Sound Change Quickie Thread

[Pabappa]

Oh, and I think the voiced velar fricative [ɣ] was encoded differently from Greek gamma because the IPA symbol is supposed to have a loop with a clear hole (or "eye" as typography people say), whereas Greek gamma very often has a tight loop with an inexistent eye.

Yeah, Unicode is inconsistent on whether it pulls from the existing characters or uses new ones for IPA. For example, the IPA phi (ɸ) is a separate character from the Greek phi, but the IPA theta (θ) simply reuses the Greek character. It's mildly frustrating because it means I can't type out IPA by just going to the "IPA Extensions" chart on Babelmap, but I dont type in IPA that often. (And when I do, I dont bother with the /g/ thing.)

[Kuchigakatai]

Yeah, Unicode is inconsistent on whether it pulls from the existing characters or uses new ones for IPA. For example, the IPA phi (ɸ) is a separate character from the Greek phi, but the IPA theta (θ) simply reuses the Greek character. It's mildly frustrating because it means I can't type out IPA by just going to the "IPA Extensions" chart on Babelmap, but I dont type in IPA that often. (And when I do, I dont bother with the /g/ thing.)

Last thing I heard, they have now relented on the ⟨ɡ⟩ thing because practically nobody cared anyway.

[Nortaneous]

Allophonic [ꭓ] or [ʁ] for /q/ is common

I'm guessing it's just a mistake, but you used U+AB53 for chi there. IPA [X] is supposed to simply be U+3C7 GREEK SMALL LETTER CHI, very commonly covered in fonts. U+AB53 is a Latin Extended-E symbol specific to German dialectology to represent palatal [ç], and it is very rarely covered in fonts. (I can't actually see your small chi on default latest Ubuntu.)

It's not - the default font for newer versions of phpBB has x χ as homoglyphs, so Latin chi is necessary to work around that.

Since a homoglyph of Latin <x> is a valid style of Greek chi from the perspective of general-purpose font design, the IPA symbol should be represented by Latin chi in general once font support catches up.

Last thing I heard, they have now relented on the ⟨ɡ⟩ thing because practically nobody cared anyway.

For databases, PHOIBLE uses the single-story g, but eurasianphonology.info uses ASCII g.

One of the reasons for the single-story g thing was that they used to distinguish it from double-story g. They got rid of that a long time ago, although unfortunately they still distinguish <a> from <ɑ>. (LaPolla's grammar of Qiang transcribes the low vowels as /a ɑ/, and it's pretty annoying to try to accurately transcribe the text.)

[bradrn]

Oh, and I think the voiced velar fricative [ɣ] was encoded differently from Greek gamma because the IPA symbol is supposed to have a loop with a clear hole (or "eye" as typography people say), whereas Greek gamma very often has a tight loop with an inexistent eye.

Yes, I believe this is correct. Also capital Greek gamma is ⟨Γ⟩ and capital Latin gamma is ⟨Ɣ⟩.

[bradrn]

Allophonic [ꭓ] or [ʁ] for /q/ is common

I'm guessing it's just a mistake, but you used U+AB53 for chi there. IPA [X] is supposed to simply be U+3C7 GREEK SMALL LETTER CHI, very commonly covered in fonts. U+AB53 is a Latin Extended-E symbol specific to German dialectology to represent palatal [ç], and it is very rarely covered in fonts. (I can't actually see your small chi on default latest Ubuntu.)

It's not - the default font for newer versions of phpBB has x χ as homoglyphs, so Latin chi is necessary to work around that.

I don’t know if you’re aware, but zompist eventually changed the forum font because of this exact problem (plus others).

(This was supposed to be in the same post as the previous one, but I only saw your comment after I pressed ‘submit’, and I didn’t want to go back and edit this in…)

[Kuchigakatai]

It's not - the default font for newer versions of phpBB has x χ as homoglyphs, so Latin chi is necessary to work around that.

Since a homoglyph of Latin <x> is a valid style of Greek chi from the perspective of general-purpose font design, the IPA symbol should be represented by Latin chi in general once font support catches up.

I agree with the sentiment, in fact, back in the days of Microsoft Trebuchet (the horrible font that phpBB selects by default and is responsible for that Greek chi mess), I used to simply mix X-SAMPA inside. I'd transcribe "joder" in Spain as [Xoˈðeɾ].

