantal segment r – produced either at the roots of the teeth or the alveolae – flanked by segments ă (one-fourth of a mora) which are produced further forward than a due to the tongue positioning for r: ără. This accords with the statement that r̥ has the same place of production as velar (jihvāmūlīya) stops k and so on – in contrast to a, said to be produced further back – and with the R̥gvedaprātiśākhya order of vowels: r̥ follows a ā. There is also evidence that r̥ was pronounced ĕrĕ and ĭrĭ. In descriptions where the place of the consonantal segment is assigned to r̥, both this vowel and r are said to be retroflex elements. Modern pronunciations of r̥ vary; e.g. ri in the midlands, ru in the southwest.

Yes, it is a good point, but the source clearly says that, Modern pronunciations of r̥ vary; e.g. ri in the midlands, ru in the southwest. see, it says ′Modern Pronunciation′. About Classical or Earlier pronunciation, it says, flanked by segments ă (one-fourth of a mora) which are produced further forward than a due to the tongue positioning for r: ără. This accords with the statement that r̥ has the same place of production as velar (jihvāmūlīya) stops k and so on – in contrast to a, said to be produced further back – and with the R̥gvedaprātiśākhya order of vowels: r̥ follows a ā. There is also evidence that r̥ was pronounced ĕrĕ and ĭrĭ. Can you assert any one transcription to ′r̥ ′? As it has several transcriptions, not one, ără or ĕrĕ or ĭrĭ. And none of them could be added as the IPA key, it is not suitable. ′ɹ̩′ or ′r̩′ is its common transcription. it clears 'ri' transcription is never suitable, as it is modern pronunciation. The pronunciations ără or ĕrĕ or ĭrĭ could be true, but it is secondary pronunciation. Originally it is indeed syllabic. And you can't assign thse in transcription. And about 'ri', it is not even uniform, it is 'ru' in southern parts.Ashutosh Jha (talk) 10:43, 27 June 2020 (UTC)

@Ashutosh Jha The only problem with your edits is that there is no attempt by you whatsoever to make your edits on this IPA help page consistent with the Phonology section in the Sanskrit article. Have you even read that section before making your changes to the r symbol in this IPA page? Because I have read that section and I have not seen any mention to [r] being pronounced or transcribed as [ɹ] (in fact the information in that section suggests it can be pronounced as either [ɾ̪], [ɾ] or [ɽ] without any mention of [ɹ] and even Vedic Sanskrit#Phonology does not mention [ɹ]) along with these sections not mentioning either [ɹ̩] or [ɹ̩ː] as syllabic vowels (with Vedic Sankrit saying that it is transcribed as [r̩] or [r̩ː]). I wouldn't have had a problem with these edits if the [ɹ] symbol was mentioned in the Phonology sections in that article or if you had edited those Phonology sections first (with reliable sources) before making your changes to Help:IPA/Sanskrit. Broman178 (talk) 09:51, 29 June 2020 (UTC)

Initially I had made the same edits as you did. I had too used 'r' instead of alveolar approximant. But, it was reverted and they said that it was an untrilled approximant. I was intolerant to the transcription 'ri', and I wanted to change it to syllabic. So, I had only one option, to transcribe it as alveolar approximant. So, I did it. However, Iam not sure that is it approximant or a tap. I have no problem. But I wanted syllabic transcription, as the sources suggest. You can see it above in this section, that I had initially transcribed it as r, but I was said to transcribe it as alveolar approximant. Ashutosh Jha (talk) 14:46, 30 June 2020 (UTC)

Thanks for your reply, I have made a few tweaks to this IPA help page yesterday, While I've kept most of the changes you have made (while mentioning some of the previous mentioned phonemes in footnotes) because I prefer to avoid edit wars when possible, I have replaced the approximant symbol with the trill symbol because in my opinion, there is no point transcribing /r/ as [ɹ] if the approximant symbol is not listed or mentioned in either of the two Phonology sections I've mentioned. I believe consistency should be maintained across Wikipedia articles for Sanskrit and mentioning the approximant symbol here without mentioning it first in those pages in my opinion is extremely confusing for readers and is inconsistent with those two sections (I think whoever reverted your edit when you first added [r] was wrong to have reverted it). I think I would prefer to keep it transcribed as [r] in this page unless there are reliable sources which support the alveolar approximant [ɹ] (for which it should be mentioned in those Phonology sections first before being added to this page), otherwise I will have to revert those changes. Broman178 (talk) 08:57, 1 July 2020 (UTC)

@Broman178: Are you saying that the only hesitation to fix the transcription is to change it in the other places first? Because @Ashutosh Jha is correct here; ऋ, ॠ, ऌ, ॡ are all defined as vowels in the Pāṇinian grammar and practically every other source on Sanskrit. A sound made, linguistically speaking, when the tongue touches any part of the inside of the mouth cannot be called a vowel (which also happens to be how sounds are classified in Pāṇinian grammar). This is corroborated by the fact that you can't have short and long versions of stops/flaps (it's tautological; it's literally in the name). [r] is a trill, so it definitely does not quality as the proper transcription of ऋ or ॠ. Even the only reference on the main page corroborates the transcription of the vowels as [ɹ]. If sources are what's necessary, I can definitely provide more. Getsnoopy (talk) 01:27, 31 October 2020 (UTC)

@Getsnoopy: Even if those are correct, there is no point transcribing these symbols if they aren't supported by either the Phonology sections in Classical Sanskrit or Vedic Sanskrit (I have seen no mention of [ɹ] in the Phonology section for either the Vedic Sanskrit or Classical Sanskrit articles, just [r] was mentioned in Vedic Sanskrit and [ɽ] for Classical Sanskrit, so why should we transcribe it here if it isn't supported by those articles?). At the moment there seems to be a confusion whether to transcribe symbols for Classical Sanskrit or Vedic Sanskrit, because at first this article transcribed symbols for Classical Sanskrit but recent edit warring has resulted in symbols for Vedic Sanskrit being added in this Help Page as well.

What I suggest is, make sure these Phonetic symbols are mentioned in the Phonology sections for both Classical Sanskrit and Vedic Sanskrit and then mention the symbols here possibly in a separate section for Classical Sanskrit and one for Vedic Sanskrit (similar to how e.g. Help:IPA/Persian has separate sections for Farsi, Dari and Tajik) because they are transcribed differently for both and we also have to be consistent with the Phonology sections for those articles. We must remember the whole point of an IPA Help page is not to describe or just mention phonetic symbols for Wikipedia users (in any case, the Phonology sections of those articles are good enough for that), but to help editors transcribe IPA symbols for pronunciation across Wikipedia articles for these languages, and having symbols which aren't supported by the Phonology sections in these languages is bound to confuse editors. Broman178 (talk) 06:36, 31 October 2020 (UTC)

@Broman178: Agreed. We should change the other articles first and then the IPA page. I like your idea of separating the two à la the Greek page or the Portuguese page, though it wouldn't change the outcome either way in this particular case. Getsnoopy (talk) 06:03, 18 November 2020 (UTC)

Unnecessary reversions of edits.

