Talk:Hawaiian phonology

Starting the "Hawaiian phonology" article
Since the Wikipedia "Hawaiian language" article has become longer than desired for an average Wikipedia article, I'm starting up this "Hawaiian phonology" article. The "Phonology" section of the "Hawaiian language" article is currently big enough to be its own article. So to reduce the size of the "Hawaiian language" article, it's appropriate to export its "Phonology" section to become this "Hawaiian phonology" article. Of course, we don't want (1) the same information to be needlessly duplicated in two different articles, nor (2) a complete lack of information on Hawaiian phonology in the "Hawaiian language" article. Therefore, some time will be needed to make appropriate adjustments to the "Phonology" section of the "Hawaiian language" article. Until/While that is done, please have patience with the temporary duplication of material.

Agent X 16:56, 27 July 2006 (UTC)

Ideas for improving the article
Great article. I took the liberty of replacing ˡ (U+02E1 MODIFIER LETTER SMALL L) in the article with ˈ (U+02C8 MODIFIER LETTER VERTICAL LINE) where it is used to indicate a stressed syllable. The former character is incorrect since although it appears as a vertical line in sans-serif fonts, it actually appears like a superscript lowercase l or a superscript 1 in serif fonts like Charis SIL and Doulos SIL. I only noticed this because I have a CSS hack that applies to all uses of the IPA template and forces a different font for them. I’m taking phonology at U. Hawaii at Mānoa so I’ll be sure to add anything interesting about Hawaiian that comes up. — Jéioosh 02:45, 23 August 2006 (UTC)


 * Oh crap! Didn't even catch that. I hope I haven't done it elsewhere. AEuSoes1 03:23, 23 August 2006 (UTC)


 * Well, like I said, I only noticed it because I have a CSS hack in place. I only wish that the rest of the world had the SIL IPA fonts installed... — Jéioosh 04:14, 23 August 2006 (UTC)


 * Again, glad you like the article. The replacement symbol you used looked alright to me, but so did the original symbol. It's a stretch to claim "[t]he former character is incorrect" on the basis of its appearance in fonts that 99.99% of users do not have. Most people have never heard of the Summer Institute for Linguistics. The character that I used to mark syllable stress is the one that Wikipedia presents for use in the IPA list of symbols below the edit box. So if you feel you have a "better" symbol, then you should contact Wikipedians who are in a position to consider implementing your suggestion(s). You might try Brion Vibber. If you learn something new about Hawaiian phonology at UH, do let us know. But remember, Wikipedia does not want anything that is not already published. So if it's "new", it might be unpublished, and therefore undesirable under Wikipedia policy. Agent X 15:15, 26 August 2006 (UTC)


 * The font used to display the characters is unimportant. Their character encodings are specific. On the basis of the two Unicode names, I argue that U+02E1 MODIFIER LETTER SMALL L is inappropriate as the stress marker in IPA. Checking the appropriate Unicode chart on their website I find that U+02C8 MODIFIER LETTER VERTICAL LINE is listed under the section “Miscellaneous phonetic modifiers” with the comment “primary stress, downstep; precedes letter or syllable modified”. On the other hand, U+02E1 MODIFIER LETTER SMALL L is given in the section “Additions based on 1989 IPA” just below U+02E0 MODIFIER LETTER SMALL GAMMA which has the comment “these modifier letters are occasionally used in transcription of affricates”. So it’s fairly clear that the Unicode standard makes a distinction between their intended purposes. I note that in the bottom of the edit box the two are given in sequence, with U+02C8 immediately following U+02E1. It’s easily possible to press the first of the two accidentally. I will bring this up at m:Help talk:Special characters and ask someone to reorder them for clarity. — Jéioosh 00:23, 27 August 2006 (UTC)


 * Your explanation and argument sounds good. I don't know the layout of the "insert box" characters on other users' screens. When I edit, the IPA characters are displayed in two lines. The first line ends with tiny lowercase superscripts: h, w, a j-like character, gamma, "lefty" glottal, n. The second line begins with the character I used for stress mark, followed by other stress marks, length mark, something too obscure to make out, a mid-height black square (apparently used to mark the end of a series of related symbols), and then the line ends with the double-curly-bracketed IPA. Based on your explanation, it appears that the first line of IPA characters is broken in a way that misled me into thinking that the last member of the series of superscripts is the first member of a series of stress marks. Bad line break. And there is no mid-height black square, nor significant white space, separating the superscript series from the stress mark series. Not seeing that previously, I used the first mark because it looks the same as the second, but darker. Of course, it also happened because a tiny superscripted sans-serif L looks nearly identical to a tiny superscripted vertical line.


