Sound Change Quickie Thread

[Tropylium]

I've got a /i e a u/ vowel system that I'm trying to collapse into /i a u/. The easy route is just to merge /e/ with /i/ and /a/ in various environments, but that's kinda boring. Any ideas as to something more interesting I could do?

i u > ɨ ʷɨ
e > i
a > o > u (maybe remaining in some positions)
ɨ > ə > a (maybe in some positions > i / u, e.g. jɨ wɨ > i u)

[Knit Tie]

If I have voiced, plain, aspirated and ejective series all contrasting, how plausible would it be to do this:

>voiced prenasalised merge with voiced and nasals into nasals, i.e. [mb b] > m

Then,

> aspirated shift into plain and plain shift into voiced

[Raphael]

What is, generally, more common - intervocalic consonant loss or loss of individual consonants from consonant cluster as those clusters simplify?

[Raphael]

If my language has /s/, /z/, /ʃ/, and /ʒ/, would it be plausible to have /s/ turn into /ʃ/ except before other consonants, turn /z/ into /ʒ/ except before other consonants, and leave "pre-existing" /ʃ/ and /ʒ/ unchanged?

[bradrn]

What is, generally, more common - intervocalic consonant loss or loss of individual consonants from consonant cluster as those clusters simplify?

I’d imagine loss of consonants from clusters, but I’m not sure. (Warning: I know pretty much nothing about diachronics, so this may be unreliable.)

If my language has /s/, /z/, /ʃ/, and /ʒ/, would it be plausible to have /s/ turn into /ʃ/ except before other consonants, turn /z/ into /ʒ/ except before other consonants, and leave "pre-existing" /ʃ/ and /ʒ/ unchanged?

Minor clarification: why would ‘pre-existing’ /ʃ/ and /ʒ/ be changed in the first place given the change you’re describing?

[Raphael]

If my language has /s/, /z/, /ʃ/, and /ʒ/, would it be plausible to have /s/ turn into /ʃ/ except before other consonants, turn /z/ into /ʒ/ except before other consonants, and leave "pre-existing" /ʃ/ and /ʒ/ unchanged?

Minor clarification: why would ‘pre-existing’ /ʃ/ and /ʒ/ be changed in the first place given the change you’re describing?

They wouldn't; I just wanted to clarify that I'm to some extent merging /s( and /z/ with /ʃ/ and /ʒ/.

[dɮ the phoneme]

So apparently VOT is correlated with the high of the following vowel. Stops have longer VOT before higher vowels. Has vowel hight ever been attested as the source of an aspiration contrast?

[Nortaneous]

So apparently VOT is correlated with the high of the following vowel. Stops have longer VOT before higher vowels. Has vowel hight ever been attested as the source of an aspiration contrast?

Lovegren 2011 claims that Mungbam high vowels condition aspiration on preceding plosives:

To take an example, let us consider a pair of words ... /ídi̋/ 'candle sap' and /íde̋/ 'bean'. In the first word, the consonant is apico-alveolar, prevoiced and aspirated. In the second word, and in all other contexts, it is lamino-dental and fully voiced. We might transcribe these phonetically as [íd̥ʰi̋] and [íd̪e̋], respectively.

In some Ryukyuan languages, however, the opposite happened, and unvoiced plosives became aspirated between nonhigh vowels, or in word-initial position preceding a nonhigh vowel. From Thorpe 1983:

[-son -cont -voice] = [+aspiration] / {## [+syll -high]}_[+syll -high]

[StrangerCoug]

When a language undergoes tonogenesis, is there a general trend for what happens to the tone following ejective consonants?

[Frislander]

I'm on Frislandic iteration number umpteen, and as it's been for the past few times it's an Indo-European language, with a few particular distinctive sound changes, including merging the *T and *D series while retaining the *Dh series as aspirates. Now I've decided to add an extra distinctive sound change, namely the realisation of *l being as a velar nasal. Now I should imagine that this is a possible sound change by going through a velar lateral stage, especially considering how common the velarisation of laterals is in Indo-European s a whole, and it apparently has been attested in Rennellese for a lateral to turn into a prenasalised velar stop, so I'm just checking that people would think this a natural sound change in IE.

When a language undergoes tonogenesis, is there a general trend for what happens to the tone following ejective consonants?

The data are sparse and inconclusive on this one, due to the fact that we don't appear to have any attested cases of languages with ejectives that have undergone a phonation-based tonal split like that seen in East Asia (in fact I'm not even sure phonation-based tonal splits are at all common outside of Asia). My instinct on this one is that like implosives they'll pattern with plain voiceless plosives when such things do occur, though I can also see a situation where an ejective leads to creaky voice on the following vowel and thereby lowers the pitch. But again, the data is too sparse to draw any firm conclusions. There might be something in the history of Oto-Manguean that might be suggestive but that would probably require much more complete understanding of the history of Oto-Manguean prosody than we currently have.

[Pabappa]

maybe look up Athabaskan languages .... sorry Im lazy so I didnt do it myself ..... but they have ejectives and tones and i think may have come from a toneless ancestor. i know that high tones in some Athabaskan languages correspond to low tones in others, which suggests you might be able to do it whichever way you want, though its possible that rather than two separate developments there was a single development and then a second conditioned change that swapped them around. its also poissible that onset consonants had nothing to do with the tonogenesis in Athabaskan, sincve its usually the coda that has the greatest effect.

an ejective *following* a vowel is almost sure to be correlated with high tone, as are voiceless consonants generally.

there might be some tonal languages with ejectives in Cushitic, Omotic, etc ..... hard to research how they got that way, though.