Many times, my edits are being reverted, without any reason. And the reason given here is that, it must be discussed. If we have all the sources, it is known to every person being concerned to the language, why discussion is needed. And about 'r̥', I can't understand, what's the problem? All the grammar books, whether traditional and modern, states that it is a syllabic vowel. But it is again and again being reverted to 'ɽi'. If it is pronounced so today, that doesn't mean it is it's original pronunciation. As श is pronounced nowadays as ʃ, but it's original pronunciation is ɕ. Even if in classical Sanskrit, it was 'ri', we must be nuetral. In English, the most common pronunciation of 'r' is 'ɹ', but it is transcribed as 'r'. To stay consistent with the script, we must follow 'ɹ̩' or 'ɹ', and even r̩ and r are acceptable. But noway, it could be ri. It not confuses the reader, but also makes it hazardous for the Sanskrit literature, mostly in verse, as they are set in prosodic meter which is very sensitive to the length of the vowel, as 'ri' will make previous consonant conjunct, making that syllable heavy. And, the other edits. I have added a note to coronal consonants, that they are dental, but originally may have been alveolar. I've changed r to ɹ, and of course ready to discuss on that. Added other pronunciation of Visarga, we could add it in a note, it could be discussed. And also, changed the text 'Classical Sanskrit' to just 'Sanskrit' to make the IPA key broader, as the page is named 'Help:IPA/Sanskrit', not Classical Sanskrit. In short, I'm ready for any discussion, instead of the pronunciation of syllabic r as 'ri'. I'm afraid, if this behaviour is not going to stop, no further information could be put on it. It's not the single page, where it is done. I hope, this is not going to repeat. Ashutosh Jha (talk) 19:31, 26 June 2020 (UTC)

The reasons have been given. I'll restate them briefly

The focus of the key is classical Sanskrit because our transcriptions are classical. You're taking out "classical" to try to change the scope without consensus.

There is not a consensus to change the syllabic r as you have changed them. This is the subject of the above discussion

We try to avoid unnecessary phonetic details. The difference between a dental and alveolar consonant has already been considered unimportant enough to notate and also unimportant enough to mention. It won't help readers understand our transcriptions or editors in how to transcribe.

As far as I can tell, the note about postvocalic [h] doesn't help editors transcribe, since the writing system already indicates this sound directly.

The statement An allophone of [e] and [o] after [e] and [o] respectively doesn't make any sense. I was under the impression the sound in question, ɐ, is simply just short a.

— Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 02:38, 27 June 2020 (UTC)

I have just reverted your edits again because I checked the Phonology section in the Sanskrit article carefully and I saw no mention of symbols like [x], [ɸ] and [ɹ] anywhere in that article which means adding those symbols here now creates a lot of inconsistencies with the Phonology section in that article since they are not listed phonemes of Sanskrit. Unless we establish consensus here first on this matter and unless there are strong references to support those symbols, I definitely oppose the mention of them here, especially if there is no mention of them in Sanskrit Phonology. Broman178 (talk) 09:05, 27 June 2020 (UTC)

The allophones [x] and [ɸ] were present in the early language (you can read about them at Vedic Sanskrit#Phonology), but as far as I'm aware we don't go that far when transcribing on Wikipedia, and if these symbols aren't used in transcriptions here, there's no need to list them on this page. The statement An allophone of [e] and [o] after [e] and [o] respectively makes perfect sense: just think of the language not liking [ee] and [oo] clusters. And as for [ɽi], as clear from the section above, there's absolutely no reason why we should continue to be using that. – Uanfala (talk) 10:24, 27 June 2020 (UTC)

We are transcribing Classical Sanskrit, not Vedic Sanskrit.

Also, the reason that the business about e and o doesn't make sense is that a) the bracketing is incorrect, as you don't have allophones of phones but phonemes and, more importantly, b) the sound in question is short a. If that is the allophone of these two sounds in that context, what is the symbol we're using for short a? — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 15:21, 27 June 2020 (UTC)

ɐ is definitely an allophone of e and o. As in ते अब्रुवन् is तेऽब्रुवन्, and नित्यः अहं is नित्योऽहं. Avagraha is used so that the reader could know that it is e, but here an allophone of ɐ. Ashutosh Jha (talk) 11:24, 28 June 2020 (UTC)

And if you think that allophones of Visarga should be removed, then, provide it ina note. If you think I'm wrong, why don't you correct it, why you just revert it. Ashutosh Jha (talk) 11:27, 28 June 2020 (UTC)

This is getting tiresome. I'll say it again: Our notes should be limited to what helps:

1. editors transcribe Sanskrit from Devanagari or IAST romanization to IPA
2. readers understand our transcriptions

The note in question regarding Visarga does neither. It's just an interesting factoid about phonemicity. The note on [ɐ] is also not helpful in the ways I've already identified. If there's some kernel of truth to a statement about allophony of [eː] and [oː] (and if so, citation needed), it's muddled by poor wording. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 13:24, 28 June 2020 (UTC)

I have reverted the help page, if you think anything is wrong, or against 'so called' consensus, please remove those one by one manually and not revert that. And make seperate sections in the talk page to discuss that. And please, even if you think it necessary to revert the edit, remove all the edits, but provide correct transcription to 'ṛ,ṝ and ḷ'. Ashutosh Jha (talk) 15:27, 28 June 2020 (UTC)

I'm not sure what you're saying here. The opposition to your edits has been explained already. It kind of looks like you've ignored most of what people are saying, so I'm not really up for laying out these arguments again for you to ignore them again. Read and respond to what people have already said. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 15:58, 28 June 2020 (UTC)

Ignore all my arguments, but please give me the reason of maintaining the transcription of 'ṛ' as 'ri'. It seems you have not read the previous section [of Syllabic r in Sanskrit|]. I have cleared all the arguments. For my other edits, I'm surely ready for discussion in the talk page. I'm now only changing the transcriptipn for 'ṛ,ṝ and ḷ' and going to correct formatting, and I'll make any further edit only after discussion, such as notes on Visarga, or on allophones of 'ɐ'.Ashutosh Jha (talk) 16:23, 28 June 2020 (UTC)

Australian, Scottish, etc.