 * If you have the motivation, maybe you can suggest more than just reordering in your communication with the Special Characters people. The addition of a separation character between the series of superscripts and the series of stress marks will help. Here's my best suggestion to offer on this matter --- persuade the Special Characters folks to implement appropriate new text for the hover boxes which appear when the mouse pointer "touches" the IPA characters. As of today (August 27th 2006), the hover text "Pronunciation in IPA" is the only text that appears, no matter which character is pointed to. It will be waaaay better to have hover text which provides the correct identity of each individual character. By "correct identity", I do NOT mean unicodese, but rather, the name of the IPA character. For a character with no "name", an identifying phrase is fine, such as "superscript L", "primary stress", etc. If you like this idea, please be my guest and try to get it implemented. Maybe the other user above, AEuSoes1, can join you in promoting this. He seems to love IPA-related stuff. Agent X 12:47, 27 August 2006 (UTC)


 * That is a good idea, but I’m afraid there might be technical problems with it. I know that the current hover text (link title) is from the IPA template. It’s possible that each link could have its own style overriding the link title, though. I’ll bring this up with the appropriate people. Thanks for the great suggestion. — Jéioosh 20:20, 27 August 2006 (UTC)

I have done some phonological analysis of the stress system of Hawaiian (in the context of a paper on Hawaii Creole English stress patterns); however, I am unsure how appropriate it would be but it would assuredly be better than saying the stress system is unpredictable in phrases or words with more than three morae, since it's a quantity-sensitive system. I would be more than happy to share the paper. Naea &#124; Hale o Keawe 10:49, 28 March 2019 (UTC) — Preceding unsigned comment added by Naea-a-liloa (talk • contribs)

Edit 1

 * I am very impressed with the work that you've done. I must say, though, that a lot of it goes over theoretical ground already covered in other articles.  Informing a reader of things like allophones, prosody, and synchronic linguistics can be accomplished with a simple interwiki link.  I've trimmed the article down considerably.  I don't believe I've taken anything important but if you feel that something I've taken out should be in the article don't be afraid to put it back in.  AEuSoes1 08:39, 22 August 2006 (UTC)


 * Glad you like the article. In accordance with Wikipedia policy on article format, I wrote the lead section as a summary of the article. So that serves as a "trimmed down" version. I also used the summary as the basic replacement for the Phonology section in the Hawaiian Language article, as indicated above. That was finished on August 1st 2006. So this is the place for the full-blown version, rather than a minimal version.


 * One of the Wikipedia Guides is to "make technical articles accessible". I emphasized that on the talk page for the Hawaiian Language article. (I suggest you read that talk page in its entirety.) That's why I took the trouble to "explain jargon" in this Hawaiian Phonology article. It's Wikipedia policy. I'm well aware of links, as can be seen by the many links which I wrote into this article and the Hawaiian Language article. (However, most linguistics articles in Wikipedia lack sources, so they are actually not "legitimate" in that respect. By contrast, the Hawaiian Language and Hawaiian Phonology articles are fairly well supported, as a result of my efforts.) Several of the links in the Hawaiian Language article, which I carefully matched to headings in this article, are no longer working properly as a result of the editing that you did. But don't worry, my last version will be restored.


