Template talk:my-IPA

New template

@Angr, Atitarev I created this template for Burmese. Most results seem to be correct, although I haven't checked it thoroughly. I'm sure there's plenty of room for improvement (such as voicing in Okell). It would be great if you could take a look at this template and kindly provide your thoughts. Thanks! Wyang (talk) 11:39, 31 January 2017 (UTC)

• Thanks, Wyang! One thing that would make it easier to use is if it could take a Latin-alphabet (ASCII characters) transcription as the phonetic respelling (probably in addition to, not instead of, the Burmese-alphabet respelling). —Aɴɢʀ (talk) 12:46, 31 January 2017 (UTC)

@Wyang Thanks! You've actually fulfilled my wishes! (Remember how I just casually asked you about creating a module?) – AWESOME meeos ！ * (chōmtī hao /t͡ɕoːm˩˧.tiː˩˧ haw˦˥/) 12:49, 31 January 2017 (UTC)

• What does the word= parameter do? Is it ever anything besides the pagename? —Aɴɢʀ (talk) 12:51, 31 January 2017 (UTC)
• In hsibouñ̀ and ouñ̀zi, the grave accent needs to be over the u, not the ñ. —Aɴɢʀ (talk) 12:53, 31 January 2017 (UTC)

@Angr, Atitarev, Awesomemeeos

Thanks for the comments. I added a bit of documentation to the template page, and corrected some of the errors (voicing in Okell, accent before nasal ending in Okell, -wu- rhymes, glottal stop, and args["word"] logic). Auxiliary Latin-alphabet respelling is a good suggestion; I wasn't sure what the Latin respelling should look like but I created a Burmese-alphabet respelling table, and I will work on the Latin respelling if we can have a ASCII transcription system parallel to the Burmese one.

|word= only represents the page title, so it shouldn't be used outside of this documentation page. I think I've fixed everything I can fix at the moment. Please let me know if there are more suggestions or comments. Thanks! Wyang (talk) 08:25, 1 February 2017 (UTC)

@Wyang Just like the manual transcriptions, do a [phonemic] transcription, based on the /phonetic/ one, where further subsitution rules should be made to render the phonemic version – AWESOME meeos ！ * (chōmtī hao /t͡ɕoːm˩˧.tiː˩˧ haw˦˥/) 08:31, 1 February 2017 (UTC)

@Wyang: For an ASCII-based respelling, we could use Okell's transcription with the following changes (for example, these can obviously be tweaked):

Okell	ASCII
ʔ	?
ñ	~
ă	a'
á é eí í ó oú ú	a/ e/ ei/ i/ o/ ou/ u/
à è eì ì ò où ù	a\ e\ ei\ i\ o\ ou\ u\
p etc.	+p etc.

I picked a' for ă since you're already using the apostrophe to mark minor syllables in the Burmese respelling; likewise I used + for voicing since you're already using it in the Burmese respelling. —Aɴɢʀ (talk) 10:06, 1 February 2017 (UTC)

Congratulations on another accomplishment, Frank! I have acquired a few books on Burmese but haven't actively started learning it. I am very glad there is now a way to generate a pronunciation and some more phonetic transliterations based on phonetic respellings just like Thai. Pity, it's not available in other dictionaries for look-up. I find the Burmese script and reading rules the hardest in the region (after Tibetan). Burmese is on my to-do-list, anyway, but I'm not sure I'll be able to assist much at this stage, we'll see. I will learn from you guys (Wyang and Angr). --Anatoli T. ^{(обсудить}/^вклад) 11:35, 1 February 2017 (UTC)

@Angr Thanks for the suggestions. I incorporated the ASCII respelling as well, and added a few cases using ASCII respellings as input. There are some multiple-to-one MLCTS-Okell mappings (such as ky / kr -- c) and I think we should try to make sure the MLCTS and ALA transliterations are correct when using ASCII respellings. Please kindly review this again. Do you think this could replace {{my-roman}} and be placed in the Pronunciations section if it is mature enough?

