Module talk:nan-pron

Quanzhou pronunciation for Min Nan

@Wyang At this revision Min Nan has also Quanzhou pronunciation. The automatically produced only has Taipei and Zhangzhou. --Anatoli ^{(обсудить}/^вклад) 07:05, 2 May 2014 (UTC)

Only Taipei and Zhangzhou pronunciations can be inferred from POJ. Quanzhou pronunciation has tone sandhi rules not inferrable from POJ orthography. I'm not sure. I prefer dropping it to make things simple. Wyang (talk) 07:09, 2 May 2014 (UTC)

I agree with you but I wonder if native speakers could add their local pronunciation? Well, it won't be in the same format and it's unsourced, anyway. OK, dropping. --Anatoli ^{(обсудить}/^вклад) 07:25, 2 May 2014 (UTC)

Variant pronunciations

@Kc_kennylau, Justinrleung Since the interest is here, I would like to point out one inaccuracy in the present module, which is the handling of variant pronunciations. A large part of variation in Min Nan readings is not in-dialect variation, but rather inter-dialectal variation. e.g. 日 (ji̍t/li̍t), 月 (go̍eh/ge̍h), which means the IPA pronunciations for dialects/patois will be incorrect in these cases (i.e. Zhangzhou uses ji̍t and Quanzhou uses li̍t for 日). I had the thought of using additional symbols to represent this, similar to what is done in 日期 (Mandarin), such as "ñit" for 日 which will be interpreted by the module as ji̍t (Zhangzhou) and li̍t (Quanzhou). Some explanation for the interdialectal variations can be found here. I haven't got around to doing this and this has been a problem that's long-standing now. It would be great if you guys are interested and we can try to fix it together. Wyang (talk) 23:47, 24 June 2015 (UTC)

I've noticed this problem for a while, so it would be great if we could fix it together. I think we should include Xiamen (Amoy), Zhangzhou, Quanzhou, Taipei and Kaohsiung as the main varieties for Quan-Zhang Min Nan. However, POJ does not have enough coverage of symbols, since it was made based on the Xiamen dialect. The 6th tone unique to Quanzhou can perhaps be represented by ˇ, as suggested by 泉州話. For the vowels POJ does not have, I think we should adapt some of the conventions used in Tâi-lô, which has more coverage in terms of phonemes. For example, 豬 in Quanzhou can be written as tir.

I think the proposal of using additional symbols to represent the inter-dialectal differences is a good idea. However, I think it might be a problem for characters that have completely different pronunciations between the dialects. For example, 奶 is pronounced as ni in Xiamen and Taipei, leng in Zhangzhou and Kaohsiung, nee (IPA^(key): /nɛ˦˦/) in Zhangzhou and lin in Quanzhou, A solution for this problem can be something like the Mandarin template, where a regional pronunciation can be marked with a note, as done in 老撾. Justinrleung (talk) 05:06, 25 June 2015 (UTC)

Justin, thanks for the reply! Would you be able to provide more details on the interdialectal variations? You seem very knowledgeable and resourceful. Ideally we probably would like to have compiled tables listing the dialectal correspondences for the five varieties proposed, in terms of initials, finals, tones and tone sandhi rules. As for the second part, I agree using characters may be the best approach. Wyang (talk) 12:08, 25 June 2015 (UTC)

I've compiled tables showing interdialectal variations in Min Nan. They're by no means exhaustive, especially for the finals, but I think there is sufficient information to work with.Justinrleung (talk) 05:14, 30 June 2015 (UTC)

Tone marks for POJ

@Wyang, Kc kennylau, which set of rules are we basing poj_format on? Are we using the rules given on the Wikipedia article for POJ? — justin(r)leung _{{ (t...) | c=› }} 06:14, 18 January 2016 (UTC)

@Justinrleung: I believe so. --kc_kennylau (talk) 11:45, 18 January 2016 (UTC)

@Kc kennylau: Then we might have to change the code a bit, since for example, the current code forces goán to be góan. — justin(r)leung _{{ (t...) | c=› }} 13:38, 18 January 2016 (UTC)

