Talk:阿莎力

a-sa-lih

@Mlgc1998 Please stop reverting my edits. The dictionary clearly writes it as a1-sa1-lih4. There's no way that's toneless writing. — justin(r)leung _{{ (t...) | c=› }} 14:54, 6 December 2022 (UTC)

@Justinrleung He asks, which dictionary are you talking about? or can just send me a screenshot of it to show him. Mlgc1998 (talk) 14:58, 8 December 2022 (UTC)

"There's no way that's toneless writing."

What's the exact basis for this statement?

How often do you hear native speakers say this word with a mid-level tone contour on the second syllable? Three-syllable Japanese loans mostly follow a T1-T2-T4 pattern in mainstream Taioanese. (The mid-west coastal towns are exceptional.) Some words go T1-T2-T2 or T1-T3-T4. T1-T1-T4 is not a pattern that exists. What dictionary did you get that from? Isn't it a red flag that they didn't record the T1-T2-T4 pronunciation at all? What tones do they have marked on OA-SÁ-BIH, KHI-MÓ͘-CHIH, SE-BÍ-LO͘H & THO͘-MÁ-TO͘H?

Also, as you probably know, romanized Hokkien & Taioanese have long had a convention of omitting tone marking on certain words (always marked with a final "-h"). 釆 (talk) 09:59, 10 December 2022 (UTC)

@Justinrleung @Mlgc1998 釆 (talk) 13:39, 10 December 2022 (UTC)

@釆: I'm not sure if Mlgc1998 has sent you a screenshot from the dictionary, but 臺灣閩南語辭典 in the reference section writes it as "a1-sa1-lih4" with the tone numbers. How is that toneless writing? — justin(r)leung _{{ (t...) | c=› }} 18:14, 10 December 2022 (UTC)

@Justinrleung That's not toneless writing. (As I hope you realize, it's not writing at all, just as "ɪɡˈzæmpl" is not writing.) And no, I haven't seen the screenshot.

How often do you hear native speakers say this word with a mid-level tone contour on the second syllable?

Isn't it a red flag that 臺灣閩南語辭典 didn't record the T1-T2-T4 pronunciation at all? What tones do they have marked on OA-SÁ-BIH, KHI-MÓ͘-CHIH, SE-BÍ-LO͘H & THO͘-MÁ-TO͘H? 釆 (talk) 01:55, 13 December 2022 (UTC)

@釆: 臺灣閩南語辭典 doesn't include too many loanwords. Usually it really doesn't show the tones when it includes a loanword, such as "【西米羅】se-bi-loh" or "thoo-ma-tooh". 阿莎力 seemed to be an exception to this. — justin(r)leung _{{ (t...) | c=› }} 02:43, 13 December 2022 (UTC)

@Justinrleung Does the compiler give a good explanation for this inconsistency? Can we really rely on his tone marking for non-literary loanwords?

He omits the dominant (& possibly only) -SÁ- pronunciation. What do you make of that, in the context of the above? 釆 (talk) 06:12, 13 December 2022 (UTC)

The dictionary could be off, or it could be that it's a rare pronunciation. — justin(r)leung _{{ (t...) | c=› }} 16:56, 13 December 2022 (UTC)

@Justinrleung Good analysis. Which do you think is the case?

If he had recorded both the T1-T2-T4 & T1-T1-T4 pronunciations, it might be the latter. Given that he only recorded T1-T1-T4, while somehow not marking the tones on SE-BÍ-LO͘H, etc., it seems safe to say that he was just being sloppy. And we have to wonder where else he was sloppy.

Keep the socio-cultural context in mind: many scholars are — evidently — only interested in the "Sinitic" aspects of Taioanese & Hokkien (& Hakka & so on).

If the T1-T1-T4 form *A-SA-LIH really exists, you should be able to track it down elsewhere, no?

