Sound Change Quickie Thread

[Kuchigakatai]

How common is phonemic /d͡z/ without phonemic /z/?

Proto-Semitic is usually reconstructed like that these days, with a /ts dz tsʼ/ series (parallel to /k g kʼ/ and /θ ð θʼ/) and a single /s/ sibilant. The more traditional reconstruction had /s z sʼ/ and /ʃ/ instead. I'm not familiar with the literature at all to know the arguments about why.

[Travis B.]

How common is phonemic /d͡z/ without phonemic /z/?

Proto-Semitic is usually reconstructed like that these days, with a /ts dz tsʼ/ series (parallel to /k g kʼ/ and /θ ð θʼ/) and a single /s/ sibilant. The more traditional reconstruction had /s z sʼ/ and /ʃ/ instead. I'm not familiar with the literature at all to know the arguments about why.

The reason for reconstructing Proto-Semitic the way that it is now is that what many often now call /ts dz/, traditionally /s z/, have evidence from loans from Semitic of being affricates, while there is no real evidence for that what was traditionally called /ʃ/ was really postalveolar when taking Semitic as a whole into account (there is little real evidence that it wasn't postalveolar, though, either; it likely was postalveolar in some portions of early Semitic and non-postalveolar in others). Note though that there is general agreement that "/sʼ/" was /tsʼ/, based on evidence from both Hebrew and Ethio-Semitic.

[anteallach]

How common is phonemic /d͡z/ without phonemic /z/?

Proto-Semitic is usually reconstructed like that these days, with a /ts dz tsʼ/ series (parallel to /k g kʼ/ and /θ ð θʼ/) and a single /s/ sibilant. The more traditional reconstruction had /s z sʼ/ and /ʃ/ instead. I'm not familiar with the literature at all to know the arguments about why.

The reason for reconstructing Proto-Semitic the way that it is now is that what many often now call /ts dz/, traditionally /s z/, have evidence from loans from Semitic of being affricates, while there is no real evidence for that what was traditionally called /ʃ/ was really postalveolar when taking Semitic as a whole into account (there is little real evidence that it wasn't postalveolar, though, either; it likely was postalveolar in some portions of early Semitic and non-postalveolar in others). Note though that there is general agreement that "/sʼ/" was /tsʼ/, based on evidence from both Hebrew and Ethio-Semitic.

Sometimes in languages with only one sibilant fricative (or only one PoA for sibilant fricatives) it can be somewhat retracted and sound a bit /ʃ/-like to speakers of languages with an /s/-/ʃ/ contrast. E.g. Greek and Castilian Spanish, both of which also have a dental non-sibilant fricative, as did reconstructed Proto-Semitic.

As to the original question, UPSID has the following languages listed with /dz/ but no /z/: Chamorro, Cherokee, Dahalo, Hadza, Kashmiri, Kwakw'ala, Lelemi, Lushootseed, Nepali, Ocaina, Resigaro, Sandawe, Telugu, Tsimshian, Yareba. Some of these analyses are probably wrong or at least controversial, of course.

[Zju]

Given bog-standard T D and Tː (mostly) intervocally, how plausible is the shift T Tː > T Tʰ, instead of the more common T Tː > D T or T Tː > T?

[WeepingElf]

Given bog-standard T D and Tː (mostly) intervocally, how plausible is the shift T Tː > T Tʰ, instead of the more common T Tː > D T or T Tː > T?

I see no problem with that.

[chris_notts]

Given bog-standard T D and Tː (mostly) intervocally, how plausible is the shift T Tː > T Tʰ, instead of the more common T Tː > D T or T Tː > T?

I see no problem with that.

It's possibly more likely than if the voiced stop slot weren't already occupied, I would guess.

[conlangernoob]

Sorry if this sounds dumb, but how does non-phonemic stress evolve?

[Zju]

There are various types of non phonemic stress. E.g. stress can gradually weaken and then cease to be phonemic, out of which you can get fixed stress (word initial or word final).

[conlangernoob]

What about one system of non-phonemic stress to evolving into another?

[chris_notts]

Sorry if this sounds dumb, but how does non-phonemic stress evolve?

There's no reason contrastive stress should be the default. You could equally well ask how words distinguished by stress placement evolved (and I think the answer is often merger of syllables or deletion of some peripheral vowels or syllables, without the previously regular stress then shifting to where it "should" be, or possibly also compounding of stressed words producing a different pattern to the default).

Also worth mentioning that not all languages have a strong word stress. Some have more of a phrasal accent, or even no obvious accent defined by constituency at all.

[chris_notts]

What about one system of non-phonemic stress to evolving into another?

I don't know if this is the general answer, but in language families that show diversity in fixed accent placement, I get the impression the cause is often due to language contact. I.e. those families where the daughter languages show a number of systems tend to be the ones surrounded by other language families showing lots of options.

I suspect accent placement is more borrowable than a lot of other phonological phenomena.

A good example of this might be Basque, which maintains a fairly distinct non-Spanish phonology, but where I've read that a fairly large group of speakers have adopted a more Spanish stress pattern compared to the native Basque one due to contact and bilingualism.

[conlangernoob]

What about one system of non-phonemic stress to evolving into another?