@Justinrleung: Better now. --kc_kennylau (talk) 14:41, 18 January 2016 (UTC)

華僑

@Wyang, Kc kennylau, Justinrleung, Some latest changes cause problems for華僑／华侨 (huáqiáo), POJ - "Hoâ-kiâu". --Anatoli T. ^{(обсудить}/^вклад) 04:16, 20 January 2016 (UTC)

This was intentional; see the edit summary of [1]. Originally Tailo-influenced technically-incorrect POJ (hoâ instead of hôa) was automatically and silently fixed by the module but not anymore. The error message is pretty self-explanatory. —suzukaze (t・c) 04:20, 20 January 2016 (UTC)

Thanks. I hope someone was going to address the pages with module errors. --Anatoli T. ^{(обсудить}/^вклад) 04:23, 20 January 2016 (UTC)

Teochew Pengim

Apparently there's a Pengim that has a vowel "eu" (possibly devised by gaginang.org) and a Pengim that has "e" (潮州話拼音方案)????possibly??

• [2] says that 汝 is le2,
  • with their chart saying that e=/ɯ/
  • and an audio recording that sounds like /lɯ/;
  • w:zh:潮州話拼音方案 also has e=/ɯ/;
• but there are websites that say that 汝 is leu2,
  • such as this one (which seems to be associated with gaginang.org)
  • which has this audio recording...

The module currently converts le2 to /lə⁵²/ (/ə/ does not appear on either [3] nor w:zh:潮州話拼音方案) and leu2 to /lɯ⁵²/. IDK. —suzukaze (t・c) 10:39, 17 February 2016 (UTC)

@suzukaze-c I think we should follow w:zh:潮州話拼音方案 (i.e. use ⟨e⟩). Mogher and 潮州音字典 both have ⟨ke3⟩ for 去 and ⟨le3⟩ for 你/汝. However, both these sites have ⟨e⟩ in the recording sounding more like /ə/ than /ɯ/. In fact, Mogher has the IPA /ə/ and some variant of POJ using ⟨eu⟩ corresponding to ⟨e⟩ in 潮州話拼音方案. — justin(r)leung _{{ (t...) | c=› }} 15:13, 18 February 2016 (UTC)

At any rate, ê does not coexist with e and eu :/ —suzukaze (t・c) 18:48, 19 February 2016 (UTC)

line 611

Could text = mw.ustring.toNFD(text) be substituted for text = gsub(mw.ustring.lower(text), ".", split_tone) and table split_tone? —suzukaze (t・c) 00:00, 21 June 2016 (UTC)

Thanks! Very clever. It works well. Wyang (talk) 02:28, 21 June 2016 (UTC)

Not sure if the edit here caused a bunch of errors. See CAT:E. — justin(r)leung _{{ (t...) | c=› }} 06:56, 21 June 2016 (UTC)

Previewing a random assortment of CAT:E pages using previous module revisions suggests that it is not at fault... —suzukaze (t・c) 07:00, 21 June 2016 (UTC)

Poking around using mw.log() suggests that errors are being raised because a tone_value['Quanzhou'][4] is not found. —suzukaze (t・c) 07:04, 21 June 2016 (UTC)

@Mar vin kaiser changed the tone sandhi of tone 2 in the Quanzhou dialect: diff. I'm not sure if this is correct - he may have wanted to change the sandhi value of tone 2 when it is followed by 仔. There is no tone 4, only 4A (-p/t/k) and 4B (-h), although they have the same tonal value in isolation. Wyang (talk) 07:56, 21 June 2016 (UTC)

He may have meant 5 -> 6 (which was already correct). Minnan Fangyan Da Cidian uses 2 instead of 5 and 4 instead of 6. — justin(r)leung _{{ (t...) | c=› }} 08:19, 21 June 2016 (UTC)

Yeah sorry, I was trying to fix the Quanzhou pronunciation of 姨仔, but it didn't work when I tried fixing it, as I don't understand the coding much. It's ok now though. Thanks to whoever fixed it. --Mar vin kaiser (talk) 10:35, 21 June 2016 (UTC)