 * Just to let you know, one should be very careful when making major edits to the work of other writers. This is especially important when you are not a specialist in the topic area of the article. Your edits resulted in the article making some incorrect statements, especially regarding Hawaiian word stress. Try discussing in advance your ideas for improvement, here on the talk page. I'll put in a heading, below, for that discussion. Agent X 14:49, 26 August 2006 (UTC)

Edit 2

 * I'm reverting your revert. If you check other phonology pages on Wikipedia you'll see that they are accessible in that they are not bloated and so detailed as to be overwhelming, even to those who know little about phonology.  A few reminders here and there about what symbols represent or what concepts mean is one thing, but going into detail about phonological theory and terminology so that it constitutes such a sizeable chunk of the article is off-topic to a certain degree, just as going into detail about election theory is off-topic in an article about George Bush. In addition, I find the tone to be condescending to the reader's intelligence in its simplicity.
 * Other than the deleting of material I find tangential, my other edits are good quality edits that do not need to be reverted. I improved the way in which the article indicated the phonemes that have no allophones, changed the tense in the references to sources to past participle, moved the glottal stop section into the consonants section (since it is a consonant), cleaned up the way in which allophones are indicated, and put the IPA notice at the top.  I also changed the representation of D (as in (C)D) to V, since diphthongs are vowels.  If you have a beef with this last item, we can have a discussion on that too.
 * I wouldn't be at all surprised if there were errors in my edits, but those are easily fixed without reverting. Do not revert my edits because a revert means you blatantly disagree with all of my edits which I think would be unreasonable. AEuSoes1 00:01, 1 September 2006 (UTC)

To AEuSoes1 --- Your user page indicates that you are a native speaker of English, and a 24-year-old in California who has a BA in English from Fresno State. It is very presumptuous for any person to deem himself qualified to rewrite an article on the phonology of a language which he does not speak. I am a fluent speaker of Hawaiian, and a long-time resident of Hawaii with a PhD in Linguistics. I reverted your edits because I am in a position to know that they were not good, in spite of your high opinion of your own edits. Yours were very major edits, not minor ones. Fixing the damage you did was best accomplished by simply undoing it in one step. Your threat, "Do not revert my edits", is contrary to Wikipedia policies. It is not your place to proclaim that I cannot disagree with your edits. I encouraged you to discuss your ideas for improvement, in advance, in the talk section created for that purpose. When you are specifically so encouraged, you should pay attention. I am the originator of, and major contributor to, the Hawaiian Phonology article. That is a fact which you need to heed, because when certain disputes occur, Wikipedia defers to an article's originator or major contributor. The "IPA notice" you added did not damage the article, but the rest of your edits did.

Agent X 04:54, 1 September 2006 (UTC)

Edit 3

 * I really don't see how your knowledge of Hawaiian is relevant to my edits. As far as I can tell, I didn't omit any important information and if you see anything, you're welcome to put it back in after my edits.  My edits were all about format and not substance so my knowledge of the language (which you correctly point out is very little) is not relevant and my knowledge of phonology (which I know quite a bit about, although I've got no degree) is good enough that you needn't brandish your PhD when what we're talking about is page layouts.  I'm aware of Wikipedia policies and my point in telling you not to blindly revert is that just because you created the page doesn't mean that this is your territory and it certainly doesn't mean that you can blindly revert quality edits with edits you disagree with.  Don't forget WP:be bold is a policy as well
 * I tell you what I'll do, I'll put back the edits that I feel are uncontroversial. If you'd like to gain consensus on the issue you can do so in the appropriate ways.  I'll be relying on precedent in pages like Russian phonology, Arabic phonology, Romanian phonology, Polish phonology, Danish phonology, and English phonology.  I recommend that you take a look at them since their present layouts are typical of how phonology pages are constructed at Wikipedia.  AEuSoes1 06:57, 1 September 2006 (UTC)