If we mechanically entered data into Wiktionary from dozens of wordbooks, that would be ... a mess, probably. But that would still be way better than mechanically entering data into Wiktionary from just one wordbook. Let us contribute to Wiktionary with our knowledge. Wiktionary should be a repository of our knowledge, not a record of our learning. That goes for all of us. 釆 (talk) 12:27, 14 December 2022 (UTC)

@Justinrleung (Is there further reasoned opposition to removing the T1-T1-T4 form for now?) 釆 (talk) 15:38, 16 December 2022 (UTC)

@Mlgc1998 釆 (talk) 15:39, 16 December 2022 (UTC)

@Justinrleung 釆 (talk) 12:31, 16 May 2023 (UTC)

etymology

The Mandarin word 阿莎力 is clearly a loanword from Taioanese, not a direct loan from Japanese. We should edit this entry to reflect that. 釆 (talk) 09:46, 10 December 2022 (UTC)

@Justinrleung (In case you haven't seen this yet.) 釆 (talk) 02:22, 13 December 2022 (UTC)

@釆: I've attempted to make this a little clearer. — justin(r)leung _{{ (t...) | c=› }} 02:32, 13 December 2022 (UTC)

@釆: What are your thoughts on where Hakka would fit? Would Hakka borrow directly from Japanese or via Taiwanese? — justin(r)leung _{{ (t...) | c=› }} 22:01, 13 December 2022 (UTC)

This is a question of fact. From what little we know, either could've been the case, or both. If it turned out that non-Japanese-speaking Hakka speakers didn't use this word till the 1970s, for example, there would be a strong case for Taioanese. I have no knowledge of this, though. It's plausible that the word entered Hakka from both languages generally, or that East Coast Hakka got it from Japanese while Southern Hakka got it from Taioanese. "A loan from Japanese and/or Taioanese" might be the most hygienic formulation for now. 釆 (talk) 03:28, 15 December 2022 (UTC)

@Justinrleung 釆 (talk) 03:29, 15 December 2022 (UTC)

a̍t-sá-lih & *at-sá-lih & *à-sa-lí

@Justinrleung What is the basis of these two pronunciations?

"*À-sa-lí" doesn't seem unreasonable (but I don't speak much Hakka). Does nobody say "*à-sa-li"? And what's the Hoiliuk pronunciation?

"*At-sá-lih" seems to be in error.

"*A̍t-sá-lih" seems to be a written representation of a pronunciation where the first syllable takes a checked mid-level surface tone. In other words, I'm not sure there's a pronunciation where the first syllable takes a low tone, and I'm asking you for solid evidence of it. 釆 (talk) 02:16, 13 December 2022 (UTC)

• What is the basis of these three pronunciations?

釆 (talk) 02:17, 13 December 2022 (UTC)

Does nobody say "*à-sa-li"? → Does nobody say "*à-sa-lì"? 釆 (talk) 02:18, 13 December 2022 (UTC)

@釆: Please see the quotations listed under the definition. There is a chance that people might be slightly less "accurate" when writing in POJ, so it might need more checking. As for a̍t-sá-lih, I think tone 8 (-p/t/k) usually sandhis to tone 4, so it's not entirely low. Right now the module shows it as 32 in Chao tones, which is pretty close to a mid-level. The 海陸 pronunciation should be about the same as the 四縣 pronunciation, aˇ sa li` in the Taiwanese Hakka Romanization System according to 客語外來語（含原閩客國(語)互借詞）. I might have to check other sources like 徐兆泉's dictionary to confirm if other pronunciations are recorded for 四縣 and 海陸. — justin(r)leung _{{ (t...) | c=› }} 02:29, 13 December 2022 (UTC)

> There is a chance that people might be slightly less "accurate" when writing in POJ

No. Rather, POJ — whether for Taioanese, Hokkien, or Hakka — has rarely been "100% phonologically faithful", for better or worse. Like any script used en masse.

> As for a̍t-sá-lih, I think tone 8 (-p/t/k) usually sandhis to tone 4, so it's not entirely low.

Running T8 is always low, across all dialects of Taioanese. (I can't speak for the far-flung dialects of Hokkien.) Running T8 is not identical to standing T4 in any dialect of Taioanese, although the tones are sometimes taught that way as a shortcut. (Remember, their target audience is native speakers looking to gain literacy, not people trying to learn the language itself.)

If T8 in running position represented a Chao 32 (which it does not), we might be OK here, but we'd be in trouble everywhere else throughout the language, where running T8 is a low tone.

A mid-level checked contour in running position doesn't fit w/i traditional Taioanese phonology. Such syllables only occur in non-literary loans from Japanese. This is why 白話小詞典 represents that pronunciation as "ÂT-SÁ-LIH", using an expanded version of POJ that has yet to gain any traction or acceptance. We probably need to borrow the IPA tone marking apparatus. However, I'm not 100% sure people still say "AT˧-SÁ-LIH" (of course, the -T assimilates to -S). I don't "doubt" it — I just can't "guarantee" it.

