Template talk:IPA pulmonic consonants

RfC: chart transcription and placement of creaky-voiced glottal approximant

1. Should the creaky-voiced glottal approximant (currently listed as ⟨ʔ̞⟩ ⟨˷⟩) be removed from this chart template?

a. If yes, should glottal approximants be greyed out in the chart?

b. If no, how should it be transcribed in the chart?

2. How should the consonant be transcribed in articles (outside of this template)? consensus for #2 from discussion below: ⟨˷⟩ (as in /ha˷oʔ/)

(This RfC pertains to the discussion above, #Template-protected edit request on 31 October 2025, which appears to have reached a stalemate.) ~ oklopfer (💬) 17:30, 13 November 2025 (UTC)

Discussion (creaky placement and transcription)

I argue that it should not be listed in the chart, and that glottal approximants should indeed be greyed out. ⟨ʔ̞⟩ is never used in any source literature at all, and both the official IPA chart and John Esling's expanded chart doi:10.1002/9781444317251.ch18 (which this template is directly based on) mark glottal approximants as impossible articulations - this is therefore largely WP:OR, and is directly contradictory to the widely recognized sources. In my opinion, it does not matter if it theoretically 'fits' in the chart; readers look at this template as an authoritative source, so we should not be giving them information that goes against sourcing universally; no source has ever placed the consonant in the pulmonic chart, nor used the transcription ⟨ʔ̞⟩, except for Wikipedia. Its description is largely interpreted from a single paragraph of SOWL:

The so-called voiced glottal stop is signaled by slight irregularities in the glottal pulses, and a considerable decrease in the amplitude of the pulses which is probably due to an increase in the glottal stiffness; but there is no indication of anything that would normally be called a stop, glottal or otherwise. There is, however, a clear distinction between * and a simple transition between adjacent vowels as shown in the lower part of the figure. The vowel-to-vowel transition reflects simply the waveform changes that can be associated with the changes in the formant frequencies. In summary, it seems that * in Gimi is voiced, and involves some glottal activity; but it might better be called a creaky voiced glottal approximant rather than a stop.

— Ladefoged, Peter; Maddieson, Ian (1996). The Sounds of the World's Languages. Oxford: Blackwell. p. 76. ISBN 0-631-19815-6.

If it must remain in the template, then I insist that a sourced transcription is used, not a presumptive and made up one (by made up, I mean it does not exist in any literature). doi:10.1017/S0025100321000116 uses the transcription ⟨ʔ̬⟩; doi:10.1017/S0952675704000302 and both use ⟨ʔ̰⟩; and SOWL and doi:10.21437/Interspeech.2021-1417 both use ⟨*⟩ (⟨ʔ̬⟩ and ⟨ʔ̰⟩ are technically impossible by definitions of glottal stops and {creaky,} voicing). ~ oklopfer (💬) 17:30, 13 November 2025 (UTC)

I still argue what Kwami think about, he said ⟨ʔ̞⟩ but I feel if glottal stop lowered, it may be equal or adjacent to ⟨h⟩ instead of what known as creaky-voiced glottal approximant. señor verde⸻pregúnteme 10:56, 14 November 2025 (UTC)

Based on strict transcription definitions yes, it probably would, but not in the reality of laryngeal action. And it is the descriptions of the sound we are going off of, not the transcriptions of it. If h is placeless aspiration and ɦ is placeless breathy voice, * is placeless creaky voice. ~ oklopfer (💬) 13:42, 14 November 2025 (UTC)

Copying from above:

Heard back from Esling: some indication of [˷] would be necessary, but ⟨ʔ⟩ is not appropriate. In his conception there are laryngeal approximants, but not glottal ones, as [ʔ] is the relaxed state of the glottis.