Bug for 姨仔: fails PSDB, gives "iar", but needs apostrophe "i'ar" Hongthay (talk) 15:20, 21 June 2016 (UTC)

Ideas

First, the recent developments are great! I would like to put on the table: having regional info in a centralized table vs. at the article page...does this make more sense than the current way? The entire MoE regional tables (two big ones) are available. Also, instead of typing 仔, could an alternate code be used with ASCII entities or a modified á? Thanks! Hongthay (talk) 14:47, 21 June 2016 (UTC)

More of a question: Are literary readings more or less uniform across regions? Hongthay (talk) 00:18, 24 June 2016 (UTC)

I guess they are quite similar, but there are still differences. For example, 丈 is tiōng (Xiamen, Taiwan) / tiāng (Zhangzhou) / tiǒng (Quanzhou). — justin(r)leung _{{ (t...) | c=› }} 09:14, 27 June 2016 (UTC)

Phofsit Daibuun

@Hongthay, Kc kennylau Should Phofsit Daibuun only be used for Taiwanese (and not for other varieties, especially Quanzhou)? Also, should we show different tone sandhis for the fifth tone? — justin(r)leung _{{ (t...) | c=› }} 00:27, 30 October 2016 (UTC)

My experience is with MLT. The sandhi for POJ to MLT conversion is based on southern convention (in line w/ Zhangzhou, I think). I would add it might be nice to directly input the MLT in case the automatic conversion is not right...there are writing exceptions as well. Thank you Hongthay (talk) 04:01, 30 October 2016 (UTC)

@Hongthay I wouldn't think that it's beneficial to directly input the MLT. It's better if it's done automatically unless MLT/PFDB has many exceptions. Could you point out some exceptions? — justin(r)leung _{{ (t...) | c=› }} 05:47, 30 October 2016 (UTC)

@Justinrleung, I've noticed many cases where an apostrophe is left in though there is no technical ambiguity with it removed. Kex + oe (計劃) could be keaoe but it's kea'oe. Siong'iong, not siongiong (常用). Heng'ioong, not hengioong (形容). O'of, not oof. Another case is the hyphen is left in (though these are often sum of parts): juxnpviar-kauq (潤餅𩛩), olioong-tee (烏龍茶) Hongthay (talk) 06:12, 30 October 2016 (UTC)

@Hongthay There must be rules for apostrophes and hyphens, no? — justin(r)leung _{{ (t...) | c=› }} 06:16, 30 October 2016 (UTC)

@Justinrleung There may be rules that are followed in most cases but I think there is some discretion involving readability/aesthetics (human factor) as well. Hongthay (talk) 07:32, 30 October 2016 (UTC)

Splitting |mn-t= by location, similar to |mn=?

@Justinrleung, @Suzukaze-c e.g. 潮汕. Wyang (talk) 23:38, 20 January 2017 (UTC)

We definitely should. —suzukaze (t・c) 23:43, 20 January 2017 (UTC)

Great - we should try to create a page like User:Justinrleung/Interdialectal differences in Quan-Zhang Min Nan to describe the inter-dialectal differences in Teochew. It may be better to move the Teochew code here to a separate module. Wyang (talk) 23:47, 20 January 2017 (UTC)

@Justinrleung [in case you didn't get the ping] —suzukaze (t・c) 19:37, 2 April 2017 (UTC)

@Suzukaze-c, Wyang I got the ping before. I don't know why I didn't reply then. I found this website, which should be sufficient to create something like we did for Hokkien. — justin(r)leung _{{ (t...) | c=› }} 19:51, 2 April 2017 (UTC)

That's a great site! I'm a bit short of time these days - so please feel free to go ahead with the split if anyone is keen. Wyang (talk) 09:25, 3 April 2017 (UTC)

好惡

When creating 好惡 with {{zh-new}}, it reported an error in the Min Nan pronunciation: The Min Nan pronunciation defaultly provided was hò͘ⁿ-ok/hò͘ⁿ-ò͘ⁿ. --Dine2016 (talk) 02:08, 17 August 2017 (UTC)

@Dine2016: it's not hò͘ⁿ-ok, but hòⁿ-ok (without right dot on o). — justin(r)leung _{{ (t...) | c=› }} 02:10, 17 August 2017 (UTC)