Note that many linguists (to use the word broadly) over the years have chosen to just ignore such questions. This is why, for example, dictionaries offer spotty or unreliable coverage of non-literary Japanese loanwords, with 白話小詞典 being the exception. 釆 (talk) 06:40, 13 December 2022 (UTC)

@Justinrleung (Hoiliuk POJ "*À-SA-LÍ" means high-level on the first syllable, low-level on the middle syllable, and low-to-mid rising on the last syllable. I'm sure Hoiliuk Hakka has this item, but there's no way it's pronounced that way. I think we're OK omitting it for now though.) 釆 (talk) 06:44, 13 December 2022 (UTC)

@Justinrleung > Please see the quotations listed under the definition.

Good hard work. Are all those exact quotations, BTW? 釆 (talk) 06:53, 13 December 2022 (UTC)

@釆: About the quotations, yes, they are exact quotations from the source. We generally do not change the orthography employed in the primary sources that we are quoting from.

We currently do not support non-四縣 varieties of Taiwanese Hakka in the module, which is unfortunate. It is, however, on our list of things to do.

I'm wondering if you have any sources (acoustic preferably, or impressionistic if that's all we have) for the claim that say sandhied tone 8 is low in all dialects of Taiwanese (or at least in the so-called 通行腔). — justin(r)leung _{{ (t...) | c=› }} 15:31, 13 December 2022 (UTC)

@Justinrleung Any recording will back that so-called claim.

https://www.youtube.com/watch?v=o4EIGd6PoHA

We may have to ask questions about the learning materials & processes that (mis)led you to think otherwise.

> We generally do not change the orthography employed in the primary sources that we are quoting from.

Interesting.

Why are the Taioanese & Hakka examples mixed in together as if they were from the same language? (Serious question.) 釆 (talk) 16:27, 13 December 2022 (UTC)

@釆: I cannot trust my ear, so I would like some study on this phenomenon. I know you're a native speaker, so you would know how the language works, but you are only one datapoint. In the same way, one video does not show all of Taiwanese.

As for why Taiwanese and Hakka examples are mixed together, they are both put under "Chinese" and quotes are sorted chronologically. — justin(r)leung _{{ (t...) | c=› }} 16:49, 13 December 2022 (UTC)

@Justinrleung (Thanks, but) I'm not a native speaker, neither technically nor practically. Pretty much any recording in natively-spoken Taioanese will be what you're looking for. On Youtube alone there's enough material to keep you occupied for years. The practice can't hurt. You need to be able to trust your ear eventually for this work that we (but mostly you-all) are doing.

Can you find recordings of natural native speech with mid-level running T8? If not, do you have other conclusive proof that such dialects exist?

BTW, the recordings from the ROC MOE dictionary are unnatural, but they do show how running T8 is always low.

https://www.moedict.tw/'%E7%99%BD%E7%85%A0

https://www.moedict.tw/'%E7%86%9F%E4%BC%BC 釆 (talk) 15:57, 16 December 2022 (UTC)

@Justinrleung > they are both put under "Chinese" and quotes are sorted chronologically

Are you far enough along in your studies of Hokkien or Taioanese to perceive that Hakka is a totally different language(s) even in terms of intelligibility? 釆 (talk) 16:00, 16 December 2022 (UTC)

@釆: Of course, but please see WT:About Chinese. — justin(r)leung _{{ (t...) | c=› }} 16:36, 16 December 2022 (UTC)

@Justinrleung Is this the "state of the art"?

https://en.wiktionary.org/wiki/Wiktionary_talk:About_Chinese#Language_treatment:_Only_the_macrolanguage_is_treated_as_a_language? 釆 (talk) 01:30, 21 December 2022 (UTC)

@釆: Things on the talk page are not official, but shows ongoing discussion. Please refer to the main page for clarification. (It's also getting outside the scope of this entry. WT:BP is the place for larger discussions.) — justin(r)leung _{{ (t...) | c=› }} 07:28, 21 December 2022 (UTC)

> Can you find recordings of natural native speech with mid-level running T8?

Rather, can you find recordings of natural native speech with mid-to-low falling or mid-level running T8? 釆 (talk) 16:01, 16 December 2022 (UTC)

彈了袂 bái，介 A-SA-LIH。

@Mlgc1998 @Justin I added this example from Twitter. For some reason, the back-end took the spaces out and I haven't been able to put them back in. (They need to be in there.)