@Uanfala, since the North American pronunciation was used for consonants like ɖ why not the Australian pronunciation for /ɐ/ and /a:/? AE has /ɐ/ but NAE doesnt have [ʈ, ɖ] etc and words with /kʂ, tr, ɟɲ, ɕr/ are common so why not those? ऌ was also the short version of ॡ so it should be /l̩/ AleksiB 1945 (talk) 20:50, 1 August 2021 (UTC)

I have no strong opinions about whether we should make reference to Australian or Scottish varieties – my inclination is to avoid anything other than North American and RP as that may not be familiar to all readers, but I'd happily defer to the editors who look after the broader set of IPA pages. As for /tr/, /kʂ/ and the like – these are consonant clusters, and there's no reason to list them – the fact that the corresponding ligatures in Devanagari are less straightforwardly decomposable is beside the point. Thanks for pointing out the error with /l̩/ – I'll fix it now. – Uanfala (talk) 21:00, 1 August 2021 (UTC)

English approximations

@Aeusoes1, why? as i said in the summary 1) most people (if some do) dont pronounce canyon with a [ɲ] or ahead with a [ɦ] its either [n] or [nʲ] for canyon and [h] for ahead 2) burnt has a partial retroflex n only in rhotic accents, even the rhotic accents dont have a pure [ɳ] 3) t in stable or table isnt retroflex in any of the dialects 4) ch in chew is aspirated and c/ch are distinguished in sanskrit, how is the inaccurate? AleksiB 1945 (talk) 21:14, 2 August 2021 (UTC)

The column in question is "English approximations." Saying an approximation is "roughly like" is redundant. Taking a closer look at the state of the guide, there were other instances of "roughly like" or "somewhat like" that I'd also like to see removed.

/t/ and /d/ are not retroflex in any referent dialect we use for these guides. Trying to imply some sort of allophonic retroflexion after a coda /r/ is poorly motivated. On top of that, portal features flapping.

I dislike that so many of the sounds have "no English equivalent" on them with the logic that e.g. the gh of loghouse is not exactly the same as the [ɡʱ] of Sanskrit, Hindi, etc. This fetishization of accuracy comes at the expensive of helpfulness. If it's good enough for the source cited in this edit summary, it's good enough for us. I move we restore the earlier examples for the murmured consonants. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 04:00, 3 August 2021 (UTC)

@Aeusoes1, what about rechew and port? — Preceding unsigned comment added by AleksiB 1945 (talk • contribs) 23:10, 3 August 2021 (UTC)

I don't know about other people, but for me rechew features aspiration. Port is pronounced with a final glottal stop (or prominent glottalization). — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 03:57, 4 August 2021 (UTC)

Bear in mind the English approximations for the murmured stops were removed in most of the Indic language IPA Help pages (Help:IPA/Hindi and Urdu, Help:IPA/Bengali, Help:IPA/Nepali, Help:IPA/Odia etc.) by Sol505000 (would be interesting to hear what Sol505000 thinks about this) so I would suggest that if they are restored here, it would probably have to be restored in all those other Help pages too. However, that being said, I do really personally dislike the dental stops being approximated with alveolar stop sounds from English without any clarification that the sound is dental because that to me is just completely inaccurate and misleading (anyone who thinks the dental and alveolar stops are the same should get their ears checked) - it would be more accurate in my opinion to say e.g. for [d]: "do, but dental" or "do (dental)" in the approximations for the dental stops similar to what Help:IPA/Irish has done. Broman178 (talk) 09:25, 10 August 2021 (UTC)

Do for Sanskrit /d/ (which is dental, as you say) is about as inaccurate as birdhouse for /ɖʱ/ (IMO birdhouse is further away from the Sanskrit target, but anyway). If birdhouse is close enough (which I'm not gonna argue that it is not), then do is close enough as well, but ado would be even better as the Sanskrit sound is fully voiced, like intervocalic /d/ in English. The lack of full voicing (especially a complete lack of voicing) is just as much of a pronunciation mistake as the alveolarity of /d/. Sol505000 (talk) 09:41, 10 August 2021 (UTC)

I'm fine with ado etc to reflect voicing of English's lenis consonants. I'm also warm to ideas that will help us parse the dental/retroflex distinction, though it's a tall order to do so when most of our audience won't be primed to understand that distinction. It would be a good idea to hash something out here in the talk page and get explicit consensus than go back and forth with edits. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 04:16, 13 August 2021 (UTC)

What about (American English) parts and partake for [ʈ] and [ʈʰ]? — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 16:40, 13 August 2021 (UTC)

"parts" to be seems okay but I'm not really sure about "partake" when you consider the flapping which usually occurs between vowels in American English (I will admit, I've rarely heard American English speakers not flap the "t"s in the middle of words). Regarding the murmured consonants, I would agree with Sol505000 that words like "birdhouse" would not have full voicing of the "h" but I probably would still say its a bit more accurate than using alveolar stop approximations for the dental stops, especially if there is no clarification that it is dental (one of the reasons why I decided to add that in most of the Indic Language IPA Help pages after looking at the IPA Help page for Irish), because I'd say in fast speech, there might be some voicing in that word which would sound similar to the murmured consonant while dental stops to my ears sound quite distinct from the alveolar stops in English whether its slow or fast speech. Regarding the retroflex/postalveolar stops, I think it would be better to say for the English approximations e.g. "do, but retroflex" or "redhead, but retroflex" similar to what I did for the dental stops rather than those American English examples (especially the second one for the murmured consonant) although if others support the examples which have been added, then it might as well stay. Now it seems Aeusoes1 has restored all the murmured approximations in most of the Indic Language IPA Help pages so if that is to change, it would probably need a further discussion here or elsewhere. Broman178 (talk) 09:11, 16 August 2021 (UTC)

As a speaker of American English, I can assure you that partake does not typically feature flapping (I think if someone said it with flapping, I would probably hear it as party ache). The flapping that occurs in dialects that feature it has a complex relationship with stress.

I would rather avoid jargon like "retroflex" in the approximations column if we can. It really will not help lay readers understand. It feels a little strange not to reflect such a salient contrast in our examples, but we may be depending too much on the approximations to convey the contrast when a footnote might be a better choice. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 19:54, 16 August 2021 (UTC)

I guess because you speak American English, you'll be more familiar on that area than me so I won't argue on that as in any case, I'm from England so I may not be as familiar on that area. Now regarding whether to add things like retroflex for the English approximations, I personally have no problem with that but one alternative I would suggest is adding "Indian" before approximations like "do" for the retroflex stops because looking at the retroflex stop article, the only example of English given there is "dine" for Indian English (with "time" being otherwise used for the voiceless retroflex stop) and there is no example of e.g. "partake" for American English so I personally would prefer something like that (or "do" but with the tongue further back") if we are to avoid terms like "retroflex" in the approximations.