Just to be clear, here are the edits that I've just done. I'll be explicit in what I’ve done since the edits can be overwhelming. These are changes that don't require me to know the language or even to look at any of your sources. I've taken the information you put in and removed a lot of the redundancy. Now just check here to see the paragraphs that I would like taken out. Check the other phonology pages and you won't find any lengthy sections on "terminology and conventions"AEuSoes1 07:10, 1 September 2006 (UTC)
 * Replaced ε with . It’s a common mistake at wikipedia and it gets on my nerves.
 * Changed the introductory paragraph to be more direct in what this article is about right off the bat.
 * lumped together D (for diphthong) with V (or VV) since diphthongs are vowels (I also realize that my original edits made slight errors in the prosody section, I have fixed these errors)
 * shortened all the titles, especially since they were only lengthy due to the addition of synonyms and including the adjectival "Hawaiian" is rather redundant.
 * reworded the comments on allophonic variation for the consonants and vowels
 * moved the glottal stop section up to be a subsection of the consonants section
 * merged the "why does Hawaiian have diphthongs and long vowels" section up to the vowels section since it was only two paragraphs and could fit easily into the section
 * changed past tense to past participle such as "X has conjectured" rather than "X conjectured." I hope I don't need to justify such a change.

To Aeusoes1 --- Rewriting content, with no knowledge of the particular topic, and without looking at any sources, as you admit to doing, is the ANTITHESIS of what Wikipedia contributors should be doing. The majority of the article's content is careful paraphrasing of information from the sources. That is supplemented by quotations. Paraphrases are essentially careful re-wordings of quotes. When you, with no knowledge of the source texts, significantly change the paraphrases, you might as well be changing the direct quotes.

As for WP:BOLD, it states:
 * "... but don't be reckless!"
 * "If you expect or see a disagreement with your version of the article, and you want to change or delete anything substantial in the text, it's a good idea to list your objections one by one in the talk page, reasonably quoting the disputed phrases, explaining your reasoning and providing solid references." [italic added]
 * "Then, wait for responses for at least a day: people edit Wikipedia in their spare time and may not respond immediately. If no one objects, proceed, but always move large deletions to the Talk page and list your objections to the text so that other people will understand your changes and will be able to follow the history of the page." [italic added]
 * "... avoid deleting information wherever possible" [italic added]
 * "Be bold in contributions, but not in destructions. Editing is a collaborative effort, so editing boldly should not be confused with reverting boldly." [italic added]

You wrote: "I really don't see how your knowledge of Hawaiian is relevant to my edits". Because I have near-native fluency in Hawaiian, as well as a PhD in Linguistics, I know that your uninformed attempts to rewrite the description of word stress, throwing out "D" as a symbol for diphthong, are doomed to failure. For example, you made it have "3. (C)VV" as a stress unit. If V expands as a diphthong, then "wiuiu" can result --- an impossible stress unit in Hawaiian. That's only ONE of MANY errors your edits cause.

More importantly, your rewrite is not supported by any sources. You failed to cite any references, and admitted that you don't care to look at any.

You wrote: "My edits were all about format and not substance". That is false. See the preceding two paragraphs. Your rewrite of the word-stress content is substance, not format.

You wrote the phrase "blindly revert" to describe my restoration of the original version. My revert of your edits is informed, not "blind". I've seen an edit summary you wrote in the history of another phonology page where you made the same accusation against another user who undid your edits.

You point to the "layout" of other phonology pages, calling it "precedent". However, the Russian (RUS), Arabic (ARA), Romanian (ROM), Polish (POL), and Danish (DAN) phonology articles are not in compliance with Wikipedia guides. POL and DAN have no lead section. The leads of RUS, ARA, and ROM, violate WP:LEAD. None of those articles use boldface properly. ROM and POL have no references at all. RUS has only 2 old refs, ARA only 1 old ref. DAN has 8. None of those articles have even ONE CITATION to a reference. The "layout" of those pages is not consistent.

In contrast, the Hawaiian phonology article (HAW), as originated by me, has a proper lead section, proper use of boldface, 8 references, and 70 citations to those references. HAW is one of the best-supported linguistics articles currently on Wikipedia.

I will undo your edits by restoring my last version of the article. Because you have persisted in "forcing" your unsupported and objected-to edits into the article, I will ask a Wikipedia administrator to take a look at the situation. I will SUPPORT good changes, such as Jeioosh's idea about the stress-mark character. But I will OPPOSE bad changes, such as your uninformed edits. Agent X 16:28, 2 September 2006 (UTC)


 * Let me get this straight. Because I made a slight error with diphthongs, you're reverting everything else I've done?  a blind revert is when you "throw the baby out with the bathwater" so to speak.  That is, you revert an edit because you dislike one small part of it but also remove other parts of it that are either good or that you didn't even consider.  Here are some examples from your reversion.