The problem is that U+AB53 (Latin small chi) is rarely supported, so all I saw in your post were squares with the letters "AB53" instead (because that's Ubuntu Linux's fallback font). So I checked what letter that was supposed to be on a Unicode analyzer. On Windows, I imagine many would see plain squares (?).

For databases, PHOIBLE uses the single-story g, but eurasianphonology.info uses ASCII g.

Lovely.

One of the reasons for the single-story g thing was that they used to distinguish it from double-story g. They got rid of that a long time ago, although unfortunately they still distinguish <a> from <ɑ>. (LaPolla's grammar of Qiang transcribes the low vowels as /a ɑ/, and it's pretty annoying to try to accurately transcribe the text.)

I know typography is for nerds that should consider getting a life, but I'm still not used to the terrible typography you often see in Chinese dictionaries. Apparently many Chinese people believe that pinyin has to be written with <ɑ> instead of <a> (I think it's an influence of their childhood books from elementary school), so in publications from China (and a few western ones, as in some books I have from Oxford University Press) I often find pinyin <ɑ> in a different font (larger or smaller than the other letters) or otherwise accents horribly placed for <ɑ> (like "zhɑˇn") while the other vowels are fine ("zhǔn").

[Nortaneous]

I don’t know if you’re aware, but zompist eventually changed the forum font because of this exact problem (plus others).

Oh - it's just not changed in the post textarea.

[Knit Tie]

Can I have voiceless nasals become /h/, while voiceless semivowels /j/ and /w/ become a glottal stop, realistically?

[Nortaneous]

Can I have voiceless nasals become /h/, while voiceless semivowels /j/ and /w/ become a glottal stop, realistically?

Probably not, but maybe. Some Maori dialects have h wh > ʔ ʔw. It would probably help to introduce a contrast between h h̃, and then h > ʔ (also in hj hw = j̥ ʍ) h̃ > h. But then you have to collapse all the voiceless nasals and remove the semivowels from the clusters with ʔ. Nuosu has ŋ̥ > h but preserves m̥ n̥.

[Knit Tie]

Natch, I'll skip the ejectives, then. Another question I have is can postalveolar, strongly palatalized affricates become true palatal stops?

[Nortaneous]

Natch, I'll skip the ejectives, then. Another question I have is can postalveolar, strongly palatalized affricates become true palatal stops?

Ufa Udmurt *dʑ > ɟʝ

[Tropylium]

Another question I have is can postalveolar, strongly palatalized affricates become true palatal stops?

Insofar as "true palatal stops" are a thing at all, yes.
– Hungarian gy, though phonologists still debate if it's better to call it /ɟ/ or /ɟʝ/ or /dʲ/, has *ntɕ as one source.
– Old Yukaghir *ntɕ (spelled ndsch) > modern Yukaghir ď.
– Eastern Khanty has *tɕ > ť.
I do not know if these latter two are closer to [tʲ dʲ] or [c ɟ], but in any case not affricates.

[Kuchigakatai]

Answering this post from early December last year:
http://verduria.org/viewtopic.php?f=3&t ... 297#p22299

How sure are we about the pronunciation of Middle Chinese? I mean, I know I asked this exact question before, but it was for one specific word, so im not repeating myself .... this time I mean, are we sure that there were ever voiced aspirates in Chinese to begin with? Specifically with that unpaired /bʰ/ we see in some reconstructions, exactly as in PIE, ... I wonder if the IE-ists and the Sinologists were borrowing ideas from each other. I should mention Im skeptical of the reconstruction of PIE as well.

Yeah... The idea that Late Middle Chinese had the series of stops Tʰ T Dʱ has been espoused since Karlgren's early reconstructions in the 1910s and 1920s of the sound of the 11th-century Guangyun rime tables, and I'm sure they're inspired off PIE. In fact, he did reconstruct the MOA series Tʰ T D Dʱ for the stops of Old Chinese, just like in the PIE of that time.

Tʰ T Dʱ is not very satisfying typologically, but it doesn't help that Wu and Xiang show all three of the Early Middle Chinese Tʰ T D as Tʰ T D (e.g. /kʰ k g/), Mandarin and generally Cantonese conditionally merge the Late Middle Chinese Dʱ into Tʰ before the old level (平) tone and into T before the other tones (top/rising 上, departing 去, entering 入), and apparently in a lot of the remaining dialects Dʱ simply becomes Tʰ.