And for the dental stops, I'd probably also suggest adding a dialect with actually uses dental stops in place of the usual alveolar stops used for most English dialects (the only one which actually seems to use a dental stop in place of the usual alveolar stop is the Ulster dialect although that only in words like "dream"), although for that, I'd also suggest e.g. "do, but with the tongue touching the front two teeth" though that probably would add clutter. However, if others agree that "partake" is a suitable example for the retroflex stop or that the alveolar stop approximations are okay on their own without "dental" written by them, then I'm happy for them to stay. Broman178 (talk) 16:44, 17 August 2021 (UTC)

"Irish thing" etc. for dental stops may indeed be a good idea. For a dental nasal we may use tenth. For that matter, we can also use eighth for the plosive, but it's difficult to distinguish it from the fricative in orthography. Using breadth or width avoids this problem, but I wonder how many speakers realize the d is voiceless in those words. Nardog (talk) 17:00, 17 August 2021 (UTC)

Regarding dental stops, if by chance we decide not to use alveolar stop approximations for them, we could also use "Indian" for words like "thing" as well as "Irish" and like I said above using "Indian" for words like "dine" or "time" for retroflex stops would also be okay because words like "art" "partake" are not mentioned in any of the retroflex stop articles but Indian English examples have been provided for them. The only problem though is in most other articles, alveolar stop approximations are used for dental stops in non-Indic language IPA Help pages (like Help:IPA/Spanish, Help:IPA/French, Help:IPA/Italian, Help:IPA/Japanese etc.) without much difficulty (even though for them I still think its inaccurate - French though would be a different case because the pronunciation of the dental stop does at times lean more towards an alveolar stop at times from what I've listened) while for the Indic languages and even Irish, this issue would be more noticeable because in many Indian English and Irish English dialects, the dental stops are used in place of the dental fricatives in words like "the" "thin" etc, which is I believe neutrality in the approximations would be important for them. Broman178 (talk) 09:36, 19 August 2021 (UTC)

Many mistakes

इ is listed as [ɪ] while ई is listed as [i:]. Similarly, उ is given as [ʊ] and ऊ as [u:]. This is demonstrably untrue. They have the same vowel quality. इ is [i] and ई is [i:], same with उ [u] and ऊ [u:]. Panini establishes this in the Astadhyayi. अ and आ do differ in vowel quality as well as length and due to how Panini formulates his rules he has to account for this fact. The sandhi rule for joining like vowels is formulated thus: When two homogeneous vowels meet (two vowels with the same quality, irrespective of length), replace them both with one long homogeneous vowel. So इ/ई + इ/ई will become one ई. However, as अ and आ differ slightly in pronunciation this creates a problem, as they are not technically homogeneous. Panini has to work around this fact in the Astadhyayi to allow अ/आ + अ/आ to yield आ, first pretending at the onset of the rules that they were homogeneous, then in the very last verse resetting अ to its normal pronunciation. He does not do this with any other vowels but treats इ/ई and उ/ऊ as identical in pronunciation with no problems. Nor does the Paniniya Shiksha mention any difference in vowel quality between them, only अ and आ. Verse: अ अ । (Ast. 8.4.68)

As for ए, ऐ, ओ, and औ, Panini and the Pratishakhyas disagree slightly. According to Panini, ए is one half matra in the throat and one and a half on the palate (so closer to [ɐi:]) and ओ is one half matra in the throat and one and a half on the lips (so like [ɐu:]). The Pratishakhyas disagreed and said that they each have one place of articulation which two matras (thus [e:] and [o:]). They are in agreement regarding ऐ and औ. Both assert that ए is one matra in the throat and one on the palate [ai] and औ is one in the throat and one on the lips [au]. Both Panini and the Pratishakhyas are quite old, so in my opinion either pronunciation of ए/ओ is acceptable. Verse: अर्धमात्रा तु कण्ठ्यस्य एकारौकारयोर्भवेत् । ऐकारौकारयोर्मात्रा तयोर्विवृतसंवृतम् ॥ (PS. 19)

Transcribing ऋ as [r̩] is a bit puzzling. The description of the sound is correct, but I don't see why it should be regarded as a syllabic consonant at all and not a true vowel, as according to the Shiksha it is अस्पृष्टः (untouched). It approaches the place of articulation of र् but does not make contact. That many modern Indian languages pronounce it as ri or ru is unsurprising, given that even in Classical Sanskrit ऋ is a weak vowel that tends to make this change in some derivations, such as कृ becoming क्रियते and कुरुते. Also many roots which end in long ॠ tend to change to ईर् (as in तीर्यते from तॄ) or ऊर् (as in पूरयति from पॄ). As for ऌ, there was again some disagreement. Some grammarians simply didn't recognize it as a vowel, some regarded it as an untouched (अस्पृष्टः) vowel, and some as more like a syllabic ल् that was touched (which might explain why some did not regard it as a vowel at all). Verse: अचोऽस्पृष्टा यणस्त्वीषन्नेमस्पृष्टाः शलः स्मृताः । शेषाः स्पृष्टा हलः प्रोक्ता निबोधानुप्रदानतः ॥ (PS. 38)

As for र्, many places of articulation are given. Panini regards it as a cerebral and does not recognize the alveolar place of articulation at all, placing र् further back in the mouth than the Spanish r, for example. However, the Pratishakhyas gave various places of articulation for it. Some considered it dental, alveolar, or even velar. But the Paninian pronunciation is more like [ɽ]. Verse: अष्टौ स्थानानि वर्णानामुरःकण्ठः शिरस्तथा । जिह्वामूलं च दन्ताश्च नासिकोष्ठौ च तालु च ॥ (PS. 13)

The source provided is filled with many other errors and should not be regarded as accurate at all. The stress rule it gives is simply nonexistent. It states that Vedic pitch accent is unpredictable (which is true) but then says that there is a stress accent present in Classical Sanskrit which follows a rule. However, the rule it gives does not exist in Sanskrit. It is the Latin stress rule that western scholars adopted for Sanskrit as a matter of convention, but it has never been true for Sanskrit. In fact, stress accent in Sanskrit is a rather inconsequential thing for which there does not exist any simple rule at all. Furthermore, it states many false things about the anusvara. There is a long-standing conflation of the terms anusvara, anunasika, and rangavarna that has confused many people. The anusvara is not a nasalization of the vowel, it is a nasal sound pronounced after the vowel. Hence it is referred to in the old grammatical literature by its full name "अनुस्वारनासिक्यः" literally meaning "nasal sound after the vowel". According to the Paniniya shiksha and the Pratishakhyas, the anusvara (along with the yamas) has only the nose as its place of articulation, not the nose and mouth as it would be in a nasalized vowel. Thus it is a "pure nasal" sound in which the sound is uttered directly and entirely from the nose. The term "anunasika" has also lead to confusion. Panini only uses it to mean any nasal sound. Indeed, the nasal consonants are called the anunasika consonants, but people have gotten confused by the term thinking it too means a nasalized vowel. However, that is not how Panini uses the term. The nasalization of a vowel is called the रङ्गवर्णः in the Shiksha. Verse: अनुस्वारयमानां च नासिका स्थानमुच्यते । (PS. 22a)

Another egregious error is transcribing the original pronunciation of the visarga as [x] which is absurd. [x] is the जिह्वामूलीयः (literally meaning "root of the tongue") which is a breath restricted by the back of the tongue and occurring as an allophone of the visarga when followed by प् or फ्. The unvoiced visarga [h] is contrasted with the voiced hakara [ɦ]. Some regarded the hakara as only being half voiced, but in either case, it is in the throat when followed by a vowel and in the chest when combined with a nonvowel. Verse: हकारं पञ्चमैर्युक्तमन्तःस्थाभिश्चसंयुतम् । औरस्यं तं विजानीयात्कण्ठ्यमाहुरसंयुतम् । (PS. 16) 2607:D580:162:BB00:EC0A:B4B7:C8BC:BDAE (talk) 07:17, 25 November 2022 (UTC)

Consensus?