Also, this is an example of our word written tonelessly.

The POJ line automatically converted 袂 to BĒ, which is nifty, but how do we indicate that this could be either BĒ or BŌE?

I'll lay off the question of the "trad." & "simp." lines for now. 釆 (talk) 16:56, 13 December 2022 (UTC)

@釆: We generally don't indicate all possible pronunciations. People can just click into the link on 袂 and see that for themselves. As for spaces, you could maybe put in &{}n{}b{}s{}p{};{} to preserve spaces. I haven't been very concerned about spaces unless it's between two POJ words. — justin(r)leung _{{ (t...) | c=› }} 17:13, 13 December 2022 (UTC)

Thanks for that.

Intuitively — in line with custom — we realize there has to be spaces between POJ words & the Hindu-Arabic numbers (0-9). The same is true with POJ words & Hanji, at all times — although Mandarin-educated *learners* may not observe this (apparently always unwritten) rule. The need for spacing "emanates" from the POJ side; it's a Latin-script thing. And it does seem to be more than justified aesthetically, although custom alone should be reason enough.

Given the context (the orthographic tradition, etc.), it would be misleading to convert 袂 in the original to just BĒ. If there's such a thing as over-transliteration, that would be it. In this as in many cases, the original doesn't specify which of the (in this case two) orthographically represented pronunciations "applies", so it's "up to the reader", in a sense. We have to "pass that ambiguity (a.k.a. choice) on" to the Wiktionary user. 釆 (talk) 16:18, 16 December 2022 (UTC)

@Justinrleung 釆 (talk) 16:18, 16 December 2022 (UTC)

@Justinrleung @Mlgc1998 I added another example. Now there should be no spaces before the hyphens, and "kā chò/chòe-lo̍h-khì" should be "kā chò/chòe--lo̍h--khì" (or "kā chò/chòe--lo̍h-khì").

若我ê作法tō khah a-sah-lih，kā作 --落 -去。 [Taiwanese Hokkien, trad. and simp.]

Nā góa ê chò/chòe-hoat tō khah a-sah-lih, kā chò/chòe-lo̍h-khì. 釆 (talk) 05:20, 17 December 2022 (UTC)

@釆: Should be fixed (with a temporary patch solution) because I can't think of a way to make the module (as-is) cooperate with the input. As I've said before, we do not usually show all possible pronunciations. It's just like in the English entries where we would not show both colour and colour in examples. It is enough to provide one possible reading. If you have an issue with that, it would be best to take that to WT:Beer Parlour for wider discussion since it's not an issue particular to this word. — justin(r)leung _{{ (t...) | c=› }} 06:20, 17 December 2022 (UTC)

@Justinrleung Not showing all possible pronunciations on our own examples is common sense. 👍

In this case we're not creating an example. We're citing & transliterating something somebody else wrote. It would be presumptuous & sloppy to project pronunciations onto someone else's written expression. I see no problem in arbitrarily picking pronunciations for our own creations. But if it's someone else's existing written expression — and we ARE presenting it as their written expression — we can't just project an arbitrary pronunciation onto it. 2001:E68:542A:D581:356E:712:EC32:EA0A 07:00, 4 February 2023 (UTC)

@Justinrleung It seems sensible to recognize that whereas we can create our own examples using any pronunciation or spelling that has currency, we can't project an arbitrary pronunciation onto somebody else's work. If this discussion belongs elsewhere, please share a link to that space. 釆 (talk) 18:03, 15 May 2023 (UTC)

@釆: WT:BP is the place to discuss this. Thanks. — justin(r)leung _{{ (t...) | c=› }} 20:28, 15 May 2023 (UTC)

@Justinrleung New topic? 釆 (talk) 01:24, 16 May 2023 (UTC)

@釆: Yes, click on "Click here to start a new Beer parlour discussion." and start a new discussion in BP. — justin(r)leung _{{ (t...) | c=› }} 02:10, 16 May 2023 (UTC)

erroneously reverted edit?

@Justinrleung's note: "removed out of process, and some support is seen in print/quotations (see talk); please consider an rfv-pron if this is to be done in process)"

It's great that you have "processes". Right now you're just using procedure your knowledge of Wiktionary to block people from correcting (your) errors.

