Sound Change Quickie Thread

[Porphyrogenitos]

Are there any cases where a /p/ > /t/ or /b/ > /d/ shift has been attested before, whether conditional or unconditional?

[Pabappa]

Ancient Greek had /pj/ > /pt/, possibly through /pč/ or something similar as an intermediate. This /pt/ could then feasibly contract to just /t/. There is also /kʷ/ > /t/ before front vowels, which may have gone through /kp/ and then the same change as above, though it was probably just /kʷ/ > /čʷ/ > /tʷ/. This change also had a voiced version, gʷ > d, but apparently only before /e/ and not /i/. There may have also been /bj/ > /bd/, but there isnt traditionally a /b/ in PIE.

One change I've used a lot in conlangs is /f v/ > /þ ð/, which Ive never seen before in a natlang, but these are very labial-heavy lanfuages and so it could be seen as a way to spread out the phonemes more. But if this change does exist in a natlang, it could easily be extended to stops, at least voiced stops, as an intermediate change (e.g. /b > v > ð > d/). /p/ > /fʷ/ > /f/ > /þ/ > /ṭ/ (dental t) > /t/ is another idea, only because I suspect p > f directly is likely to be conditional.

Another change Im doing is consonant cluster simplification. Of course a change like /pt/ > /tt/ would be commonplace but its application would be pretty limited. If you had a protolang that offered words like /plam/ and you wanted to get /tã/ you could have a rule that all consonants change to match the PoA of a following sonorant, and then delete the sonorant. Thats what Im doing in pre-Proto-Tapilula.

[vokzhen]

Labial to dental in a small cluster of Oceanic languages, before unrounded vowels via linguolabials.

By my understanding (from someone else on the board) some Polish dialects have or are in the process of innovating a similar change to Greece, where /pj bj mj/ fortify to [pc bɟ mɲ], to give another example of that stage happening. Sotho has *pj *bj to /pʃ bʒ/, but you're probably less likely to get stops from that.

[Nortaneous]

aren't there some other African languages that have pj bj > ptS bdZ? Kinyarwanda has stop+semivowel > stop+stop fortition in general: bw > bg, bj > bJ\

[StrangerCoug]

A sound change I've done before in past conlang tinkerings is /ɺ ɾ/ /ɾ/ (the /ɺ/ is supposed to be perceived as a lateral, not a rhotic, before the merger starts happening). Suppose, though, that it's getting on my conlang speakers' nerves that the two sounds are so similar—what are some possible dissimilation routes they could go?

[----]

Turn one of them into /d/.

[vokzhen]

Redistribute them, with a lateral appearing next to back or open vowels and a rhotic next to front or close. Fortition to /d/ (a combination of this and the open/close distinction turned into /d~l/ in some Bantu languages). Turn the central one into an approximant (esp intervocally or in the coda), which then opens up further changes (deletion, to /j/, to a voiced fricative). Pretty sure spontaneous nasalization isn't out of the realm of possibility.

[Qwynegold]

A sound change I've done before in past conlang tinkerings is /ɺ ɾ/ /ɾ/ (the /ɺ/ is supposed to be perceived as a lateral, not a rhotic, before the merger starts happening). Suppose, though, that it's getting on my conlang speakers' nerves that the two sounds are so similar—what are some possible dissimilation routes they could go?

Just turn /ɺ/ into /l/.

[Qwynegold]

I'm developing a lang from Proto-Circassian and I need to ask if a few of these changes are plausible before I go any further:
*qʷ h
*ʔ~qʼ h~∅
*ʔʷ~qʷʼ h~∅

I would expect these to have /X(_w)/ as an intermediary, though that would make the rest of your sound changes implausible.

[Haplogy]

How likely is χr→ʁ/_V ?
And ɹχ→ʁ/V_ ?

[Arzena]

How likely is χr→ʁ/_V ?
And ɹχ→ʁ/V_ ?

It's not a far stretch at all: /χr/, /rχ/ > [χʁ], [ʁχ] > [ʁ]. You're moving the rhotic to a matching uvular POA and then deleting the fricative element, leaving the trill.

[vokzhen]

Or simply voicing adjacent to a particular sonorant, and deleting the sonorant Strikes me as very similar to something like tn > dn > d.