@1Firang: There is no consensus in Help:IPA/Hindi_and_Urdu. You have just kept on adding your changes with "brute force" without consensus. So please don't change anything before reaching a genuine consensus (not just among you and yourself). Stop your edit spree and read Help:IPA and some introductory books about IPA and phonetics first. Competence is required. Austronesier (talk) 14:30, 11 June 2023 (UTC)

Austronesier, I did ask at Help_talk:IPA/Nepali#Corrections and Broman178 responded after which I carried out the corrections.-1Firang (talk) 15:39, 11 June 2023 (UTC)

Kwamikagami, with this you reverted the approximations, "other, the (with th-stopping)" with "do", removed the "(but dental)" after the approximation "redhead", replaced "jazz" with "juice", "derrickhand" with "kin", "panther" with "tenth", "minion" with "enjoyable", "under" with "burn", "pot (but more plosive)" with ""pan", "borrow" with "roti", "birth and three" with "stable", "think" with "table", "anthill" with "table", "beachhead" with "'chew" and "high" with "hi" but I want to discuss each of those.-1Firang (talk) 16:40, 11 June 2023 (UTC)

If you want to discuss you changes, remember that the onus is on you to show that these are improvements. So explain for every single instance why you want to change A_old to A_new, B_old to B_new etc. and wait to gain consensus. Until then, don't change anything. You have turned four pronunciation keys into trainwrecks. –Austronesier (talk) 16:50, 11 June 2023 (UTC)

@1Firang: Please change the phrasing of Kwamikagami replaced... etc. You are the one who wants to introduce changes here. Kwamikagami and I have merely salvaged the last good verion. And don't forget to explain your changes (onus, remember?). "More appropriate" and "unacceptable" is not enough. –Austronesier (talk) 18:01, 11 June 2023 (UTC)

 Done - see below.-1Firang (talk) 18:31, 11 June 2023 (UTC)

Requesting changes (with reasons)

@Austronesier: Like I said, Broman178 responded and I carried out the corrections after that. Now please be kind enough to discuss these changes:-

The approximation for द should be "other, the (with th-stopping)" or "do (but dental)" as the IPA for that is "ð",

The "(but dental)" after the approximation "redhead" should be re-instated for the Sanskrit alphabet ध as it will become ढ otherwise,

Kwamikagami replaced "jazz" with "juice" for the Sanskrit alphabet ज and both are fine, so I will not ask you to change the same,

Kwamikagami replaced "derrickhand" with "kin" for the Sanskrit alphabet ख but if "derrickhand" is unacceptable, we should have "kettle (but more plosive) as the "k" in "kin" is not plosive,

For the Sanskrit alphabet न, "panther" is better than "tenth" as the "n" in the former is more pronounced and was already accepted by consensus in the Hindi and Urdu IPA help page (Help:IPA/Hindi_and_Urdu) before I started editing these pages,

For the Sanskrit alphabet ञ, "minion" is more appropriate, as compared to "enjoyable" as the "n" in the latter is more like न,

For the Sanskrit alphabet ण, "under" is more appropriate, as compared to "burn" as the "n" in the latter is more like न,

For the Sanskrit alphabet प, "pot (but more plosive)" is more appropriate, as compared to ""pan" as the "p" in "pan" is not plosive,

For the Sanskrit alphabet र, "borrow" is more appropriate, as compared to "roti" as "roti" is not a proper English word,

For the Sanskrit alphabet त, "birth and three" are more appropriate, as compared to "stable" as the "t" in the latter is more like ट,

For the Sanskrit alphabet थ, "think" is more appropriate, as compared to "table" as the "t" in the latter is more like ट,

For the Sanskrit alphabet ठ, "anthill" is more appropriate, as compared to "table" as the "t" in the latter is more like ट,

For the Sanskrit alphabet छ, "beachhead" is more appropriate, as compared to "'chew" as the "ch" in the latter is more like च and

For the Sanskrit alphabet ऐ, "high" is more appropriate, as compared to "hi" as the "i" in the latter is more like आइ (IPA: aɪ) according to this.-1Firang (talk) 17:50, 11 June 2023 (UTC)

I suspect the problem may be that you do not know English as well as you think you do. For instance, in your last item you say that "high" is closer in pronunciation to ऐ than "hi" is, when the two are in fact perfect homophones -- as you could verify by checking the very source that you cited.

I object to any disyllabic equivalents to what are single consonants in Sanskrit. I remember seeing such things in highschool and being completely confused by them.

Then there's your claim that थ is like 'think' but त is like 'three' -- neither are fricatives, and what is supposed to be the difference between the 'th' of 'think' and of 'three'? Certainly not aspiration!

You might have a point about adding a note to the existing key word (not replacing it) about th-stopping, for those readers who know what it is, but really, anyone who knows what th-stopping is already understands the IPA and doesn't need this key.

The rest is mostly nonsense.

Consider also that we don't even have "a" pronunciation of Sanskrit: we have reconstructions and modern approximations, the latter of which are influenced by the speaker's native language. — kwami (talk) 21:34, 11 June 2023 (UTC)

@Kwamikagami: The IPA for"high" is "hʌɪ" according to this and the IPA for"hi" is "haɪ" according to this so they are not "homophones". The 'th' of 'think' is more plosive than that of 'three'. You say, "You might have a point about adding a note to the existing key word (not replacing it) about th-stopping", so please let me know if you agree with the addition of "(with th-stopping)" after "the" for the approximation for द, with another approximation being "other, or adding "(but dental)" after the "do" that we have as the approximation for द at present, as the IPA for that is "ð" and adding "(but dental)" after the approximation "redhead" for the Sanskrit alphabet ध as it will become ढ otherwise.-1Firang (talk) 04:32, 12 June 2023 (UTC)

You should not argue about the phonetics of a language that you do not speak fluently. You should also read your sources before using them to claim that native speakers pronounce their language wrong. You quite literally do not know what you're talking about.