Anyway, Ladefoged's standard IPA ⟨*⟩ would still seem to be appropriate. If not that, extIPA-style ⟨˷⟩ would seem to be the way to go. Phonetically we'd have [a̰] etc., depending on the surrounding vowels, but for phonemic notation we'd need to abstract away from the vowel, so our choices would seem to be /ha*oʔ/ or /ha˷oʔ/. — kwami (talk) 02:17, 16 November 2025 (UTC)

Is actually pre-creaky /o/, post-creaky /a/ or both? señor verde⸻pregúnteme 06:24, 16 November 2025 (UTC)

We could probably track down Ladefoged's recordings, but then we'd be substituting our own judgement. Evidently it's not easy to hear. It might be a creaky transition from [a] to [o], which would be hard to transcribe phonetically, which would be a reason to stick to a phonemic transcription. — kwami (talk) 08:12, 16 November 2025 (UTC)

Unless there are others that exist, Ladefoged's recordings can be heard here, and the word in question, /ha˷oʔ/, (I believe) can be found in gim_word-list_1976_01 @ 0:32-0:35 and maybe one of the examples @ 0:20-0:25. I agree, we should not substitute our own judgement, but we do at least have some tangible documentation. ~ oklopfer (💬) 13:31, 16 November 2025 (UTC)

Does /ha˷oʔ/ work for you then? Would the tilde in [ɦ̰] cancel the murmur and leave just creaky-voiced phonation?

Though we're not addressing whether in should be included on the navigation chart. Bare phonation doesn't really fit the layout, though of course that's precisely what [h] and [ɦ] are commonly analyzed to be. — kwami (talk) 08:26, 17 November 2025 (UTC)

Yes, /ha˷oʔ/ works for me. I think as you stated before, [ɦ̰] still doesn't really make sense, and kind of implies a neutralization.

If we are leaving it in the chart, is it remaining as [ʔ̞]? ~ oklopfer (💬) 13:20, 17 November 2025 (UTC)

We should be consistent: if it's there, it should be the same <˷>. I don't know about including bare phonations, though. Do we have any others that serve as consonants? — kwami (talk) 13:26, 17 November 2025 (UTC)

I can't find any other examples in Module:IPA symbol/data. IPAlink-ing ⟨̰⟩ takes you to creaky voice, and I am not sure we want to disrupt that navigation; the closest linking we have might be ⟨˞⟩ (r-colored diacritic), but that is obviously not listed in either the vowel or consonant chart, and is far more ambiguous. ~ oklopfer (💬) 13:43, 17 November 2025 (UTC)

My mistake, ⟨˷⟩ is not the same as ⟨̰⟩. Right now it does not link to anything, so it would not break symbol data:

⟨˷⟩ (I've added it now in Special:Diff/1322693287, please revert me if you disagree with the change) ~ oklopfer (💬) 16:54, 17 November 2025 (UTC)

@Fdom5997 are you okay with this transcription being used for Gimi language? ~ oklopfer (💬) 17:01, 17 November 2025 (UTC)

@oklopfer Actually, sure. Why not? It's a creaky-voice symbol, so for that, I'd suggest we use it instead of the dreaded asterisk! So yes! Fdom5997 (talk) 20:07, 17 November 2025 (UTC)

As shared by Esling, Whalen et al. (2023) is also related and quite interesting, and gives food for thought. It is evident that the status of these sounds is quite uncertain, and very much still being actively explored and expanded on in field work. ~ oklopfer (💬) 14:04, 16 November 2025 (UTC)

Yeah, I'd vote for /ha˷oʔ/. That's agnostic as to the analysis. — kwami (talk) 20:45, 16 November 2025 (UTC)

@Kwamikagami Same. We should use it! Fdom5997 (talk) 20:13, 17 November 2025 (UTC)

After reading through this discussion, I too support the use of /˷/ rather than an asterisk. The wildcard character should be avoided as much as possible, as it is pretty abstract in nature and primarily should serve as a placeholder when no other symbols are available. – PharyngealImplosive7 (talk) 03:38, 18 November 2025 (UTC)

Are we going to change this template to use ˷? - BᴏᴅʜıHᴀᴙᴩ (talk, contributions) 00:48, 18 November 2025 (UTC)

• Use ˷ for this sound. - BᴏᴅʜıHᴀᴙᴩ (talk, contributions) 05:22, 17 November 2025 (UTC)

Discussion for some changes

I am discussing which changes we should import from the sandbox. See #Template-protected edit request on 27 August 2025. @Oklopfer, PharyngealImplosive7, and Kwamikagami: pinging these users. - BᴏᴅʜıHᴀᴙᴩ (talk, contributions) 19:48, 17 December 2025 (UTC)

some of those don't link to an article or even a section about the sound, so they're just clutter — kwami (talk) 21:45, 17 December 2025 (UTC)

I've now fixed up the ones that had page sections. Going to list the consonants from the sandbox diff so we can discuss them all (see below) ~ oklopfer (💬) 15:56, 18 December 2025 (UTC)