I don't know if this is the general answer, but in language families that show diversity in fixed accent placement, I get the impression the cause is often due to language contact. I.e. those families where the daughter languages show a number of systems tend to be the ones surrounded by other language families showing lots of options.

I suspect accent placement is more borrowable than a lot of other phonological phenomena.

A good example of this might be Basque, which maintains a fairly distinct non-Spanish phonology, but where I've read that a fairly large group of speakers have adopted a more Spanish stress pattern compared to the native Basque one due to contact and bilingualism.

Thanks!

[Creyeditor]

Sorry if this sounds dumb, but how does non-phonemic stress evolve?

One possible source is intonation. You have a strong-weak (or high-low) pattern at the end of a phrase and extend it to words (which can have effects on subsequent changes). Of course, the question is then how intonation comes about and I don't know if there is a good answer to that (yet).

[conlangernoob]

Two questions:
How do length distinctions evolve in vowels?
How do voiced distinctions evolve in consonants from a Proto-Language for which voicing is not phonemic?

[Zju]

How do voiced distinctions evolve in consonants from a Proto-Language for which voicing is not phonemic?

Vowels and sonorants are both voiced, so it follows from laziness linguistic economy that a single voiceless segment inbetween can get voiced. In particular, an obstruent assimilating to the voicing of two neighbouring vowels is a type of lenition.

How do length distinctions evolve in vowels?

Hoo boy. Take your pick:
- Some marginally more difficult to articulate consonant drops out of coda position, so compensatory lengthening is used to maintain minimal pairs: /tox/ → /toː/
- Vowels in open syllables tend to be slightly longer and that distinction can become phonemic: /tadam/ → /taːdam/ (cf. Germanic languages)
- Vowels in front of voiced segments tend to be slightly longer and that distinction can become phonemic: /dam/ → /daːm/ (cf. Proto Slavic)
- Diphthongs can monophthongise: /aw/ → /oː/
- Take a hint from BrE: further /ˈfɜː.ðə/
- Nasal vowels can lengthen in the process of losing their nasality (looking at you, Lithuanian!)
- Sometimes, entire syllables can fall off with compensatory lengthening: PU *päŋə → Finnish pää 'head'
- Whatever other transphonologisation process I couldn't think of.

[Rounin Ryuuji]

Two questions:
How do length distinctions evolve in vowels?

Length can evolve in a number of ways, notably:

• Monophthongisation, for example /ai ei eu oi ou/ > /aː iː oː eː uː/, or various other possible shifts;
• Regular coda consonant loss, as with /s/ at earlier stages of French, or /r/ in some variety of English (this often triggers a quality shift, too);
• Medial consonant deletion, with two of the same vowels next to each-other coalescing into a long vowel, or subsequent diphthong or vowel-sequence smoothing.
• Lengthening in open syllables followed by loss of final vowels (this is the origin of the use of orthographic silent -e to mark long vowels in many English words).

[chris_notts]

Two questions:
How do length distinctions evolve in vowels?

Length distinctions can evolve a number of ways. Some examples:

1. Merger of two vowels to create a long vowel, often either with a weak consonant (j, w, ...) or after a consonant has been lost to create a vowel sequence

e.g. VV -> V:, VjV -> V:

2. Compensatory lengthening if a segment elsewhere is lost or shortened

VCV# -> V:C#, VCCV -> V:CV, ...

3. Lengthening of open stressed or prominent syllables, followed by loss of coda from closed syllables creating a length contrast

stressed CV -> CV:
some stressed CVC -> CV (coda loss)

How do voiced distinctions evolve in consonants from a Proto-Language for which voicing is not phonemic?

Some ways it can happen:

If there are already multiple series distinguished by phonation or airstream then they can just shift, e.g. t -> d, tʰ -> t.

A length distinction can become a voice distinction, e.g. t -> d (especially intervocally), t: -> t. This can also happen due to minimum weight constraints on syllables. Some languages lengthen consonants after open syllables with short vowels, especially stressed syllables, so e.g. you have:

CV:CV
CVC:V (onset of next syllable also closes first syllable)
*CVCV (not allowed because first syllable has insufficient weight)

So then you can have a shift where stops after a short stressed initial syllable remain voiceless, whereas stops after a long voweled initial syllable become voiced, and the original vowel length distinction can be lost. I think this happened in some Australian languages.

Similarly, since intervocalic voicing is so common, loss of codas which protected a stop from voicing can produce a phonemic surface contrast. E.g.:

kata -> kada
kasta -> kasta -> kata

Such a contrast can then spread to initial/final position by loss of some initial/final vowels or by various other means.

[Zju]

If I'm not mistaken, there was even the case of pitch accent turning into vowel length in West Slavic languages.

[conlangernoob]

Four dumb questions: In general, how many sound changes does a language go through every, say, 1,000 years? Also, how much variation is there? Also, do some types of sound change occur more than others? Finally, how many years does it take between a sound change to begin and a sound change to finish?

Thanks,

Conlangernoob

[Travis B.]

Four dumb questions: In general, how many sound changes does a language go through every, say, 1,000 years? Also, how much variation is there? Also, do some types of sound change occur more than others? Finally, how many years does it take between a sound change to begin and a sound change to finish?

Thanks,

Conlangernoob

The variation can be quite considerable - contrast English and Old English with Icelandic and West Old Norse, for instance.