I don't agree with any of your suggestions. — kwami (talk) 04:37, 12 June 2023 (UTC)

You need to read all my suggestions instead of dismissing them as, "The rest is mostly nonsense".-1Firang (talk) 04:41, 12 June 2023 (UTC)

When I come across nonsense, I generally skip the rest as a waste of time. How about you present two suggestions that you are satisfied are sensible. — kwami (talk) 04:54, 12 June 2023 (UTC)

@Kwamikagami: Please let me know if you agree with the addition of "(with th-stopping)" after "the" for the approximation for द, with another approximation being "other, or adding "(but dental)" after the "do" that we have as the approximation for द at present, as the IPA for that is "ð" and adding "(but dental)" after the approximation "redhead" for the Sanskrit alphabet ध as it will become ढ otherwise.-1Firang (talk) 05:04, 12 June 2023 (UTC)

The IPA for द is not [ð]! Not even close.

Currently we use GA rhoticism to carry the difference between dental and post-alveolar. That's not a very satisfactory solution IMO, but I can't think of anything better.

It would probably be okay to add "(but dental)" after the existing examples.

I wonder why we don't distinguish the series for the voiceless plosives. — kwami (talk) 05:15, 12 June 2023 (UTC)

I added (but dental). I also tried "width" for the dental d, but that's probably not good, because it varies between /d/ and /t/. Maybe we can find a better example? — kwami (talk) 05:20, 12 June 2023 (UTC)

@Kwamikagami: अदा is written as ʌðɑː according to this article, so please accept that the English IPA for द is ð.-1Firang (talk) 05:32, 12 June 2023 (UTC)

I deleted that unreferenced pronunciation as most likely wrong. And even if she does have an odd English pronunciation to her name, that doesn't mean that basic English words are the same. — kwami (talk) 06:03, 12 June 2023 (UTC)

Comment: As of my making this comment I am uninvolved in this particular dispute. But I felt the need to mention that firang themself added the unreferenced IPA pronunciation they attempted to cite. This is, if not a violation of policy (such as OR), at the very least an incredible instance of editorial misconduct. Googleguy007 (talk) 21:08, 15 June 2023 (UTC)

repetitive, WP:BLUDGEON

The approximation for द should be "other, the (with th-stopping)" or "do (but dental)" as the IPA for that is "ð", The "(but dental)" after the approximation "redhead" should be re-instated for the Sanskrit alphabet ध as it will become ढ otherwise, You replaced "jazz" with "juice" for the Sanskrit alphabet ज and both are fine, so I will not ask you to change the same, You replaced "derrickhand" with "kin" for the Sanskrit alphabet ख but if "derrickhand" is unacceptable, we should have "kettle (but more plosive) as the "k" in "kin" is not plosive, For the Sanskrit alphabet न, "panther" is better than "tenth" as the "n" in the former is more pronounced and was already accepted by consensus in the Hindi and Urdu IPA help page (Help:IPA/Hindi_and_Urdu) before I started editing these pages, For the Sanskrit alphabet ञ, "minion" is more appropriate, as compared to "enjoyable" as the "n" in the latter is more like न, For the Sanskrit alphabet ण, "under" is more appropriate, as compared to "burn" as the "n" in the latter is more like न, For the Sanskrit alphabet प, "pot (but more plosive)" is more appropriate, as compared to ""pan" as the "p" in "pan" is not plosive, For the Sanskrit alphabet र, "borrow" is more appropriate, as compared to "roti", as "roti" is not a proper English word, For the Sanskrit alphabet त, "birth and three" are more appropriate, as compared to "stable" as the "t" in the latter is more like ट, For the Sanskrit alphabet थ, "think" is more appropriate, as compared to "table" as the "t" in the latter is more like ट, For the Sanskrit alphabet ठ, "anthill" is more appropriate, as compared to "table" as the "t" in the latter is more like ट, For the Sanskrit alphabet छ, "beachhead" is more appropriate, as compared to "'chew" as the "ch" in the latter is more like च and For the Sanskrit alphabet ऐ, "high" is more appropriate, as compared to "hi" as the "i" in the latter is more like आइ (IPA: aɪ) according to this.-1Firang (talk) 06:11, 12 June 2023 (UTC)

You're wasting our time by repeating the same nonsense, without knowing what you're talking about. I'm done here. — kwami (talk) 06:26, 12 June 2023 (UTC)

@Kwamikagami: I know what I am talking about. You either need to read and respond to your fellow editors and build a consensus or stop reverting their edits. Please read WP:IDONTLIKEIT.-1Firang (talk) 07:07, 12 June 2023 (UTC)

No, you very obviously don't have a clue what you're talking about. This was confirmed by your inability to transcribe the simple sound recording that you selected. The fact that you are too ignorant to recognize your own ignorance, and to remedy it by educating yourself, is not my problem. — kwami (talk) 07:13, 12 June 2023 (UTC)

I will answer only two points now the rest to follow later. Don't make change unilateral changes until then:

• hi → high for ऐ: N The change is pointless, as both are pronounced in the same way [hai]. This is quite symptomatic for your overall competence: you look up https://dictionary.cambridge.org/dictionary/english/hi, but not https://dictionary.cambridge.org/dictionary/english/high, which would immediately tell you how pointless this request is.
• pan → pot for प: N Again, pointless. Both words have an initial aspirated plosive [pʰ] in the major varieties of English (UK/US) that most readers are familiar with. When you say that the "p" in "pan" is not plosive, this is plainly wrong.

More later. –Austronesier (talk) 07:36, 12 June 2023 (UTC)

@Austronesier: With this revert, you replaced the approximations "other, the (with th-stopping)" with "width", so please explain how that (the latter) is an approximation for द (English IPA: ð)?-1Firang (talk) 11:20, 12 June 2023 (UTC)

There is no ambiguity in "high" but with "hi" there is some ambiguity (the IPA for "hi" can be "hʌɪ" or "haɪ")-1Firang (talk) 11:32, 12 June 2023 (UTC)

For the record: द is a stop (= plosive) and has been described as such literally for millenia (since Pāṇini). [ð] is a fricative. द = IPA [ð] is wrong.

There is no ambiguity in "hi". It is a homophone of "high". –Austronesier (talk) 11:54, 12 June 2023 (UTC)

Part2

• kin → kettle for ख: N Analogous to प. Both words have an initial aspirated plosive [kʰ] in the major varieties of English (UK/US) that most readers are familiar with. The statement the "k" in "kin" is not plosive is wrong.
• Adding "(but dental)" to the approximation of ध: Y The "width" trick (where the following dental fricative [θ] triggers /d/ to be pronounced with dental [d̪]) doesn't work here, so have to make the dental pronunciation of the plosive in "redhead" explicit. It seems that Kwami has already implemented the change.
• (American) burn → under for ण: N The nasal in "under" is alveloar in the major varieties of English. In order to trigger the more retracted pronunciation of the retroflex nasal (IPA [ɳ]), the general American rhotic R-sound is a helpful trick. You're right insofar that the final nasal in "burn" will sound like न in non-rhotic accents; but that's why we specify the American pronunciation here.