Allophone of. Always. Emphasis on.
 * 1) .  This is the correct form for the IPA open-mid vowels.  Your revert has turned it to.
 * 2) Your edits have put for the long diphthongs:

Allophone of. Always. Emphasis on.

Allophone of. Always. Emphasis on.

Allophone of. Always. Emphasis on.

Allophone of. Always. Emphasis on.

Allophone of. Always. Emphasis on.

Note 1. There is ordinarily no variation in the sounds of the long diphthongs.

I have shortened this to

''All long diphthongs are falling. There is typically no allophonic variation for long diphthongs.''


 * There is absolutely no information loss in this edit. I've taken the information that you put and made it more accessible by reducing a redundancy.  All my most recent edits are an attempt to do this with the entire article and the only sources I need to reorganize the page layout is Wikipedia policy pages.  My error in the "re-write of the word-stress content" is just that, an error.  You, with your knowledge of Hawaiian and linguistics simply need to tweak my edit to make it accurate, which I think you're able to do.
 * Since we're citing Wikipedia policy, I'll cite WP:REVERT:
 * What's important is to let people know why you reverted. This helps the reverted person because they can remake their edit, but fixing whatever problem it is that you've identified
 * Your opposition to the rest of my edits is unclarified. You've listed objection to one of 8 items that I've listed.  What is your objection to the other 7?  I'd like to assume that you don't have any objection, so I can put them back in, right?
 * As for the other phonology pages, my point was that none of them have a large section dealing with terminology and conventions. None of them do.  I've never seen any Wikipedia language page that does so.  The citations are irrelevant to this discussion since this is about format.
 * But anyway, I'll wait for other people to come and deliver some sort of third opinion. I'll leave the article as is (as I recommend you do as well) until somebody says something.  AEuSoes1 23:04, 2 September 2006 (UTC)
 * Well, nobody seems to want to touch this page. I've restored my edits and tweaked the stress section so that (I hope) it's now correct.  AEuSoes1 23:34, 16 September 2006 (UTC)


 * Looks like most of AEuSoes1's edit is valid, the only things I see wrong is the mistakes that AEuSoes has already acknoledged and corrected. Remember nobody owns any page on wikipedia. Of oourse I would suggest that AEuSoes1's edits be improved, but not by reverts... but by going through and changing the bad parts. —— Eagle (ask me for help) 21:42, 23 September 2006 (UTC)

Phoneme inventory organization
I think a phonemical, not phonetical phoneme table with suitable underspecification would benefit this article. Eg. explicitely leaving an empty alveolar stop slot, as done currently, doesn't really capture the phonological structure. I suggest something along these lines:

Any objections? I'm a bit torn if there are any OR concerns here (tho I suppose a sorce for this kind of an analysis might be possible to locate, anyway). The POA fluctuation is accounted alreddy by the prose, the part I'm more iffy about is the continuant categorization. Having /h/ as fricativ, /l/ under approximant, and /w/ under both might be less controversial (possibly allowing for extending /k/ to fric, too). At the least I oppose classifying /l/ as "lateral" if it can be realized as non-lateral.