Standard/prestige accent

@Justinrleung: Hi, which variety is most standard/most common or prestige, the one users are most likely to hear (if there is such an accent)? If it is, e.g. Taipei, can it be moved to to the top in the display, please, rather than showing them alphabetically? --Anatoli T. ^{(обсудить}/^вклад) 11:21, 21 May 2019 (UTC)

@Atitarev: It depends on who you ask. In Mainland China, the Xiamen (Amoy) dialect definitely has the prestige. In Taiwan, the Kaohsiung dialect seems to be the basis for General Taiwanese (通行腔/普通腔/優勢腔). Other dialects have some regional prestige, particularly the Taipei dialect in northern Taiwan and the Tainan dialect in southern Taiwan. We currently aren't showing them alphabetically, but essentially based on the order of the input. — justin(r)leung _{{ (t...) | c=› }} 19:41, 21 May 2019 (UTC)

IPA

I fixed the fake IPA tone numbers to actual IPA. It might be better to do the conversion at [local tone_superscript], though that would still require getting rid of the double digits further up (except for things like '224', where the double digit '22' is used to indicate contour rather than dab'ing tone #2).

For sandhi, rather than using the superscript minus (as a sub for the en dash), it would be more professional to use reversed Chao tone letters. But that would require more tinkering with the code. kwami (talk) 01:40, 21 August 2019 (UTC)

@Justinrleung: Hi Justin. Could you tell me what needs to be fixed for sandhi? We really shouldn't use regional ad hoc conventions and call them IPA. Once this is fixed up, the IPA template should no longer accept digits or other regional conventions that are unintelligible internationally, so all of these transcriptions will generate errors. kwami (talk) 02:52, 21 August 2019 (UTC)

@Kwamikagami: It should be fixed. The only problem is the double digits. — justin(r)leung _{{ (t...) | c=› }} 02:56, 21 August 2019 (UTC)

@Justinrleung: Could you clarify? What's the problem with the double digits? kwami (talk) 02:58, 21 August 2019 (UTC)

@Kwamikagami: The problem you mentioned above, that's the only thing that's not fixed. Your edit before didn't use reversed Chao for sandhi; that's why I've reverted your edit. — justin(r)leung _{{ (t...) | c=› }} 03:03, 21 August 2019 (UTC)

@Justinrleung: Thanks for taking care of that, Justin. It looks beautiful. That's a much more elegant coding solution. I got rid of the 'unreleased' diacritic as minor phonetic detail -- if we're not going to bother to indicate that vowel letters are not syllabic, it seemed like overkill, plus it's distracting next to the tone letters, and in some fonts (e.g. Gentium) it displays poorly, with the width of a full letter.

There's a similar sandhi-coding issue at Module:cdo-pron, and I'm worried that I may have introduced a sandhi error at Module:cjy-pron (which I mentioned on the talk page) that your solution should avoid (if we then change <¹¹> to <˩> before <¹> to <˩> in the output). I'm not comfortable enough with the coding to want to mess around with anything other than trivial changes,, if I tried to fix it, I might introduce additional bugs that I wouldn't notice. kwami (talk) 03:57, 21 August 2019 (UTC)

@Kwamikagami: I'm not sure if all the editors will agree entirely with removing the unreleased diacritic - this should have had more discussion. Different modules have varying degrees of phonetic detail, e.g. Sixian Hakka does have diacritics for non-syllabic vowels. I personally don't think we need so much phonetic detail, but I'm not sure if other editors agree. I'll take a look at the other modules soon. — justin(r)leung _{{ (t...) | c=› }} 06:21, 21 August 2019 (UTC)

If you want to revert, I'm not going to quibble. It seems like a distraction to me, and since it's completely predictable -- it would be like adding aspiration to English plosives. But the main reason was that it looks really bad in some popular IPA fonts. kwami (talk) 07:31, 21 August 2019 (UTC)

@Suzukaze-c How many books do you see use ꜖, and how many use "fake IPA"? Wyang (talk) 04:20, 25 August 2019 (UTC)