More to come. –Austronesier (talk) 18:33, 12 June 2023 (UTC)

द (English IPA: ð) can have the approximations "thy", "breathe", "father"" like in the English IPA Key. Any objection?-1Firang (talk) 19:30, 12 June 2023 (UTC)

Digest this[1] first before repeating the same error over and over again. –Austronesier (talk) 19:48, 12 June 2023 (UTC)

Perhaps you should show where you got that information from.-1Firang (talk) 20:10, 12 June 2023 (UTC)

I am waiting for your "more to come".-1Firang (talk) 19:32, 12 June 2023 (UTC)

@Kwamikagami and Austronesier: You may want to read this and this - the latter shows that the IPA for द is "ð".-1Firang (talk) 19:55, 12 June 2023 (UTC)

The first source is good. You should read and understand it first before writing inaccuracies like द = [ð] again. The article Hindustani phonology has lots of wonderful sources which might elevate your grasp of phonetics to a level that will eventually enable you to make suggestions that are actually helpful. –Austronesier (talk) 20:10, 12 June 2023 (UTC)

If you hear the audio at Hindustani_phonology#Consonants, you can hear the pronunciation of दाल, the IPA for which is "d̪aːl" (and English IPA is "ðaːl").-1Firang (talk) 20:24, 12 June 2023 (UTC)

1Firang, that's not [ðaːl]. Perhaps you should learn IPA before instructing people on something you don't understand. — kwami (talk) 00:55, 13 June 2023 (UTC)

@Kwamikagami and Austronesier: At the moment, we have;-

Consonants
IPA	Nagari	IAST	English approximation
d	द	d	width
ɖ	ड	ḍ	(American)bird

The "d" in bird and width are not pronounced differently as per the chart above (copied from the Help IPA Sanskrit page) but they should be, which is why I am trying to change the "width" as an approximation.-1Firang (talk) 03:25, 13 June 2023 (UTC)

Actually, they are pronounced differently. This key is a simplistic aid for people who already know how English is pronounced but aren't familiar with the IPA. Everyone else can go by the IPA, which is primary. — kwami (talk) 03:29, 13 June 2023 (UTC)

द can have the approximations "thy", "breathe", "father"" like in the English IPA Key as the approximation "width" makes the "d" sound like ड.-1Firang (talk) 03:43, 13 June 2023 (UTC)

A challenge: Please give me an example (e.g. on YouTube) of द pronounced like "thy", in any modern Indic language. If you can't do that, you're wasting our time, and I won't bother to respond further. — kwami (talk) 03:55, 13 June 2023 (UTC)

If you hear the 2nd audio at Hindustani_phonology#Consonants, you can hear the pronunciation of दाल, the IPA for which is "d̪aːl" (no need to go off Wikipedia).-1Firang (talk) 04:56, 13 June 2023 (UTC)

So, you concede that you were wrong. I consider this thread closed. — kwami (talk) 05:04, 13 June 2023 (UTC)

@Kwamikagami: Do you perceive the difference between द and ड ? Do you understand that the "d" in "width" and "bird" are the same (pronounced similarly)?-1Firang (talk) 05:35, 13 June 2023 (UTC)

Maybe you should suggest an alternative to the approximation "width" to represent द better.-1Firang (talk) 05:38, 13 June 2023 (UTC)

The difference is obvious; the problem is that you either don't know how English is pronounced in the US/UK, or don't know the IPA. Whenever you present evidence, it proves you wrong, but you don't see it. As Austronesier pointed out above, you won't even accept the dictionaries you yourself cite as evidence. There's no point discussing this with you if you refuse to learn. — kwami (talk) 05:39, 13 June 2023 (UTC)

I have understood the IPA. Please answer the question above. Do you perceive the difference between द and ड ?-1Firang (talk) 05:45, 13 June 2023 (UTC)

@Austronesier: Can we at least change the approximation "width" to "ado" (but dental), like it was before if "thy", "father" etc. are unacceptable as approximations for the indic alphabet द?-1Firang (talk) 05:51, 13 June 2023 (UTC)

@Austronesier and Kwamikagami: You still have not explained your objections to what I asked at Help_talk:IPA/Sanskrit#Requesting_changes_(with_reasons) (although @Austronesier: attempted to) - please do.-1Firang (talk) 12:26, 13 June 2023 (UTC)

You're just repeating yourself. I'm not going to waste my time repeating myself if you can't be bothered to read what I wrote the first time. — kwami (talk) 12:57, 13 June 2023 (UTC)

You need to read all my suggestions instead of dismissing them as, "The rest is mostly nonsense".-1Firang (talk) 15:29, 13 June 2023 (UTC)

@Kwamikagami: With this revert, you changed "ado (but dental)" to "width"as an approximation for the Indic alphabet द but it is wrong, while the original approximation that was there (ado) is correct, so please self revert your edit or provide a reliable source for the same.-1Firang (talk) 10:18, 14 June 2023 (UTC)

No. — kwami (talk) 10:19, 14 June 2023 (UTC)

First of I all, I have not just attempted to answer your points. I have answered five of them in all necessary detail.

Secondly, you should be aware that we are not approximating script letters, but the sounds of languages. The Sanskrit sound that is spelled with an द when Sanskrit is written in Devanagari is a voiced dental plosive (as described by Panini and still pronounced as such in the contemporary pronunciation of Sanskrit). A voiced dental plosive is transcribed in narrow IPA as [d̪] which is the symbol [d] (by default representing a voiced alveolar plosive) with the dental diacritic ◌̪) since the is no dedicated IPA symbol for dental plosives.

Thirdly, the purpose of the fourth column is to provide English approximations for every Sanskrit sound. In some cases, we can make "perfect" approximations when English has a corresponding sound in the two common accents that we use in these approximations (US and UK), as in the case of the sound [s] स which is perfectly matched by English [s]. In other cases, the sound does not exist in the two common accents of English (or maximally as a positional allophone). This is the case with द [d̪] and ड [ɖ]. Basically, neither of these sounds exists in English, since the regular pronunciation of "d" in English is [d], a voiced alveolar plosive which lies inbetween [d̪] and [ɖ] with respect of the position of the tip of the tongue, but closer to [d̪]. Neither ado nor other have the sound [d̪] in UK or US English. The first has [d], while the latter has a voiced dental fricative [ð]. None of the two is identical to [d̪]. Now, there are three strategies to approximate this sound:

1. Use the [d] in ado, but tell our readers that you have to put the tip of the tongue a bit further to the front, viz. touching the teeth. That's the "but dental" solution.
2. Use the d-sound in "width". In the word "width", most English speakers pronounce the "d" by putting their tongue in the position of the following [θ], which results in [d̪] (but often also in [t̪] with devoicing). So the "d" in width is near-identical approximation of द [d̪]. Note that the "d" in width is never pronounced as [ɖ] in US and UK English.
3. Say "other", but with TH-stopping. Some regional accents of US and UK English (e.g. in the NYC accent) have "TH-stopping". For most speakers of these accents, the TH-sound in "other" is not pronounced as fricative [ð], but rather as dental plosive [d̪]. So when "doing" such accents, the TH in other is indeed a match to द [d̪]. Note however that not all TH-stopping have dental plosives. For some, the TH-stopped sound is alveolar like the regular [d], in which case the approximation is imperfect.