[bradrn]

but they have ejectives and tones and i think may have come from a toneless ancestor. i know that high tones in some Athabaskan languages correspond to low tones in others, which suggests you might be able to do it whichever way you want, though its possible that rather than two separate developments there was a single development and then a second conditioned change that swapped them around.

My (very incomplete) understanding of Athabaskan tone is that the ancestor was originally toneless (as you say), but allowed glottal consonants (ejectives, glottalised sonorants and /ʔ/) in syllable codas. However, this contrast was lost is descendant languages, with glottalisation on the syllable coda developing into tone on the vowel; in some languages, glottalisation developed into high tone, while in others, it developed into low tone. (Gordon and Ladefoged note that this is a particularly nice example of how creaky voice may be associated with either lower or higher fundamental frequency depending on the language.)

[Richard W]

I'm on Frislandic iteration number umpteen, and as it's been for the past few times it's an Indo-European language, with a few particular distinctive sound changes, including merging the *T and *D series while retaining the *Dh series as aspirates. Now I've decided to add an extra distinctive sound change, namely the realisation of *l being as a velar nasal. Now I should imagine that this is a possible sound change by going through a velar lateral stage, especially considering how common the velarisation of laterals is in Indo-European s a whole, and it apparently has been attested in Rennellese for a lateral to turn into a prenasalised velar stop, so I'm just checking that people would think this a natural sound change in IE.

Polish gets you l > w. Armenian gets you w > g, and w > gw syllable initially is common enough in Western Europe. I believe Tok Pisin (Germanic) gets you g > ^ŋg . I think Rennellese went r > ʁ > ɣ > ɡ > ŋɡ > ŋ. However, the nasalisation relies on being in an area where prenasalisation is a common concomitant of voicing.

In short, I think l > ŋ is unlikely for an IE language, but given the right neighbours, it could happen.

[Frislander]

In short, I think l > ŋ is unlikely for an IE language, but given the right neighbours, it could happen.

Or how about entirely on its own on an island for much of its earlier history?

[mae]

-

[StrangerCoug]

I have this phonology:

Vowels

Code: Select all

/i y u/ ⟨i ü u⟩
/e ø o/ ⟨e ö o⟩
/  ɛ a/ ⟨  ä a⟩

Consonants

Code: Select all

/  m   n             ŋ     / ⟨  m   n           ŋ     ⟩
/p b t d ts  dz   k  ɡ q  ʔ/ ⟨p b t d c  j   k  g q  ɂ⟩
/  ɓ   ɗ tsʼ      kʼ   qʼ  / ⟨  ḅ   ḍ cʼ     kʼ   qʼ  ⟩
/         s   z           h/ ⟨        s  z           h⟩
/  w   l      ɹ j          / ⟨  w   l    r y          ⟩

/ŋ/ shifts to [ɴ] before /q/.

Phonotactics
Syllabes are strict C(w/l/ɹ/j)V(C). A syllable cannot begin with two consecutive approximants, and the coda is morphophonemically restricted to ⫽p t k q m n ŋ s l ɹ⫽, but coda ⫽p t k⫽ become /b d ɡ/ before a voiced consonant. (/q/ becomes [ɢ~ʁ] before a voiced consonant.)

Front/back vowel harmony applies with "dark" ⟨a o u⟩ vs. "light" ⟨ä ö ü⟩; ⟨e i⟩ are neutral and transparent to vowel harmony. While most consonants are transparent to vowel harmony, /w/ is phonetically affected by it and is pronounced [ɥ] in words with light vowels; however, they are written with the same letter, meaning context is required to distinguish /we wi/ from /ɥe ɥi/ in the spelling. Vowel harmony does not apply across the boundary between free morphemes, but it does apply across the boundary between a bound morpheme and the free morpheme to which it is attached. Prefixes and proclitics are technically affected by umlaut, not vowel harmony, but the general rule remains the same.

Stress
Stress is consistently on the first vowel of the root and is not retracted by prefixes.

I want to have a descendent subbranch that gets rid of productive vowel harmony altogether. I imagined unconditional unrounding of /ø y/ to /e i/ and then chain-shifting /ɛ a/ to /æ ɑ/ since I don't see a quick-and-easy way to get phonemic /ɔ/. Is that a plausible way of doing it (or at least a step towards it)?

[Chengjiang]

That would work, yeah. Although I’m not sure what /ɔ/ has to deal with it.

——

What are some ways I could have a language gain a full series of labialized velars in a relatively short time without gaining other labialized consonants? (Except possibly labialized uvulars.) I know a lot of languages have a series of labialized dorsal consonants, but I’m having trouble finding good descriptions of sound changes producing this series. Conditional fortition of [w] is good for gaining one, possibly two such consonants, but it’s not so useful if, say, I want a language with /k g k’ x/ to gain phonemic labialized versions of all these. Is there much natlang precedent for, say, changes like [Co] > [Cʷa] to apply to dorsals while ignoring coronals and labials?

[Pabappa]

Maybe if you have vowel sequences, do things like /tuV nuV/ > /kʷ ŋʷ/. Then, dispose of /pʷ mʷ/ by shifting them back to ordinary labials. I've done similar things in conlangs.

[StrangerCoug]

Although I’m not sure what /ɔ/ has to deal with it.

Trying to jog my memory of what /ɔ/ was even there for, but I do know that I have a strong tendency not to have /ɛ/ without /ɔ/ (or in general only one tense vowel of a given height). That said, I must confess my awareness it's well-attested.

[alice]

If the voice distinction is lost in stops, how likely is that a short vowel would lengthen after one of them at the start of a word?