User talk:Babr/Sandbox

Hindustani phonology notes

Hi @Sameerhameedy!

I saw your diff. I wasn't really sure what to make of it. FWIW, I did also update WT:Urdu transliteration and added a section on nasalisation, if you want to make any improvements to that, or compare it to your side of things? نعم البدل (talk) 13:12, 23 February 2024 (UTC)

@نعم البدل Hi sorry, I tagged you a bit too early I'm still reading a bunch of research on Hindustani phonology. I was trying to tell you one thing though: my research so far has concluded that the distinction made between अं अँ in Hindi is often orthographic not phonemic, and it should be possible to predict nasal assimilation nearly 100% of the time (with only a few English loanwords needing manual transliterations).

The very in-depth book from Ohala suggests:

• All nasal vowels assimilate with the voiced obstruent consonants (g j ḍ and b);
  • Based on this information I changed it so that پَنْجاب = panjāb but پَن٘جاب = pañjab.
• short nasal vowels assimilate with voiceless plosives (k ṭ p)

I just wanted to ask you a few questions before implementing this into ur-translit. Implementing these phonological rules would mean that اِن٘ک اِن٘ٹ اِن٘پ would return iṅk iṇṭ and imp. Since Ohala suggests that short nasal vowels tend to assimilate with these consonants. However, as hi-translit does allow short nasal vowels before these consonants, I wanted to confirm if you think always assimilating these short nasals is indeed correct? (It seems like Urdu —tends to— respell Hindi अँप to اَمپ anyway so perhaps we don't need the module to deal with that at all?)

So TLDR I'm asking if you think اَن٘ک and اَن٘ٹ always returning aṅk and aṇṭ is correct. It's worth noting that hi-translit only assimilates these consonants if they are aspirated; should we do that or assimilate these consonants even when unaspirated as what my research seems to suggest (even if that means a spall departure from Hindi)? I just wanted to know if the information "felt correct" to you. This is really the final piece of information I need to enable ur-translit to accurately predict nasalization in nearly all cases. - سَمِیر | Sameer (^{مشارکت‌ها} · ^بحث) 15:51, 23 February 2024 (UTC)

Also should اَن٘ڈ return aṇḍ or ãḍ? Research suggests the first but Hindi does the second (unless aspirated). سَمِیر | Sameer (^{مشارکت‌ها} · ^بحث) 15:54, 23 February 2024 (UTC)

@Sameerhameedy: I always lose my head trying to contemplate linguistic terms relating to nasalisation/assimilation, so I'll try and simplify this. IMO اَن٘ک (ãk) should return as aṅk, but اَن٘ٹ (ãṭ) and اَن٘ڈ (ãḍ) should return as anṭ/anḍ (as a normal n, not with a dot) because I don't believe Urdu speakers understand or pronounce the retroflex nasal – and I couldn't find any mention of it in Urdu phonology references. As for اِن٘پ, initially I was opposed to 'n' returning as 'm', but honestly I don't really think there's any issue with that. The tilde, imo, should be reserved for prolonged vowels or blending vowels (I'm not sure if that's what assimilation means).

I think what you're suggesting also goes with what I've written at WT:Urdu transliteration? نعم البدل (talk) 19:48, 23 February 2024 (UTC)

Sure, Urdu doesn't normally have the retroflex N, but it's phonologically impossible to not pronounce it in the nt and nd clusters. Try doing it at normal speed and see if you can do it. 178.120.9.49 20:18, 24 February 2024 (UTC)

@178.120.9.49:. I'm a Punjabi-Urdu speaker, Punjabi clearly has a retroflex nasal. Urdu doesn't and it takes quite some effort for Urdu speakers to pronounce it. Unless I've wrongly mapped the retroflex nasal, it's much further back than /n/. I know this is, however, different in Hindi, which does have a retroflex nasal. نعم البدل (talk) 07:07, 1 March 2024 (UTC)

@نعم البدل @Sameerhameedy 178.120.9.49 20:20, 24 February 2024 (UTC)