To sum up, none of these are wrong, but neither is perfect either.

Lastly: I know that in South Asian English, ado is pronounced with the ड-sound [ɖ], while other is pronounced with a द-sound [d̪]. But we cannot take this accent as a starting point for approximations that are intended to reach a global readership. Readers from China, Brazil, Russia etc. are primarily familiar with UK/US English and take one of these as "target" accents when learning English. Not to speak of our readers who are native speakers of English. –Austronesier (talk) 17:14, 14 June 2023 (UTC)

@Austronesier: So, in short you are saying that the approximation for the Indic alphabet द can be "ado (dental) as well as "other (with th-stopping)" right (the IPA and audio for "width" can be read and heard here and I don't agree that it is a good approximation for the Indic alphabet द)?-1Firang (talk) 18:28, 14 June 2023 (UTC)

Please get it straight. We don't approximate alphabet letters, we approximate sounds. That said, I have discussed three options. They are in plain sight. My personal favorite options are "ado" (it's a voiced plosive in every accent, and the place of articulation is very close; the optional "but dental" heals the issue) and "width" (almost always pronounced as a dental plosive, but sometimes sjubect to devoicing). "Other (with th-stopping)" has two problems: a) we have to resort to regionally confined accents, and b) not every TH-stopping accent produces a dental. With many TH-stoppers, "other" and "udder" are pronounced identically with a voiced alveolar plosive. Not much gained then over "ado", innit? –Austronesier (talk) 18:59, 14 June 2023 (UTC)

In that case, you should replace the approximation "width" with "ado (dental)" which is what we had in this Sanskrit IPA help page originally for the Indic alphabet द.-1Firang (talk) 19:13, 14 June 2023 (UTC)

No. Originally, we had this. Let's restore it, and consider the case closed? –Austronesier (talk) 19:25, 14 June 2023 (UTC)

@Austronesier: My first choice is to have "ado (dental)" and second choice is to have "do (dental)" as the "d" in the latter is plosive in some dialects. I will wait for you to implement this change in this Sanskrit IPA help page (please also make time and respond to my request at Help_talk:IPA/Hindi_and_Urdu#Reversions).-1Firang (talk) 02:46, 15 June 2023 (UTC)

Name one dialect in which the "d" of either 'do' or 'ado' is not plosive.

Name one dialect in which the "d" of 'do' is noticeably different from the "d" of 'ado'.

— kwami (talk) 03:04, 15 June 2023 (UTC)

Please compare https://www.collinsdictionary.com/dictionary/english/do and https://www.merriam-webster.com/dictionary/do - the "d" in the latter is stressed (but the "d" of 'ado' is the same in both).-1Firang (talk) 03:36, 15 June 2023 (UTC)

They're both the same in both. Some British dictionaries don't mark stress in monosyllables is all. It doesn't mean the pronunciation is any different. (Similarly, some Americanm dictionaries mark unstressed syllables as having secondary stress if the vowel isn't reduced. Again, it's just a convention, doesn't mean anything.) — kwami (talk) 03:50, 15 June 2023 (UTC)

@Austronesier: My first choice is to have "ado (dental)" and second choice is to have "do (dental)" as an approximation for द. I am waiting for you to implement this change in this Sanskrit IPA help page.-1Firang (talk) 12:51, 15 June 2023 (UTC)

I have made some changes to this IPA Help page and also a few similar changes for Help:IPA/Hindi and Urdu, Help:IPA/Nepali & Help:IPA/Marathi. Because it is rather pointless to have two approximations for e.g. [t] or have brackets for approximations like American bird, no need for duplicate links like the repetition of American/American English (see MOS:DUPLINK) & Scottish Gaelic is a separate language rather than an English variety (I have also added Scottish English in for approximations like may for [eː] in places which didn't have them because in most other English varieties, it is the diphthong [eɪ]). Broman178 (talk) 10:05, 16 June 2023 (UTC)

References

Classical Sanskrit "allophones"

I won't try to edit because there are too many of them but why are those alternative pronunciations for the diphthongs inserted here and even described as being Classical Skt. when they are distinctly NIA (raising the first element of ai to /ɛ~e/, /ɔ~o/ for au, which can only be explained as monophthongization. Interestingly, that's exactly what happens in NIA) and what's more, not a single source is given. BidensHyperborea (talk) 11:54, 19 September 2023 (UTC)

The Classical Sanskrit transcriptions are taken from Sanskrit#Pronunciation, where sources are given. The first elements are described by Cardona (2003) as maximally one-moraic. Further, he cites the Taittirīya-prātiśākhya which reports a somewhat closer pronunciation of the first segment of the diphthongs when compared to short a. But Cardona does not give an IPA transcription for this; I suspect that the transcription is OR. –Austronesier (talk) 18:35, 19 September 2023 (UTC)

Looks like it. ˀanāku 00:47, 27 September 2023 (UTC)

English examples

Why is the example for the dental plosive /t/ "eighth," with just the t highlighted? Most dialects pronounce "eighth" as [eɪθ]. Wouldn't a better example be something like "stop"? Some dialects pronounce /th/ as an allophone of t, but in those cases it's always aspirated as a [tʰ]. ForestAngel (talk) 04:22, 17 December 2023 (UTC)

@ForestAngel An idea is to use 'outthink' as the Hindi/Urdu page does. Although in normal speach the [t] might disappear there too. Exarchus (talk) 10:07, 1 January 2024 (UTC)

You're right, in most dialects people would swallow the t in "outthink" with a glottal stop. Now that I say it out loud, the /t/ in "eighth" is also a glottal stop as well and most dialects would pronounce it [eɪʔθ]. Either way, not a good example and should be addressed by an Indologist of some sort. ForestAngel (talk) 20:41, 29 March 2024 (UTC)

Stress pattern?

The article now says that Classical Sanskrit had a predictable stess pattern. But in the above topic 'Many mistakes' someone says that this is just something invented by western scholars as a matter of convention. I am actually very inclined to believe this, unless someone can give a good source for this stress pattern. Exarchus (talk) 10:21, 1 January 2024 (UTC)