[KathTheDragon]

What can I do with /z/ that isn't shifting it to /s/ or /r/?

[sangi39]

What can I do with /z/ that isn't shifting it to /s/ or /r/?

/z/ > [ʑ] before /i e j/, which could then either become [j] before /i e/ and possible shift backwards to [ɣ] (following the [ʃ] > [x] shift of Spanish?) before /a u o/, then becoming [ɰ], dropping out before /a/ and shifting to [w] before /u o/. Original /z/ could then be retained before original /u o a/ that were not preceded by /j/, and this /z/ could undergo it's own changes, like shifting to [r] or [s].

It could become [dz] after nasals while shifting to [r] between vowels. If nasals drop out syllable-finally then [dz] could become a phoneme (and if /z/ appears as the initial sound of a suffix, then you could end up with some roots that would take the [r]-initial version (if the root originally ended in a vowel) while others would take the [dz]-initial version (if the root originally ended in a nasal)).

[Daedolon]

How about this?
T = Voiceless plosive
D = Voiced plosive
N = Nasal
rT lT > ʈ
Tr Tl > ʈ

rD lD > ɖ
Dr Dl > ɖ

rN lN > ɳ
Nr Nl > ɳ

rts > ʈs > ʈʂ
rdz > ɖz > ɖʐ

[sangi39]

How about this?
T = Voiceless plosive
D = Voiced plosive
N = Nasal
rT lT > ʈ
Tr Tl > ʈ

rD lD > ɖ
Dr Dl > ɖ

rN lN > ɳ
Nr Nl > ɳ

rts > ʈs > ʈʂ
rdz > ɖz > ɖʐ

I think Lhasa Tibetan has the Tr > ʈ-esque set of changes, and is it Swedish or Norwegian that has rT > ʈ?

/r/ and /l/ could merge before these changes occur, so that helps lT > ʈ by means of an intermediate rT.

[KathTheDragon]

Thanks, Sangi, but what could I do to specifically word-final /z/ (that absolutely isn't shifting to /s/ or /r/)?

[sangi39]

Thanks, Sangi, but what could I do to specifically word-final /z/ (that absolutely isn't shifting to /s/ or /r/)?

Palatalisation could still be an option if it were progressive rather than regressive (IIRC, Early Old French had both?). If you have Vj diphthongs, then the same /z/ > [ʑ] > [ɣ] > [ɰ] > [0]~[w] change could occur.

I wonder, as well, if you might be able to have it shift to [r], but then have that become [ʀ] > [ʁ] and take it from there, without having it merge with any original instances of /r/ (which could, for example, shift to [ɾ] or [ɹ] first).

It might be able to shift to [ð̠]. If something like [ʒ] shifted forward to [z], then this [ð̠] could become [ð] before moving on to [d] or [v]. Saying that, though, I'm not sure you'd necessarily need anything to cause that movement forward, becoming a non-sibilant fricative might just be enough on it's own.

Those last two changes might work in non-final position as well.

[KathTheDragon]

Thanks!

[Zaarin]

What can one do with /kʷ/ besides turn it to /k/ or /p/?

[KathTheDragon]

Greek turned it into /t/ before front vowels (filling the gap left by /t/ becoming /s/ before front vowels)

[Zaarin]

Greek turned it into /t/ before front vowels (filling the gap left by /t/ becoming /s/ before front vowels)

That could indeed produce interesting results. I think I'll run with that, thanks.

[TheCommissar]

I want to make a daughter language to Proto-Germanic, and I want to do something with the labiovelars. Is it plausible to have them become palatals along the lines of kʷ > t͡ʃ, gʷ > d͡ʒ, xʷ > ʃ ?

[GreenBowTie]

maybe before palatals/front vowels, but that seems really unlikely to happen unconditionally

[8Deer]

I want to make a daughter language to Proto-Germanic, and I want to do something with the labiovelars. Is it plausible to have them become palatals along the lines of kʷ > t͡ʃ, gʷ > d͡ʒ, xʷ > ʃ ?

Unconditionally, no. And even before front vowels it seems unlikely without them first shifting to plain velars. But I suppose you could have plain velars and labiovelars merge as palatal before front vowels.