EMC Tʰ T D
LMC Tʰ T Dʱ
--> Wu & Xiang Tʰ T D
--> Mandarin & Cantonese Tʰ T Tʰ/T
--> Hakka & Gan Tʰ T Tʰ

(Wu & Xiang may or may not be sisters or dialectal variants of LMC? Also, Min Chinese is a sister branch of EMC, and while the nature of Proto-Min's stops is unknown, it doesn't show merges of EMC's Tʰ T D, rather, it shows more distinctions, such as a "softened" series corresponding to EMC T or D...)

It seems... somewhat reasonable to think LMC had Tʰ T Dʱ? Is there anything we could apply from Grimm's law though, from the likes of *drew-o-m > *trewan 'tree' and *gōl-u-s > *koluz 'cool'? Fortition is often dirty business.

[Nortaneous]

d > dʱ > tʱ > tʰ is an areally common sound change - in Sikkimese (Denjongke) there's a Tʱ series that can appear word-initially and becomes D in compounds. the other series are Tʰ T D, but initial /D/ is often realized as [ND] or 'prevoiced' [DT]

comparing Sikkimese forms to WT, it looks like the Tʱ- series is from WT D- and the D- series is from CD-:
/da/ mda', /tʱa/ da
/gɛː/ rgal, /kʱɛː/ gas

[Richard W]

/Dʱ/ might simply represent voiced stop plus a predictably breathy vowel. That would match the development in SE Asia. The breathiness can then do all sorts of things - when plain voicing vanishes. It can change the vowels as in Khmer, it can modify the tones (it causes tonogenesis in some Khmu dialects), and it can even be interpreted as aspiration on the consonant.

[bradrn]

I have fricatives /f s ʃ h z ʒ/. What can I do to get rid of the voiced fricatives /z ʒ/?

(If it helps, the full phonology at this point is:

Code: Select all

m n   ŋ
  t   k ʔ
b d   g
f s ʃ   h
  z ʒ
w r j
  l

a ə i u
aː iː uː

)

[foxcatdog]

I have fricatives /f s ʃ h z ʒ/. What can I do to get rid of the voiced fricatives /z ʒ/?

(If it helps, the full phonology at this point is:

Code: Select all

m n   ŋ
  t   k ʔ
b d   g
f s ʃ   h
  z ʒ
w r j
  l

a ə i u
aː iː uː

)

Weakening to approximates seems like a good idea.

[bradrn]

I have fricatives /f s ʃ h z ʒ/. What can I do to get rid of the voiced fricatives /z ʒ/?

(If it helps, the full phonology at this point is:

Code: Select all

m n   ŋ
  t   k ʔ
b d   g
f s ʃ   h
  z ʒ
w r j
  l

a ə i u
aː iː uː

)

Weakening to approximates seems like a good idea.

So /z ʒ/ → /r j/ or /l j/? (Contrary to what my phoneme list above implied, /r/ is a trill rather than an approximant — sorry for the error!)

[foxcatdog]

I have fricatives /f s ʃ h z ʒ/. What can I do to get rid of the voiced fricatives /z ʒ/?

(If it helps, the full phonology at this point is:

Code: Select all

m n   ŋ
  t   k ʔ
b d   g
f s ʃ   h
  z ʒ
w r j
  l

a ə i u
aː iː uː

)

Weakening to approximates seems like a good idea.

So /z ʒ/ → /r j/ or /l j/? (Contrary to what my phoneme list above implied, /r/ is a trill rather than an approximant — sorry for the error!)

z to a trill r is attested in latin (rhotacism as they like to call it). Or you could have it shift to approximate r giving 2 rhotics if you so like its up to you.

[bradrn]

Weakening to approximates seems like a good idea.

So /z ʒ/ → /r j/ or /l j/? (Contrary to what my phoneme list above implied, /r/ is a trill rather than an approximant — sorry for the error!)

z to a trill r is attested in latin (rhotacism as they like to call it). Or you could have it shift to approximate r giving 2 rhotics if you so like its up to you.

Thanks for the recommendation! I think I’ll do /z ʒ/ → /r j/ then, since I don’t particularly like /ɹ/.