Like most other languages written with an alphabet, the romanized Taioanese script doesn't reflect current pronunciation 100%. A̍T-SÁ-LIH & AT-SÁ-LIH are just approximations of a mid-level clipped tone contour on the first syllable, which the script doesn't provide for. This is also why you'll find "BA̍K-KHÚ-MIH-Á" on the internet; nobody pronounces this word with a low tone on the first syllable.)

Likewise, A-SA-LIH is just toneless writing, which is common for non-book Japanese loans. Nobody says this word with a mid-level tone on the second syllable. But you knew that, right? 釆 (talk) 12:29, 16 May 2023 (UTC)

@釆: I'll just respond to the two pings here. While there is some reason to doubt what is recorded in 董忠司's dictionary (a¹-sa¹-lih⁴ = a-sa-lih), I don't know if the doubt is strong enough for removal of the pronunciation unless there's more input from native speakers. At best, we can qualify it in the pronunciation notes that it might be rare. As for the other pronunciations, at-sá-lih might warrant removal since it is only attested in writing, not as a pronunciation guide, but I think a̍t-sá-lih should still be kept as the closest approximation to what is recorded in 臺灣閩南語常用詞辭典 (at₃ sa₅₅ lih₃). We can add a note that the first syllable is intended to be mid level.

As for the "out of process" thing, I apologize for the procedural stuff, but please make use of it if you think it's applicable. You can tag pronunciation sections with {{rfv-pron}}. — justin(r)leung _{{ (t...) | c=› }} 22:21, 16 May 2023 (UTC)

@Justinrleung > a̍t-sá-lih should still be kept as the closest approximation to what is recorded in 臺灣閩南語常用詞辭典 (at3 sa55 lih3

This would be misleading. It could be sensible in the orthography, but not as a purported pronunciation. (Again, to unpack this statement, you have to realize that romanized Taioanese — & Hokkien — is not a system of purported pronunciations.)

Even "at3 sa55 lih3" is phonemic. No [t] is ever heard at the end of the first syllable. This phonemic analysis reflects a shortened vowel (first syllable) with a "geminated" [s]. 釆 (talk) 07:58, 17 May 2023 (UTC)

@Justinrleung > While there is some reason to doubt what is recorded in 董忠司's dictionary (a1-sa1-lih4 = a-sa-lih), I don't know if the doubt is strong enough for removal of the pronunciation unless there's more input from native speakers. At best, we can qualify it in the pronunciation notes that it might be rare.

Assuming "a1-sa1-lih4" is a tight transcription, why didn't he also include the "more mainstream" pronunciations?

Clearly, either "a1-sa1-lih4" was loose (i.e. not marked for tone), or it was erroneous; 董忠司 may have believed that non-book Japanese loanwords are not genuine "Hokkien" words and don't deserve full lexicographic attention.

Again, if you believe that "a1-sa1-lih4" was a tight transcription, how do you account for the absence of a tight transcription of the mainstream pronunciation(s)?

"A1-sa1-lih4" is clearly either loose (I say this w/o prejudice) or in error. 釆 (talk) 08:09, 17 May 2023 (UTC)

@釆: Again, "not marked for tone" is not a possibility because the tone numbers are marked explicitly there. If you have doubts about it, it is at best an error (either a typo or a genuine mistake on 董忠司's part). The argument from silence is a little weak - if a33-sa33-lih3 is indeed used by some people, then what 董忠司 recorded is still a valid pronunciation, even if it's not mainstream. (And I don't know if it's necessarily recording the mainstream, since another "non-mainstream" thing it does is Tainan -ionn.) I don't know if it's a clearly obvious assessment to assume that the editor was sloppy in this particular case, though I agree that in general the dictionary is not very helpful with loanwords. Since it's just one source, I guess it's not really a hill to die on, so I would be fine with removing this pronunciation until more evidence surfaces.

As for a̍t-sá-lih, I get that it might be a little inaccurate. That's why I said we could add a pronunciation note. Perhaps, it's best to just have /at3 sa55 lih3/ in the pronunciation notes describing the gemination as well. What do you think? — justin(r)leung _{{ (t...) | c=› }} 14:36, 17 May 2023 (UTC)

@Justinrleung Sorry for the lag! I think removal would be best.