The MLCTS and ALA transliterations can be taken directly from the pagename, can't they? They shouldn't need access to the respelling at all. I do think the ASCII respelling shouldn't be displayed, though; it's just for ease of input. Can you display the Burmese-alphabet respelling even when the ASCII is entered? I do hope that this can replace {{my-roman}} at some point. —Aɴɢʀ (talk) 11:58, 1 February 2017 (UTC)

@Angr There are some uncommon cases where MLCTS cannot be generated from page title, such as ယောက်ျား. I'm not sure what's the best way here - perhaps using |word=ယောက်ကျား. I've changed the respelling display to the Burmese-alphabet one, when the input is ASCII. Also added a few more difficult cases, and corrected some of the errors in the code. Wyang (talk) 10:18, 2 February 2017 (UTC)

@Wyang: Even |word=ယောက်ကျား doesn't seem to work at ယောက်ျား. It generates the wrong MLCTS and ALA-LC. This is such a bizarre spelling that basically breaks all the rules of Burmese spelling. Another suggestion: without changing the respelling input, maybe the respelling display could replace "+" with the voiced letter, i.e. when ဆီ+ပုံး is input, ဆီဗုံး would be displayed. The same trick could also replace -ဘ with ဖ in the display. —Aɴɢʀ (talk) 15:00, 2 February 2017 (UTC)

Oh, one more thing: wouldn't it be more usual to have all the testcases at Module:my-pron/testcases instead of on the template documentation page? —Aɴɢʀ (talk) 15:17, 2 February 2017 (UTC)

@Angr Thanks, I added the +/-... conversions. The +သ respelling is still visible though, since there is no letter for its voiced equivalent. I felt that putting the examples here helps template users choose the correct template parameters, and also because I'm not very familiar with how to produce testcases in different formats. Wyang (talk) 05:53, 3 February 2017 (UTC)

@Atitarev Thanks for the encouragement Anatoli! I find that Burmese is much more phonetic than Thai, but the relative lack of resources is indeed a disadvantage. Hopefully this template can make entry creation and maintenance a lot easier for Burmese. Wyang (talk) 11:53, 1 February 2017 (UTC)

I also find Thai reading rules much harder than Burmese, but to some extent I suppose it depends on what you're used to. —Aɴɢʀ (talk) 11:58, 1 February 2017 (UTC)

I guess my passive interest in Thai is much older. It's good that Burmese is not not as hard as it seems. I also find Thai pronunciation much clearer than Burmese but it may also to do with the actual resources I use or are available, have better (clearer) speakers. I seem to be almost OK with Thai tones, at least with words I know. I still have to tackle Burmese tones. :) --Anatoli T. ^{(обсудить}/^вклад) 12:05, 1 February 2017 (UTC)

Burmese tones are harder to distinguish for me too, especially the high and creaky tones. Wyang (talk) 10:26, 2 February 2017 (UTC)

Yeah, these tones are hard. You are probably OK with the Vietnamese pronunciation. Vietnamese tones are even harder (for me), especially "low falling constricted" and the variety of vowels is daunting. I gave up with the intention to master this later - but I hated the recordings I had at hand then - too fast and no tone drills like I used when learning Mandarin - slow, clear, tones combinations, plenty of examples. Teachers and book publishers (with recordings) of tonal languages should have more efficient methods. Current Chinese Mandarin resources are just great, e.g. New Practical Chinese Reader is all you need to become good with Mandarin tones. Even electronic dictionaries have decent text-to-speech engines, which can be used. Well, these nations are not as rich as China. ;) I have to admit that with all tonal languages I worked with, I have only some confidence with the Mandarin pronunciation in terms of reading out loud. Getting there with Thai - the tones makes sense but I often forget tone rules and exceptions without the translit. All others - I know a bit, haven't got enough exposure to the accents and pronunciation yet, even when I managed to read some simple texts in Vietnamese (with horrible tones I can imagine) - I realised that I'm not learning the right accent and stopped, as I said before. We should have Babel tables for pronunciation, he-he. --Anatoli T. ^{(обсудить}/^вклад) 11:00, 2 February 2017 (UTC)

Haha, this isn't surprising if we look at the diagrams in Vietnamese phonology#Tone. I think you are pretty good with Mandarin tones, but Mandarin tones are relatively simple (contour-based) compared to Burmese and Vietnamese, which also involve registers (creaky voice, breathy voice, etc.). I have trouble with these tones as well. Wyang (talk) 11:24, 2 February 2017 (UTC)