Additions

These ones point to sections:

• Postalveolar plosives: ⟨t̠⟩ + ⟨d̠⟩
• Retroflex non-sibilant affricate: ⟨ɖɻ᷵⟩
• Voiceless alveolar, postalveolar, and retroflex approximants: ⟨ɹ̥⟩, ⟨ɹ̠̊⟩, ⟨ɻ̊⟩
• Voiceless palatal approximant: ⟨ȷ̊⟩
• Dental trills: ⟨r̪̊⟩ + ⟨r̪⟩
• Dental lateral fricative: ⟨ɮ̪⟩
• Postalveolar lateral flap: ⟨ɺ̠⟩

These ones do not:

• Voiceless labiodental approximant: ⟨ʋ̥⟩ (only one example on the page)
• Uvular and pharyngeal approximants: ⟨ʁ̞⟩ + ⟨ʕ̞⟩ (no section, but aligns with typical theory on gutturals, as well as ⟨ð̞⟩, which is currently in the chart and also has no section)

Visual Changes

These feel less important, since they are equivalents:

• ⟨tʂ⟩ > ⟨ʈʂ⟩
• ⟨dʐ⟩ > ⟨ɖʐ⟩
• ⟨d̠ɹ̠˔⟩ > ⟨d̠ɹ̠᷵⟩
• ⟨ɹ̠˔⟩ > ⟨ɹ̠᷵⟩
• ⟨ɻ˔⟩ > ⟨ɻ᷵⟩
• ⟨tꞎ⟩ > ⟨ʈꞎ⟩
• ⟨d𝼅⟩ > ⟨ɖ𝼅⟩
• ⟨𝼈̥⟩ > ⟨𝼈̊⟩

~ oklopfer (💬) 15:54, 18 December 2025 (UTC)

What is the reason we should change tʂ to ʈʂ? - BᴏᴅʜıHᴀᴙᴩ (talk, contributions) 17:10, 18 December 2025 (UTC)

And it would make sense for us to add ɺ̪ since it has a section. I have fixed the redirect target. - BᴏᴅʜıHᴀᴙᴩ (talk, contributions) 17:40, 18 December 2025 (UTC)

Personally I don't like the redundant retroflex hooks. They make the transcription more difficult to read. Recently when the IPA officially supported adding old-style affricate ligatures to Unicode, one concern expressed by linguists was whether the retroflex affricates would have a single hook; there was an assurance that they would. Since ligatures are single letters, a second hook would be completely redundant, but the same legibility and aesthetic concerns apply here. But not a big deal either way. — kwami (talk) 21:25, 18 December 2025 (UTC)

I think including ⟨ʁ̞⟩ ⟨ʕ̞⟩ ⟨ð̞⟩ is justified, given that it's not entirely obvious how that would fall out otherwise. — kwami (talk) 21:28, 18 December 2025 (UTC)

I oppose the uptack above, as that change is unnecessary. Plus, Windows 11 (what I use) does support the extIPA symbols added to Unicode 14.0 properly. Same with ChromeOS.

However, I do not know of any browsers (including Windows 11 or ChromeOS) that render the uptack above properly, as that was only added to Unicode 17.0. SVG-image-maker (talk) 18:18, 19 January 2026 (UTC)

@SVG-image-maker the uptack above was added almost a dozen years ago in Unicode 7.0, not 17.0. It is not a matter of browser rendering but rather of having an installed font which supports it. ~ oklopfer (💬) 18:49, 19 January 2026 (UTC)

I get that, but [d̠ɹ̠˔] and [d̠ɹ̠̝] mean the same thing and have much more browser support. And I do have Catrinity installed which has it, but 1DF5 is still a box in-browser. SVG-image-maker (talk) 19:05, 19 January 2026 (UTC)

Adding some clusters

I know there are clusters and not phonemes of their own but as I've seen some languages in their phonological charts add them for example rigwe, so here is my proposal:

