Is this a reasonable sound change?

[2+3 clusivity]

Hey all, any opinions on this sound change are appreciated.

/c/ ---> /ts**/
/j*/ ---> /dz**/

*palatal voiced stop

**laminal aleovelar



I assume it would not be difficult, but you never know.

[Torco]

you mean

c/ts/_(laminal alveolar), as in a c turns into a ts before a palatal alveolar?

I don't really understand.

[linguoboy]

Hey all, any opinions on this sound change are appreciated.

/c/ ---> /ts**/
/j*/ ---> /dz**/

It would be counterintuitive in palatal environments.

[2+3 clusivity]

no, simply: /c/ turns into /ts/


Edit: the original * was merely to show I wasn't using j's IPA value.

[Aurora Rossa]

Those changes seem pretty reasonable to me and I have used them in my own conlanging projects from time to time. Palatal stops often shift to affricates and many languages merge hushed and hissed sibilants, so it seems pretty natural to me. I would expect the resulting phonemes to retain palatalized allophones in palatal environments, however.

[Torco]

yeah, like that ts still being tS before high vowels or sumin'

[Travis B.]

Hey all, any opinions on this sound change are appreciated.

/c/ ---> /ts**/
/j*/ ---> /dz**/

It would be counterintuitive in palatal environments.

It has happened, though. Something like /c/ or /kʲ/ did indeed turn into something like /ts/ in parts of Romance, particularly Old French (and similar changes showed up elsewhere, e.g. Old Castilian); this sort of change is also reflected in Romance and Latinate loans into High German.

[Ambrisio]

In Sanskrit च was pronounced /c/ and it shifted to /ts/ in Marathi.

[installer_swan]

In Sanskrit च was pronounced /c/ and it shifted to /ts/ in Marathi.

Can you give an example for this?

[linguoboy]

It has happened, though. Something like /c/ or /kʲ/ did indeed turn into something like /ts/ in parts of Romance, particularly Old French (and similar changes showed up elsewhere, e.g. Old Castilian); this sort of change is also reflected in Romance and Latinate loans into High German.

What's the evidence for [c] as an intermediate stage here?

[Ser]

Well, Penny in the book you love hypothesized an intermediate [tʃ] state between [k] and [ts], simply because it's in the middle of those two and Italian has it.

Classical Latin caelum [kaɛlũ] > [kaɛlu] > [kɛlu] > [tʃɛlu](?) > [tsɛlu] > (palatization of [t] into [tʃ] when preceded by /i/ takes place here, so possible earlier [tʃ] > [ts] doesn't affect any of these new [tʃ]) > Old Spanish çielo [tsjɛlo]. (He doesn't give a specific place for the merger of final and [o] though, may have occurred earlier than this.)

[linguoboy]

Well, Penny in the book you love hypothesized an intermediate [tʃ] state between [k] and [ts], simply because it's in the middle of those two and Italian has it.

But that won't work for Gallo-Romance which has /k/ > [c] (> [ʧ] > [ʃ]) before /a/. And I don't need to conjecture [c] as an intermediate stage here because it's preserved in Rhaeto-Romance (e.g. Surselvan [caun] "dog").

[Ser]

Well, Penny in the book you love hypothesized an intermediate [tʃ] state between [k] and [ts], simply because it's in the middle of those two and Italian has it.

But that won't work for Gallo-Romance which has /k/ > [c] (> [ʧ] > [ʃ]) before /a/. And I don't need to conjecture [c] as an intermediate stage here because it's preserved in Rhaeto-Romance (e.g. Surselvan [caun] "dog").

But just like it's hypothetically speaking possible to place it in the evolution of Latin into Old Spanish, maybe it's a question of timing, where [k] maybe became [ʧ] and then became [ts] all before [k] became ([c]?) [ʧ] in Old French.

[linguoboy]

But just like it's hypothetically speaking possible to place it in the evolution of Latin into Old Spanish, maybe it's a question of timing, where [k] maybe became [ʧ] and then became [ts] all before [k] became ([c]?) [ʧ] in Old French.

Maybe it is, but if so let's see the evidence.

[Ser]

I've never heard of any evidence, Penny himself presented it as purely conjectural. (I'm arguing it's possible to think such a thing, though I don't believe myself in it.)

[2+3 clusivity]

In Sanskrit च was pronounced /c/ and it shifted to /ts/ in Marathi.

Can you give an example for this?