@Wyang, Angr, Atitarev Here is my point of view on these languages: I don't necessarily think that Burmese has the hardest pronunciation rules; they are simply unpredictable rules. Thai, on the other hand, has lots of silent letters, added vowels, and a way more complicated tone than Burmese. Tibetan may also be irregular, but just as irregular as Burmese. Just don't criticize their spellings—it's part of their culture anyway – AWESOME meeos ！ * (chōmtī hao /t͡ɕoːm˩˧.tiː˩˧ haw˦˥/) 04:58, 2 February 2017 (UTC)

Testing

I'm testing the template on two pages where respelling is unnecessary: ပျား and ပြား. So far, so good. —Aɴɢʀ (talk) 15:32, 2 February 2017 (UTC)

ဖြည်းဖြည်း

@Wyang What is the "2" doing in the ဖြည်းဖြည်း example? That isn't explained in the documentation. —Aɴɢʀ (talk) 07:28, 3 February 2017 (UTC)

Sorry I omitted this. It is used after ည် to indicate that it is pronounced /è/. Wyang (talk) 11:12, 3 February 2017 (UTC)

OK, thanks! That's another one where the respelling displayed could use the "normal" spelling (i.e. ီ for /ì/, ေ for /è/, and ယ် for /ɛ̀/). —Aɴɢʀ (talk) 12:59, 3 February 2017 (UTC)

@Wyang, Awesomemeeos: I really think ည် shouldn't appear in phonetic respellings at all; there are simply too many ways it can be pronounced. I don't think people are likely to guess what the "2" and "3" mean, and anyway, adding the numerals before the ့ and း characters is difficult. I'd much rather that the display only shows ီ for /ì/, ေ for /è/, ယ် for /ɛ̀/; ိ for /ḭ/, ေ့ for /ḛ/, ယ့် for /ɛ̰/; and ီး for /í/, ေး for /é/, and ဲ for /ɛ́/. —Aɴɢʀ (talk) 18:48, 19 February 2017 (UTC)

Agreed. If Awesomemeeos agrees too, I will change the display to using a reduced number of representations for different vowels. Wyang (talk) 10:26, 20 February 2017 (UTC)

Stacked consonants?

@Wyang It seems the module can't deal with stacked consonants. At any rate, it fails on ကမ္ဘာ, both with and without a respelling parameter. —Aɴɢʀ (talk) 23:37, 3 February 2017 (UTC)

It can't deal with the "+" sign at ကမ်းခြေ, either. —Aɴɢʀ (talk) 23:42, 3 February 2017 (UTC)

@Angr I created Module:my-pron/testcases to make it easier to track these. I think both cases are now fixed. Wyang (talk) 04:42, 4 February 2017 (UTC)

Tall ā

@Wyang: Could you arrange it so that +ကာ displays as ဂါ (not ဂာ) and +တာ and +ထာ display as ဒါ (not ဒာ)? Likewise +ကော, +တော/+ထော, +ကော်, +တော်/+ထော် should display as ဂေါ, ဒေါ, ဂေါ်, ဒေါ်. —Aɴɢʀ (talk) 14:19, 4 February 2017 (UTC)

@Angr Fixed (e.g. တံတား). Also thanks a lot for adding failed cases to Module:my-pron/testcases - they are all fixed as well. Wyang (talk) 23:08, 4 February 2017 (UTC)

@Wyang: Thanks for your help! နေပြည်တော် is still using the "short" version instead of the "tall" version. —Aɴɢʀ (talk) 16:18, 5 February 2017 (UTC)

@Angr My bad, sorry! Fixed it now. Wyang (talk) 08:24, 6 February 2017 (UTC)

@Wyang: I hadn't thought of this before, but the ā needs to be tall even in cases where the consonant letter hasn't changed due to voicing. For example, the respell of အဏ္ဏဝါ should be "အန်န'ဝါ" (with tall ā after ဝ; not "အန်န'ဝာ"), just as it is in the page name. The letters that always take tall ā are ခ ဂ င ဒ ပ ဝ. —Aɴɢʀ (talk) 09:12, 13 February 2017 (UTC)

@Angr Sure thing, I changed it now. Wyang (talk) 09:36, 13 February 2017 (UTC)

သွင်သွင်

The plus sign for voicing syllables doesn't seem to be working for သွင်သွင် (swangswang). --Lo Ximiendo (talk) 11:27, 10 February 2017 (UTC)

I added it to Module:my-pron/testcases. —Aɴɢʀ (talk) 14:40, 10 February 2017 (UTC)