@Wyang: It's a matter of honesty, Wyang. We try not to misrepresent information. It is also consensus to use IPA on Wiktionary, since this is an international project. If you wish to use a regional alphabet instead, that should be a matter you bring up for discussion, rather than being disruptive and edit-warring over it. Really, it's far easier for each of us to learn the IPA than to have to learn half a dozen different conventions because we're all using something different for our favorite language. kwami (talk) 06:39, 25 August 2019 (UTC)

Ha. Sorry I didn't bring this matter up for discussion more than five years ago when I created this module. Sorry I didn't choose the politically correct IPA notations when I was thinking about how to design this module at the time and decided to use the much more common and less confusing number notations. Sorry I was being disruptive when someone just came along and 'fixed' everything. Sorry. Wyang (talk) 06:48, 25 August 2019 (UTC)

I think according to the rules I'm not supposed to respond to intentional nonsense, but if you think international standards are 'political correctness', then perhaps you should read up on the IPA and ISO standards before we continue this discussion. Depending on how much of this module you created, you provided a commendable service for Wikt, but you don't own it. As for "less confusing", how is it less confusing to label a tone [13] when you don't even explain if that's a high-falling tone or low-rising? Because it could be either, and several others besides. kwami (talk) 07:33, 25 August 2019 (UTC)

As I've told you, tone contours such as [13] are unambiguous in Chinese linguistics. Anyone who has read Tone (linguistics) would be able to recognise this. If you find even this confusing, I doubt you will be able to read much of the material on Chinese linguistics. Wyang (talk) 07:43, 25 August 2019 (UTC)

Of course it's unambiguous given the regional conventions that define it. But you don't provide those conventions. Instead you provide a spurious link to the IPA. Also, not everyone comes from your background. You're basically saying that only sinologists should have access to Chinese entries on Wikt, which is ridiculous. This is an international project, and at least attempts to follow international conventions, so as to reach the broadest possible audience. You argue that a well-defined and international standard is "confusing", whereas your regional and quite ambiguous notation is better just because it's more common within that regional convention. I could just as easily argue that American languages should be transcribed in Americanist notation, but that wouldn't help the majority who aren't familiar with it. kwami (talk) 07:49, 25 August 2019 (UTC)

Where is "꜖" in the well-defined and international standard? Wyang (talk) 07:51, 25 August 2019 (UTC)

As I've said, I'm fine with abandoning sandhi notation and giving the morpheme twice. But that's quite a drastic change just to satisfy IDONTLIKEIT, so you might want consensus. And it's certainly no reason to corrupt all the other supposedly IPA transcriptions, given that only a minority use sandhi notation.

Also, the Chao tone letters are intuitive, and if they're unfamiliar they only take a minute to learn. They are inherently easier than arbitrary numbers, which you have to remember to convert between every time you change languages. Let's see, I think '5' is supposed to be a high tone in this language -- oops, no, it's low tone. Damn, just screwed up my transcription. Oh well. Better that than Wiktionary use a rational notation. kwami (talk) 07:54, 25 August 2019 (UTC)

It's not about you being fine with abandoning sandhi notation. If you would like to invest your time and energy to rewrite the module to generate phonemic-phonetic IPA simultaneously and abandon sandhi notation, then I respect your sincerity. Not if you just say fine, I'm not keen on the sandhi idea when someone points out that the reverse letters are also non-IPA, and then do no work or even remove the sandhi information. Wyang (talk) 08:02, 25 August 2019 (UTC)

Then I'm afraid you're stuck with the extended IPA. That is the international standard, recognized world-round. It's the only such system. If you want to propose a modification, such as using all-rightstem letters (e.g. separated by an en dash, as you do the digits), then that might be a productive discussion. Though, AFAIK, the left-stem letters are what people who use the IPA use for sandhi. But demanding an insular convention because that's what you're personally more familiar with is not a convincing argument for a international project. kwami (talk) 18:47, 25 August 2019 (UTC)