Okay, I've had some time and looked this up.

Apparently /c/ -> /ts/ is a common shift in NIA languages like Marathi. Many keep a /c/ allophone before front vowels and /j/. Marathi apparently separated into both /ts/ and /c/ independent of surrounding sounds via borrowing. The Indo-Aryan Languages, Masica, p. 94.

Compare /ca:r/ "four" and /tsa:ra:/ "fodder". I assume the "r"s in his text are alveolar flaps if you care. Id.

In several other NIA languages /tr/ -> /c/ providing a contrast with /ts/.

[Whimemsz]

Looking at that source on Google Books, Marathi isn't actually a good example, because as you say the situation has been complicated by dialect mixture and the author implies that Marathi retains a palatal allophone in the environment of front vowels. However, he does give some examples of languages where /c/ has shifted to /ts/ (without, judging from his discussion, retaining a palatal allophone): "Nepali, Eastern and Northern dialects of Bengali (Dacca, Maimansing, Rajshahi), the Lamani and Northwestern Marwari dialects of Rajasthani, the Kagani dialect of 'Northern Lahnda', Kumauni, and many West Pahari dialects (not, however, the Chamba dialects, Mandeali, Jaunsari, or Sirmauri)."

EDIT: An example from Nepali (taken from here and here) would be: pɔstsim "west" (cf. Sanskrit paścima). Also, tsinhɔ "mark/punctuation" (cf. Sanskrit cihna).

[Nortaneous]

didn't this happen in some slavic langs that had like 267942379423 different waves of palatalization

[tezcatlip0ca]

Related question.

Is it possible for /k/ to palatalize in two different ways, depending on the following vowel? For example, I have /ke kɨ ki ky kjV/ going to /tse tsɨ tʃi tʃy tsV/, and although /ɡ/ follows the same pattern before vowels, /ɡjV/ > /dʒV/. Also, /tjV djV θjV ðjV sjV zjV/ > /tsV dzV tjV djV ʃV ʒV/. But the real problem, /jV/ and /iˈV/ are contrastive, and after /tʃiˈV dʒiˈV/ simplifies to /ˈtʃV ˈdʒV/, remaining /iˈV/ goes to /iˈjV/, and a few sound changes later, intervocalic /j/ turns to /ʑ/. Anyway, is that too many sibilants?

[Whimemsz]

It's not unrealistic to have different types of palatalization in different environments, no (see the development of Slavic for a well-known real-world example). And no, that's not an unreasonably large number of sibilants.

[Travis B.]

It has happened, though. Something like /c/ or /kʲ/ did indeed turn into something like /ts/ in parts of Romance, particularly Old French (and similar changes showed up elsewhere, e.g. Old Castilian); this sort of change is also reflected in Romance and Latinate loans into High German.

What's the evidence for [c] as an intermediate stage here?

Not [c], but I have seen [kʲ] posited as an intermediate stage, and for this purpose the two might as well be equivalent.

[installer_swan]

In Sanskrit च was pronounced /c/ and it shifted to /ts/ in Marathi.

Can you give an example for this?

Okay, I've had some time and looked this up.

Apparently /c/ -> /ts/ is a common shift in NIA languages like Marathi. Many keep a /c/ allophone before front vowels and /j/. Marathi apparently separated into both /ts/ and /c/ independent of surrounding sounds via borrowing. The Indo-Aryan Languages, Masica, p. 94.

Compare /ca:r/ "four" and /tsa:ra:/ "fodder". I assume the "r"s in his text are alveolar flaps if you care. Id.

In several other NIA languages /tr/ -> /c/ providing a contrast with /ts/.

Very interesting, I have some passing familiarity with spoken Marathi, and can sort of parse written texts using cognates with Hindi/Sanskrit but I never realised that ths /ts/ is sometimes rendered as [च] in the orthography. Somehow, I'd always assumed it would be [त्स]. Thanks for pointing that out!

[Miekko]

Those changes seem pretty reasonable to me and I have used them in my own conlanging projects from time to time. Palatal stops often shift to affricates and many languages merge hushed and hissed sibilants, so it seems pretty natural to me. I would expect the resulting phonemes to retain palatalized allophones in palatal environments, however.

any component can be lost, though, so why the palatalized allophones need to be retained is a mystery.

[spats]

I have /ke kɨ ki ky kjV/ going to /tse tsɨ tʃi tʃy tsV/

Didn't/doesn't Chinese do something like this?