I've never heard "*A-SA-LIH" with a mid-level tone on the 2nd syl. Furthermore, multi-syl. non-book Japanese loans follow certain tone patterns. There is no 3-syl. mid-mid-low pattern.

Had he recorded a mid-mid-mid pronunciation IN ADDITION to the common mid-high-lowish pronunciation, then it would appear that he had found a unique non-mainstream pronunciation. The fact that he omitted the mid-high-lowish pronunciation tells us a lot! (To figure out exactly what's going on, though, we'd have to leaf through the book a bit and see what he does with other non-book Japanese loans.)

As for "*A̍T-SÁ-LIH", I guess /at22 sa33 lih21/ would be fine as a phonemic transcription. Phonemically, there are three tone tiers in Taioanese (& many dialects of Hokkien). I guess five-tier notation is O.K. for phonetic transcription, but then the "-t" would have to go, if I'm not mistaken; I don't think there's ever a [t] there in actual pronunciation.

BTW, [ɔ̃] in (orthographic) "-iuⁿ" seems to occur in the Kagī region as well ... and maybe as far north as Táula̍k, but don't quote me on that. 釆 (talk) 13:59, 25 June 2023 (UTC)

@釆: Okay, let's remove a-sa-lih until further evidence surfaces. As for a̍t-sá-lih, does 22 33 21 mean something like 33 55 31 in Chao tone numerals? — justin(r)leung _{{ (t...) | c=› }} 01:23, 26 June 2023 (UTC)

@Justinrleung Yes, but the Chao numerals can't be used in phonemic notation. (We can go over this in more detail if necessary.)

Also, in POJ-as-phonemic-notation, it would be better to write "ât-sá-lih" than "a̍t-sá-lih." Neither is perfect, and I think it might be better to choose neither. But if we had to choose one, "ât-sá-lih" is much less misleading. It indicates that something unusual is going on; and it's intuitive in the sense that most experienced readers will be able to guess at what's being indicated.

In POJ-as-script, note how recent all internet occurrences of "A̍T-SÁ-LIH" are. It seems clear that they're all downstream of a few pedagogical sources that used the misleading notation "A̍T-SÁ-LIH," with Wiktionary possibly being one of those. What I mean is this: If it was the case that the general writership of the language has been accustomed over time to representing mid-level checked syls. as T8 syls., so be it; at the very least, it would be part of the reality that we must catalog & reflect. But dictionaries & pedagogical materials are not "entitled" to generate such imprecision a priori. In other words, roughly speaking, those deliberately engaged in acts of teaching or lexicography are held to a higher standard, just as a restaurant would tend to respond to matters of sexual assault with three different standards of seriousness depending on whether the accused is a manager, a low-level employee, or a customer. (Surely not the best example, but hopefully it illustrates the idea.) 釆 (talk) 04:17, 26 June 2023 (UTC)

Really not the best example. Better: For instance, what a lexicographer of language X writes in language X in a swim club chat group on LINE or social media is part of the usage for language X in a way that his or her lexicographical work on language X is not. And all written languages accumulate arbitrary exceptions & exceptions-to-exceptions over time. And it's the lexicographer's job to capture & catalog such exceptions, and it's the teacher's job to prepare students to deal with or even wield such exceptions. But the lexicographer or teacher may not generate such exceptions; and if they do, the adoption of those exceptions among a subset of their audience cannot be taken as evidence that the exception has taken hold generally. 釆 (talk) 04:27, 26 June 2023 (UTC)

@釆: I am wondering if you have any sources (educational/academic) that talk about the specifics of these tone patterns in loanwords (and how they differ from native word tone sandhi). As of now, our module would not be able to handle something like ât. It would be good to make changes to the module (if necessary) not just for this exceptional case but at a wider level (if this is more pervasive than this isolated case). (I will also note that the recent update to 教典 has altogether abandoned the Chao numbers for loanwords and has moved on to normal TL tone notation, which I do think is slightly problematic.) I would also like you to elaborate on why Chao cannot be used; the 3 tone tier approach would make it notationally confusing since IPA notation is usually based on a 5 tier system. — justin(r)leung _{{ (t...) | c=› }} 00:42, 27 June 2023 (UTC)

@Justinrleung Amoy Hokkien — & all modern dialects of Taioanese — have three tone tiers, phonemically. There are three contrasting tone tiers. (I’m under the impression that Penang Hokkien has two contrasting tone tiers. Some dialects of Choanchew Hokkien may have four.)