Also, neither of these captures the flap allophone of /l/, but the current table has that problem too. --Tropylium (talk) 14:50, 26 June 2008 (UTC)
 * OK, now I remember where I found this:
 * itself citing the Elbert & Pukui grammar much of our article cites too. I notice the glottals are considered separately from the other consonants here. Hmmm. Is this part Clements or E & P's interpretation? --Tropylium (talk) 09:10, 27 June 2008 (UTC)
 * Hmmm, I don't really see an advantage to underspecifying the consonants in the table. — Æµ§œš¹  [aɪm ˈfɻɛ̃ⁿdˡi] 14:17, 27 June 2008 (UTC)
 * I beg to differ. Actually this table makes a lot of sense and explains the range of consonantal variation far better. I'd even go further and use "Coronal/Dorsal" (there does not seem to be a catch-all term including both) instead of the description "Other", "Approximants" instead of "Continuants/Other" (come on, people, every phonetician knows that [h] isn't really a fricative at all, if anything, it's rather an approximant), "Nasal stops" instead of "Continuants/Nasal" and "Obstruents" instead of "Plosives" to capture the range of allophones involved. This would simplify the table even further and do the phonological system of Hawaiian more justice. And yes, it has all sorts of advantages; just check the tables at Rotokas language for a similar case with a comparable solution: strongly simplified charts.
 * But can anybody explain this part to me?
 * Roberts documented a sound between 'th' [θ] and 'k' in free variation with 'k' among elders from O'ahu and Kau'ai.
 * How can a phone be "between 'th' [θ] and 'k'"? --Florian Blaschke (talk) 21:44, 19 May 2011 (UTC)
 * I'm sorry, but not only does the above table look lazy, but it's so divergent from normal consonant tables that I'd hesitate to accept it without seeing sourcing that uses it. The Rotokas consonant chart is simple, but it doesn't actually violate the basics of consonant charts as there is still a clearly defined place of articulation labeled for each column.  Is there really something wrong with the table as it is? — Æµ§œš¹  [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi]  23:07, 19 May 2011 (UTC)
 * The point is that in Hawaiian, the place of articulation is, at least for the /t ~ k/ phoneme, simply not "clearly defined". Nor is even its manner of articulation necessarily plosive. Look at the list of possible allophones. It's insane. This is a special case. The table isn't lazy, just adequate. --Florian Blaschke (talk) 00:03, 20 May 2011 (UTC)
 * Oh, and the point about the Rotokas tables is that while the places of articulation are indeed defined, the manner of articulation is undefined (apart from voice, and nasality in the Aita dialect). This is a clear anomaly, as well. --Florian Blaschke (talk) 00:06, 20 May 2011 (UTC)
 * I don't think those are actually produced allophones. Those are sounds that are heard as the  phoneme because of diaphonic identification.  — Æµ§œš¹  [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi]  00:12, 20 May 2011 (UTC)

?
How can I get /aː/ in Monophthongs to be center-aligned? --99.190.210.221 (talk) 18:54, 2 August 2014 (UTC)

Rare combinations?
The article quotes Pukui and Ebert: "Elbert & Pukui[36] have pointed out that "Certain combinations of sounds are absent or rare." For example, no content word has the form CVVʔV". In reading the [|grammar] I don't see that reference. Besides, it seems wrong to me - what about the common words puaʻa and loaʻa? Fool4jesus (talk) 14:45, 8 October 2020 (UTC)
 * The source acutally talks about "CV̄V" being absent and "CV̄CV" being rare. Since V̄ is treated as /VV/ here, the most important point was lost, viz. that the constraint only refers to /CV$i$V$i$ʔV/. I'll fix this. –Austronesier (talk) 19:22, 8 October 2020 (UTC)

Added citation
I've inserted a citation of Parker Jones (2006), who shows how it is possible to predict stress more accurately in words of 5 syllables, but I'm not sufficiently proficient at Wikipedia editing to include the actual reference, which is to this conference paper: http://www.alta.asn.au/events/altw2006/proceedings/parker_jones.pdf So if any editors of this page can help, that would be super. — Preceding unsigned comment added by 188.29.134.150 (talk) 12:11, 5 February 2021 (UTC)
 * I added the reference Erinius (talk) 23:58, 31 July 2021 (UTC)