Again and again, you are claiming standardness on no basis whatsoever. They are not "extended"; they are "nonstandard". You brought up the option of abandoning sandhi notation to bypass the nonstandardness of symbols, then show no seriousness in helping to achieve that. Now you brought up the 'standard' argument again, despite me requesting you to provide proof for that argument multiple times previously. Are you going to give me any evidence that the symbol "꜖" you used is "international standard, recognized world-round"? Maybe not? Wyang (talk) 08:01, 26 August 2019 (UTC)

Stop tone sandhi in Teochew

@Justinrleung I wonder when can this be implemented. For instance, in 暝旰, the first syllable is not supposed to have tone sandhi, at least in Singaporean Teochew (I'm not sure about in China). See [4] (1:23) and [5] (1:50) for examples. The dog2 (talk) 19:51, 5 January 2021 (UTC)

@The dog2: I'm not sure. This case seems to be a neutral tone on the second syllable. — justin(r)leung _{{ (t...) | c=› }} 19:57, 5 January 2021 (UTC)

@Justinrleung: So how do you suggest we deal with this case then? The dog2 (talk) 22:45, 5 January 2021 (UTC)

@The dog2: I'm not sure if the romanization actually has a way to deal with neutral tone. Perhaps we could use 0 as the tone, but the module is not ready for this yet. — justin(r)leung _{{ (t...) | c=› }} 22:53, 5 January 2021 (UTC)

Teochew /-n/ coda

We need a solution for representing /-n/ for 染蠻. @Justinrleung —Suzukaze-c (talk) 06:48, 10 March 2021 (UTC)

@Suzukaze-c: I think it could be written as -nn or -N, but either could be confusing to people who are used to POJ. — justin(r)leung _{{ (t...) | c=› }} 13:03, 10 March 2021 (UTC)

I think the main issue here is that the /-n/ final doesn't exist in standard Teochew, and in fact, one way you can identify a Teochew speaker in China is that they tend to mispronounce the /-n/ final in Mandarin as an /-ŋ/ final instead. In Pontianak (and Singapore too), Teochew speakers are often able to pronounce /-n/ because of influences from Malay and in the case of Singapore, Hokkien. The dog2 (talk) 18:39, 10 March 2021 (UTC)

@The dog2: Yeah, that's the issue but it seems that loanwords in Pontianak have changed this "ban" on final /-n/, which the romanization cannot accommodate without modification. — justin(r)leung _{{ (t...) | c=› }} 18:55, 10 March 2021 (UTC)

Would adopting Tai-lo for both Hokkien (instead of using an odd POJ-TL hybrid) and Teochew (or a TL-like system, both to avord the GZ romanization that I don't really like, and for easier comparison/unity with Hokkien) be too radical? —Suzukaze-c (talk) 22:42, 10 March 2021 (UTC)

I don't know. I actually find that the Teochew IPA seems to be inacurate. For instance, the "o" represents and /o/ sound in Hokkien (except in Zhao'an, where it is closer to the Teochew vowel), but has shifted to an /ɔ/ sound in Teochew (except in Shanwei, where it still largely follows the Hokkien vowel). And likewise, the "e", which represents more of an /e/ (French é) sound in Hokkien sounds closer to a /ɛ/ (French è) in Teochew. At least that's my observation. Listen to this and you'll know what I'm talking about. The dog2 (talk) 23:37, 10 March 2021 (UTC)

@The dog2: That's just notational. I think I've told you that because Teochew doesn't distinguish [o] from [ɔ], it's not necessary to use /ɔ/ even if it's probably more open. This is also irrelevant to how we decide to romanize. @Suzukaze-c: I would support Tai-lo for Hokkien because it is probably better for dealing with different dialects, but I also see a problem with it since we have POJ entries and not Tai-lo entries. Adopting a Tai-lo-like romanization for Teochew might be good as well, but might require a lot more work (and people need to adjust to it, and we can't just take pronunciations from the Teochew dictionary websites anymore, which might be a good thing). — justin(r)leung _{{ (t...) | c=› }} 01:27, 11 March 2021 (UTC)

Teochew phrases

@Suzukaze-c, Justinrleung this module seems cannot deal with Teochew phrases properly (like 青梅竹馬 and 廣泛性發育障礙). --沈澄心✉ 10:03, 8 April 2021 (UTC)