From here, it should be clear why there’s no place for the five-tier Chao notation in PHONEMIC notation. It can be useful in a narrow transcription, though.

The “-T” in *ÂT-SÁ-LIH, OTOH, is a phonemic representation of what surfaces as gemination of the [s].

Phonemic & narrower phonetic notation can both be useful in turn, but mixing the two is no good (i.e. incorrect).

The “updates” to the Repub. of China dictionary are interesting. Their handling of *ÂT-SÁ-LIH (asterisk as this is just a tentative spelling; https://sutian.moe.edu.tw/zh-hant/su/20372) has leapfrogged ahead of Wiktionary’s.

If you click on the audio for the Repub. of China entry, you can clearly hear how the 1st syl. is mid-tier. The same is true for *BÂT-TÉ-LIH. (There are artificial elements in their audio, though, most notably their -IAN. The [t] in *ÂT-SÁ-LIH is also artificial, to the best of my not-unlimited knowledge.)

Sources are a serious matter. Lexicography is a hardcore undertaking, and the burden is on non-native speakers, especially, to be sure we know what we’re doing. Even more so for non-speakers, which you seem to be for the time being.

I sense that I’m “up against” some kind of a loop here.

I recommend Tiuⁿ Jū-hông’s PE̍H-ŌE SIÓ SÛ-TIÁN as a go-to reference, although (IMO) it’s not w/o flaws. And I recommend Âng Î-jîn’s 1980s treatise on tone in Holo, to be read immediately. (Among other things, he lists the tonemes that can’t be represented in conventional romanized Taioanese.)

You may need to order the books. But it would be mind-boggling for a non-native speaker to try to do Taioanese (or even Hokkien) lexicography w/o the benefit of these two sources — and several others, perhaps.

(While Âng I.J. is admired by the establishment (?), most Sinologists — Taioanese-speaking or not — wouldn’t touch his treatise on tone with a 10-foot pole. It’s rigorous. Tiuⁿ T.J., on the other hand, is studiously ignored by most Sinologists, for reasons easy enough to guess at.)

You can’t evaluate the mainstream Neo-Sinological takes on Taioanese or Hokkien unless you expose yourself to a version of either or both languages that isn’t curated by the same network of academics, instructors & talking heads.

Are you going to look at what is, or not?

P.S. Realize, most traceable loanwords in Taioanese don’t “look like” *BÂT-TÉ-LIH or *ÂT-SÁ-LIH…. KÀU-IO̍K 教育 is a modern loan. So is SÛ-TIÁN 詞典. “Native” is a relative & complex concept. HÔ͘-TÔ͘ 胡途 & LĪ-HĀI 利害 & even IÂ-SO͘ 耶蘇 were pre-modern loans from Mandarin. 釆 (talk) 18:49, 29 June 2023 (UTC)

@釆: (For starters, I would consider myself a non-native speaker, though at a higher basic / lower intermediate level, and I have had limited access to speakers to practise with.) I still do not really see a huge notational issue by using Chao numbers in phonemic representation; can we not pretend 3 = 5, 2 = 3, 1 = 1? Hong Kong Cantonese only has four tiers, but Cantonese linguists see no problem with using the 5-tier Chao numbers. Also, there is some discussion before for using Chao tone letters instead of tone numbers because that is more "standard" in IPA notation. Perhaps that might obviate this (minor, IMHO) issue here.

Thank you for pointing me to those two sources. For Ang Ui-jin's work, are you talking about 台灣河佬話語聲調研究? And I apologize if these are sources that we have been neglecting. These are not accessible on the Internet, which is one big obstacle to any online voluntary (i.e., amateur, non-professional, non-funded) lexicography. I will try my best to get a copy, and I'm sure these sources will enhance our work here.

And of course, it's clear to me that there are loanwords from other languages, such as Mandarin, but in context, we are of course talking about Japanese loanwords and tone assignment. — justin(r)leung _{{ (t...) | c=› }} 03:37, 30 June 2023 (UTC)

@Justinrleung

(教育 and 詞典 ARE Japanese loanwords, in Taioanese. "People" tend to forget that, & context itself does not help us remember.)

(Voluntary lexicography should be truth-oriented. Otherwise, why volunteer? Also, the herd does not cleave to state-affiliated materials only b/c they're free, but rather out of loyalty (naively) & even laziness. There are other free resources that they uniformly ignore, partly out of political animus.)