IPA: Pulmonic consonants

Place →	Labial					Coronal										Dorsal					Laryngeal
Manner ↓	Bi­labial		Labio­dental		Linguo­labial		Dental		Alveolar		Post­alveolar		Retro­flex		(Alve­olo-)​palatal		Velar		Uvular		Pharyn­geal/epi­glottal		Glottal
Nasal	m̥	m	ɱ̊	ɱ		n̼	n̪̊	n̪	n̥	n	n̠̊	n̠	ɳ̊	ɳ	ɲ̊	ɲ	ŋ̊	ŋ	ɴ̥	ɴ
Plosive	p	b	p̪	b̪	t̼	d̼	t̪	d̪	t	d			ʈ	ɖ	c	ɟ	k	ɡ	q	ɢ	ʡ		ʔ
Sibilant affricate	(ps)	(bz)					t̪s̪	d̪z̪	ts	dz	t̠ʃ	d̠ʒ	tʂ	dʐ	tɕ	dʑ	(ks)	(gz)	(qs)	(Gz)
Non-sibilant affricate	pɸ	bβ	p̪f	b̪v			t̪θ	d̪ð	tɹ̝̊	dɹ̝	t̠ɹ̠̊˔	d̠ɹ̠˔			cç	ɟʝ	kx	ɡɣ	qχ	ɢʁ	ʡʜ	ʡʢ	ʔh
Sibilant fricative							s̪	z̪	s	z	ʃ	ʒ	ʂ	ʐ	ɕ	ʑ
Non-sibilant fricative	ɸ	β	f	v	θ̼	ð̼	θ	ð	θ̠	ð̠	ɹ̠̊˔	ɹ̠˔	ɻ̊˔	ɻ˔	ç	ʝ	x	ɣ	χ	ʁ	ħ	ʕ	h	ɦ
Approximant		β̞		ʋ				ð̞		ɹ		ɹ̠		ɻ		j		ɰ						˷
Tap/flap		ⱱ̟		ⱱ					ɾ̥	ɾ			ɽ̊	ɽ						ɢ̆		ʡ̮
Trill	ʙ̥	ʙ							r̥	r		r̠	ɽ̊r̥	ɽr					ʀ̥	ʀ	ʜ	ʢ
Lateral affricate									tɬ	dɮ			tꞎ	d𝼅	c𝼆	ɟʎ̝	k𝼄	ɡʟ̝
Lateral fricative							ɬ̪		ɬ	ɮ			ꞎ	𝼅	𝼆	ʎ̝	𝼄	ʟ̝
Lateral approximant								l̪	l̥	l		l̠	ɭ̊	ɭ	ʎ̥	ʎ	ʟ̥	ʟ		ʟ̠
Lateral tap/flap									ɺ̥	ɺ			𝼈̊	𝼈		ʎ̮		ʟ̆

• IPA help
• audio
• full chart
• table
• template

Symbols to the right in a cell are voiced, to the left are voiceless. Shaded areas denote articulations judged impossible. Ekpailou4 (talk) 18:35, 28 March 2026 (UTC)

This talk page is not used for it.
If you have any question about IPA but not to be used for citing Wikipedia itself, you can ask to either user like @BodhiHarp, @Kwamikagami, or @Oklopfer. señor verde⸻pregúnteme 19:51, 28 March 2026 (UTC)

It was a proposal for the page. I think that's a great use of the talk page. Ekpailou4 (talk) 08:14, 29 March 2026 (UTC)

Clusters aren't affricates. The Rigwe example is highly rare, affricates are usually only considered as such when they are at the same point of articulation. There are of course a few other rare outliers, like the alveolar-labial trilled affricates (opposite sequence order of Rigwe's labial-alveolar), but they don't belong here. ~ oklopfer (💬) 20:36, 28 March 2026 (UTC)

Proposed changes

@Kwamikagami @Oklopfer I suggest the following additions, these seem to have been uncontroversial above. These seem to be attested.

Voiceless postalveolar plosive

Voiced postalveolar plosive

Voiceless labiodental approximant

Voiceless alveolar approximant

Voiceless postalveolar approximant

Voiceless retroflex approximant

Voiceless palatal approximant

Voiceless dental trill

Voiced dental trill

Voiced dental lateral fricative

Voiced postalveolar lateral flap SVG-image-maker (talk) 00:51, 2 April 2026 (UTC)

FYI I had mentioned all of those in § Additions above, with the only ones you didn't mention being the voiced retroflex non-sibilant affricate and the ambiguous uvular and pharyngeal approximants. We never made any movement on adding those, probably because we all forgot about it. ~ oklopfer (💬) 01:20, 2 April 2026 (UTC)