Changtai

@幻光尘, 沈澄心 I think I support using <o> for /ɔ/ in Changtai, and I think it would be better to use <o͘> for /eu/. This is because the words that are covered seem to indicate that this is how it would be distinguished in other dialects. I don't think we should just go by how things sound to match it to the POJ, but we should rather think about how the distinctions in POJ can be taken advantage of without making too many changes. One possible issue is /ẽu/ vs. /ɔ̃/, but it seems to be distinguishable with <o͘ⁿ> vs. <oⁿ>. — justin(r)leung _{{ (t...) | c=› }} 18:01, 26 January 2022 (UTC)

Changtai (Wuan Town) dialect does not have /o/, /ɔ/, /eu/ opposition, I think it is completely feasible. 幻光尘 (talk) 05:38, 27 January 2022 (UTC)

Hongmaogang

@沈澄心 Are you using 高雄市小港區大林蒲閩南話研究 as the source? I'm wondering why you picked Hongmaogang out of all the dialects that are available in the thesis. The focus of the thesis was on Dalinpu, not Hongmaogang. Also, there are some oddities with the tone notation, such as slashes. What are those supposed to indicate? — justin(r)leung _{{ (t...) | c=› }} 09:43, 5 February 2022 (UTC)

@Justinrleung: I'm using 記高雄一個偏泉腔方言－紅毛港閩南語初探. --沈澄心✉ 09:47, 5 February 2022 (UTC)

@沈澄心: I see. I wonder what the notation with the slashes mean, though. — justin(r)leung _{{ (t...) | c=› }} 09:57, 5 February 2022 (UTC)

@Justinrleung: slash separate different tone values of the same tone category: "(1) 陰平本調是中平調或高平調，是為自由變異，屬無定分音，記為[33/55]。……(6) 陽上有兩類：中降調[31]與中平調[33]，摻雜交替出現，屬無定分音。記為[31/33]。(7) 陽去調有兩類：中平調[33]與低降調[11]，摻雜交替出現，亦屬無定分音。記為[33/11]。" --沈澄心✉ 10:03, 5 February 2022 (UTC)

@沈澄心: I think the notation isn't quite intuitive. The "free variation" probably shouldn't be shown because it could be confusing. Perhaps we should follow what 高雄市小港區大林蒲閩南話研究 records instead, which is this:

• T1: 33 > 33; T2: 51 > 33; T3: 11 > 51; T4 (p/t/k): 31 > 55; T4 (h): 31 > 51; T5: 13 > 11; T6: 31 > 11; T7: 11 > 11; T8: 55 > 11. — justin(r)leung _{{ (t...) | c=› }} 10:37, 5 February 2022 (UTC)

Correct variant pronunciations for Penang Hokkien

I noticed that the 庇能 entry uses the wrong vowel letters for both POJ and Tâi-lô (the IPA looks fine). The infobox in the Penang Hokkien Wikipedia article shows the correct POJ and TL vowels: Pī-nɛ́ng and Pī-néeng respectively. I am asking an editor more familiar with this module to add the -ɛng endings as valid for Penang and to correct the correspondences to Tâi-lô. Additionally, e͘ [e + combining do above] is used for ɛ. If no one cares, I can edit myself. Thanks. AjaxSmack (talk) 16:13, 13 March 2022 (UTC)

@AjaxSmack: The TL should be corrected. As for POJ, I don't think there's a clear standard for how it should be extended. We have essentially decided to extend it following what TL does. If we decide to change this to follow other norms, such as using the combining dot instead of doubling of <e>, it will probably better to consider other instances where we would need to change this because it's a bigger change. — justin(r)leung _{{ (t...) | c=› }} 22:20, 13 March 2022 (UTC)

The ɛ is advocated at Pe̍h-ōe-jī § Regional differences. AjaxSmack (talk) 19:46, 14 March 2022 (UTC)

@AjaxSmack: No references are given there. It is used in Douglas' dictionary, which we could follow. I think Catherine Churchman's work usually uses <e͘> IIRC. @Freelance Intellectual, what do other Penang Hokkien sources that use POJ use to write /ɛ/? — justin(r)leung _{{ (t...) | c=› }} 19:56, 14 March 2022 (UTC)