> 台灣河佬話語聲調研究

Yes.

> I still do not really see a huge notational issue by using Chao numbers in phonemic representation; can we not pretend 3 = 5, 2 = 3, 1 = 1?

Where else in linguistics do people even try to knowingly (i.e. not by mistake) bring non-phonemic elements into phonemic representation?

> Hong Kong Cantonese only has four tiers, but Cantonese linguists see no problem with using the 5-tier Chao numbers.

That's fine, in non-phonemic representation.

There is no justification for using five-tier representation in phonemic representation for a language with < 5 phonemic tiers. It feels silly even talking about it.

To comprehend segmental phonemic representation, we have to familiarise ourselves with the sound system. Why should things work differently when it comes to tone? Although (very) tangential, I found that reading Said's ORIENTALISM helped me understand issues like why unscientific practices are widely & broadly tolerated in Neo-Sinology. Also see pages vi through lxiv prefacing Cikoski's LEXICON OF CLASSICAL CHINESE, freely available online but generally ignored. (+ My command of the lit. is far from complete, and I'm more than keen to find more writings in the same vein, possibly by Asian thinkers.) 釆 (talk) 08:24, 30 June 2023 (UTC)

@釆: In context, we were only concerned with loans that aren't orthographic borrowings, because the tone assignment has a particular pattern, as you mentioned. Whether those other words you mentioned are loans from Japanese, Mandarin or some other language, I would be interested in seeing the evidence, but this page is probably not the place to add yet another tangent.

And of course, we are aiming for documenting the language as truthfully as we know it. This is a no-brainer. However, sometimes I don't think lexicography is as black-and-white and there is bound to be issues of contention that we would run into. Ideally, all sides of contention are represented (as is done in Wikipedia with their WP:NPOV), and I am totally for it.

Admittedly our IPA notation is confusing, and it often involves too much phonetic detail for slashes. I have been thinking of splitting it into phonetic/phonemic notations, but I haven't thought it through properly. However, I don't really know if there's a way of representing three- or four-tier tone systems using standard IPA. I think Chao tone letters (˥ ˦ ˧ ˨ ˩) are one way to go, or we would try it out with diacritics. This again is a larger issue, and perhaps should be discussed outside of this talk page (as I've said many times). — justin(r)leung _{{ (t...) | c=› }} 08:55, 30 June 2023 (UTC)

@Justinrleung

I believe ˥ ˧ ˩, for example, could be used to represent the tone tiers in phonemic notation for a variety with three tone tiers. (Why not? They're not "5", "3", & "1"; that's at most in people's heads.)

Why not use phonemic notation all the way through?

(And, modern Sinology is generally concerned with consensus rather than truth. Sinological sources have to be used carefully if one is aiming for truth.)

> In context, we were only concerned with loans that aren't orthographic borrowings,

Sure. But it's a very, very common mental mistake to think of orthographic borrowings as not-borrowings.

> Whether those other words you mentioned are loans from Japanese, Mandarin or some other language, I would be interested in seeing the evidence,

Mostly one has to go through the written record as well as dictionaries. A supple, community-centered understanding of the history (as opposed to off-the-shelf stuff centered elsewhere) is crucial. The Tiuⁿ Jū Hông dictionary & the 台字田 are two sources that indicate the etymology of Taioanese words. The 台日 also tags Japanese loanwords. These three sources aren't always in agreement, which is interesting. The 台字田 is obviously the chattiest. The path a word takes in entering a language is "a question of fact", and sometimes the same word may have entered Taioanese from Japanese while, some years later, entering Hokkien via Mandarin (from, indirectly, Japanese). And so on. Remember that if Tagalog KAMATIS entered Tagalog from Spanish, it is first & foremost a Spanish loan even though it entered Spanish from a Nahuatl language. If PIĀN-TONG entered Taioanese via Japanese 便當 (弁当), it is first & foremost a Japanese loan regardless of whether Japanese got it from, say, Classical Chinese.

Sinological pseudo-dictionaries are Sinograph-centric; they're not concerned with whether orthographic borrowings were borrowed, or from where. I'm not aware of any Hokkien dictionary that tags loanwords systematically. There also seems to be a lack of resources that help folks understand history from a Hokkien (Banlam)-centered POV. 釆 (talk) 12:58, 1 July 2023 (UTC)