Add spaces

Hi @Wyang, I reckon you should bring back the spaces when rendering the romanisations as well as the IPA. i.e. when when doing phrases where there is more than one word, spaces occur to show the word boundaries. Ta – AWESOME meeos ！ * (chōmtī hao /t͡ɕoːm˩˧.tiː˩˧ haw˦˥/) 04:42, 18 February 2017 (UTC)

Sure, added. Wyang (talk) 04:58, 18 February 2017 (UTC)

ကြိယာ

@Wyang, Angr Hi, what trick should I use to make က to be read with the inherent vowel, not part of the cluster? --Anatoli T. ^{(обсудить}/^вклад) 03:55, 22 February 2017 (UTC)

Hi Anatoli. {{my-IPA|က'*ရိယာ}} should work there. Wyang (talk) 06:24, 22 February 2017 (UTC)

my-IPA

Is it better idea to move my-IPA to my-pron? --Octahedron80 (talk) 06:15, 7 June 2017 (UTC)

No. All (or almost all?) of the automatically generated pronunciation templates are named xx-IPA. —Aɴɢʀ (talk) 06:37, 7 June 2017 (UTC)

Yes, all new IPA templates use or should use xx-IPA. --Anatoli T. ^{(обсудить}/^вклад) 06:43, 7 June 2017 (UTC)

My related templates are th-pron shn-pron and khb-pron that are the same title as modules. Does it ever have policy/consensus about this? --Octahedron80 (talk) 06:50, 7 June 2017 (UTC)

Adapt my-IPA for Rakhine / Arakanese language?

@Wyang, Angr Hi there! I'm wondering if you would be able to adapt this template to the w:Arakanese language (ISO code rki). There are general rules at play in converting between Burmese and Arakanese pronunciations (i.e., the Burmese rhymes generally map to corresponding Rakhine rhymes), which are already laid out in the article's phonology section (I could also supply additional guidance, as I have access to native Arakanese speakers). Perhaps this could be accomodated similar to how Template:vi-pron handles multiple Vietnamese dialects. Just wanted to hear your thoughts. Thanks for your time.

Vowel conversion (simple) [my → rki]:

လေ (air) : /lè/ → /lì/

သွေး (blood) : /θwé/ → /θwí/

Merged vowels (simple) [my → rki]:

စစ် (war) : /sɪʔ/ → /saɪʔ/

ပစ် (throw) : /tɪʔ/ → taɪʔ/

Nasal finals (more complex) [my → rki]

လည်ပင်း (neck) : /lɛ̀bɪ́ɴ/ → /làɪɴbɪ́ɴ/

စဉ်းစား (to think) /sɪ́ɴzá/ → /sáɪɴzá/

This article also provides additional rhyme mapping between Written Burmese and Arakanese (pp. 7-8). --Hintha (talk) 05:15, 27 June 2018 (UTC)

@Hintha It's an interesting suggestion. I remember reading w:Arakanese language and thinking whether it is feasible and worthwhile to automate this. The feasibility side seems clear, although it is apparent that an additional parameter |rki= or |ara= would be needed to supply a respelling for Arakanese, in cases where the Arakanese pronunciation is unpredictable from the respelling for Standard Burmese. There are some irregularities that may complicate this, and automatically generated Arakanese pronunciations may be incorrect for certain words that correspond irregularly to Standard Burmese (respelling) or the Written Burmese form (e.g. pg 14 in Okell). What is your opinion on this? Should the display of the Arakanese pronunciation be suppressed by default, and allowed only after it is checked? Are there dictionaries or other resources where we could verify Arakanese pronunciations of individual words? Wyang (talk) 06:09, 27 June 2018 (UTC)

Module errors

Three entries have the module error "Initial data not found" in {{my-IPA}}: သျှင် (hsyang), အသျှင် (a.hsyang), ဗြိတိသျှ (bri.ti.hsya.). Could someone familiar with Burmese fix this or remove the templates from the entries? — Eru·tuon 21:56, 4 April 2019 (UTC)

@Erutuon There is an ongoing discussion at Module talk:my-pron. If it's a big issue and is not resolved soon, please replace with {{rfp|my|initial data missing at Module:my-pron}}. @Mahagaja, Hintha --Anatoli T. ^{(обсудить}/^вклад) 22:29, 4 April 2019 (UTC)

I guess I was being a bit too hasty. Thanks to everyone for fixing it. — Eru·tuon 20:45, 5 April 2019 (UTC)