(Non-)nasalization of vowels
After reading Donegan's paper, I wonder about two things: Since Donegan (1995) is a rather peripheral source, I'd prefer not to make use of it, and mention nasalization solely based on Parker Jones (2018). –Austronesier (talk) 15:15, 6 March 2021 (UTC)
 * 1) The paper has no explicit source about the Hawaiian data, so it must be her own research. However, she does not explicitly mention how she got the data (elicitation, analysis of recordings etc.).
 * 2) The statement is about vowels preceding nasal consonants. Following nasal consonants, Parker Jones (2018) consistently notates nasalized vowels in the narrow transcription (p.112, e.g. narrow  for broad ).
 * Ah, thanks for noticing this. I found the claim in Zsiga (2020: 120, 125) (In Hawaiʻian, denasalization applies and nasalization is inhibited, so all vowels are oral in all contexts) cited to Donegan & Stampe (2009), but I found the description in Donegan (1995) more succinct so I cited it. If I understand correctly, Donegan & Stampe do seem to be saying vowels are always oral in Hawaiian (even after nasals), but as you pointed out it's a phonology, not phonetics, paper and they don't cite the data source or method, and Parker Jones (2018) provides counterevidence. I agree the claim is dubious, but I'd be reluctant to rephrase it to be in line with Parker Jones given Donegan & Stampe and Parker Jones seem to be at odds at each other; I wonder if it's due to a dialectal or diachronic variation. Nardog (talk) 18:20, 6 March 2021 (UTC)
 * Donegan & Stampe (2009) don't specify the direction of the processes (Denas) and (Nas), but in all examples these seem to work to the left. I have read the same into Donegan (1995), so I didn't see her data as contradictory to Parker Jones (2018) as long as we assume that Donegan only discusses regressive processes. Maybe we can just say "vowels have non-contrastive nasalization following nasal sonorants", citing Parker Jones (2018:111).
 * Generally, in Austronesian languages, progressive nasalization is more commonly found than regressive nasalization (or at least described, which explains my reading "bias" above), especially in languages with a CV-syllable structure. Probably related to this is that diachronic loss of the coda with feature transfer from the coda to the peak is very rare, so we rarely observe tonogenesis or phonemic nasalization of vowels. –Austronesier (talk) 19:20, 6 March 2021 (UTC)
 * AFAICT D&S are saying Hawaiian lacks nasal vowels in the entire phonetic inventory, giving the transcription in Table 2. Zsiga certainly seems to have interpreted them that way, or otherwise she wouldn't have said "in all contexts". Parker Jones is indeed more reliable as a descriptive source, but if we're going to mention the progressive assimilation I'd prefer it be explicitly attributed to him. Summarizing what we've learned so far, we could say something like "Vowels in Hawaiian have been described as invariably oral, even when adjacent to nasal consonants (Donegan & Stampe 2009:15–6), while Parker Jones (2018), describing a native speaker who has non-native-speaking parents and acquired the language in the revitalization movement, found that they are nasalized after nasal consonants (Parker Jones 2018:103, 111)." Nardog (talk) 20:07, 6 March 2021 (UTC)
 * Yes, sounds good. –Austronesier (talk) 21:14, 6 March 2021 (UTC)
 * Thanks! I love it when a discussion results in an improvement of the encyclopedia and our collective knowledge. Nardog (talk) 22:01, 6 March 2021 (UTC)

Missing sources
Anyone know what Elbert & Pukui 1986 and Roberts 1967 are supposed to be? Erinius (talk) 01:23, 4 August 2021 (UTC)
 * Issues with Elbert & Pukui 1986 are resolved. I'm still not sure what exactly Roberts 1967 is supposed to be (the reference was added in 2008 by an IP user), though I suspect it may be Ancient Hawaiian Music by Helen Heffron Roberts. Erinius (talk) 21:11, 14 August 2021 (UTC)
 * Roberts 1967 is in fact Ancient Hawaiian Music and the original text is still pretty vague, though Roberts does actually mention a sound "in between th and z", not in between th and k. I personally suspect a voiced or voiceless slit fricative, or maybe a denti-alveolar(?) but including that in the article with no other references would be OR on my part. Source is here. Erinius (talk) 21:33, 14 August 2021 (UTC)

Allophony of /k/
Here you have added [ç, ʝ, q, ɢ, χ, ʁ, ʡ, ħ, ʕ] to the list of "acceptable" realizations of /k/, but, with the possible exception of [q], they are hard to believe. If this is based on nothing better than a literal interpretation of "any obstruent that is neither labial nor glottal", I don't think it should be here. --Theurgist (talk) 17:59, 21 January 2023 (UTC)