All of your points are well taken. I didn't know about Churchman, but <ɛ> is also used in Tan Siew Imm's much more recent Penang Hokkien-English Dictionary, With an English-Penang Hokkien Glossary (→ISBN) which I'll add along with Douglas as a citation to the Wikipedia article (with links there for online perusal). Tan also says that Luc de Gijzel uses <ɛ> in his Penang Hokkien Dictionary and she mentions Churchman's use of <e͘ >. The fact that using double letters is very "un-POJ" (double letters look very Tâi-lô and could be seen as a POJ typo [which is what led me here]), that <eng> is used for /iŋ/ in POJ and that there are several dead-tree sources using <ɛ> make it a safe choice barring other developments. AjaxSmack (talk) 02:22, 15 March 2022 (UTC)

@Justinrleung: @AjaxSmack: De Gijzel and Churchman both use e͘, and I don't think Tan claims otherwise. Tan and Douglas use ɛ, but they also have other differences from modern POJ, such as Tan using ɵ instead of o͘, and Douglas using ɵ͘ with both a dot and bar. Following Douglas would be very old-fashioned. (Henning Klöter's "Written Taiwanese" includes a summary of the history of POJ, if you're interested.) I would personally prefer e͘ to ɛ, as it's more consistent with o͘, but I don't have a strong preference as long as entries are consistent with each other. The place name "Pī-nɛ́ng/Pī-néng" is phonologically unusual, not just in the vowel's combination with the final consonant, but also with the initial consonant (normally nasal initials only occur with nasal vowels), and it could be considered a loanword (cross-linguistically, loanwords often have unusual phonology). The e͘ vowel is used in many "native" Hokkien words, and not just in Penang. As justin(r)leung mentioned, it's more important to consider the overall orthography, rather than this one word. Freelance Intellectual (talk) 11:09, 15 March 2022 (UTC)

​​Thanks. I agree that "it's more important to consider the overall orthography, rather than this one word". Either <e͘ > or <ɛ> is fine with me (though I marginally prefer <ɛ> because it's comprehensible to readers who know IPA but not the intricacies of POJ and because it's easier to differentiate from unadorned <e>); I simply oppose the current situation where unsourced use of a Tâi-lô replacement in POJ is used. AjaxSmack (talk) 15:32, 16 March 2022 (UTC)

Changtai /ŋẽu/

Changtai IPA has /ŋɔ̃⁵³/ for ngó͘, when it should actually have /ŋẽu⁵³/ (e.g. here 午). — 59.21.138.53 18:28, 21 December 2023 (UTC)

Penang -iauⁿ

@Freelance Intellectual Right now, we input -iauⁿ for many syllables in Penang Hokkien that have -ioⁿ in Zhangzhou, e.g. 張 and 羊. While some sources such as Penang Hokkien Dictionary by Timothy Tye and 槟城闽南语语音研究 list those syllables as having the final /iãu/, other sources such as 馬來西亞檳城閩南語音韻研究 and 馬來西亞檳城福建話研究 list /iɔ̃/ as a phoneme distinct from /iãu/. 馬來西亞檳城閩南語音韻研究 (p.137-138) states iãu is a free variation of /iɔ̃/. This would cause 貓 and 娘 to share the same final and differ only in tone. I wonder how common the realization of /iɔ̃/ as iãu in Penang Hokkien is, and whether editors should input -ioⁿ instead of -iauⁿ into {{zh-pron}} for these syllables, given iɔ̃ and iãu represent distinct morphemes in some sources, although this may pose challenges when readings are directly taken from Penang Hokkien Dictionary which has only -iãu and no -iɔ̃. Thanks! RcAlex36 (talk) 17:20, 21 September 2024 (UTC)

@RcAlex36 My understanding is that ioⁿ is the more conservative pronunciation still used by some speakers, but it has merged with iauⁿ for most speakers. It's discussed in this video lecture (from 14 to 18 minutes in). Freelance Intellectual (talk) 21:43, 23 September 2024 (UTC)