Sound Change Quickie Thread

[----]

A) kt>s ?

That could work. It could happen in just a few changes, too. How about kt > ks > s?
Hanoi Vietnamese had kr > kʂ > ʂ > s, which is a similar overall change.

[Melteor]

^^Wow, really? Could kt>xt>s also work? I don't know which I prefer.

[WeepingElf]

^^Wow, really? Could kt>xt>s also work? I don't know which I prefer.

I see no problems with that.

[Thry]

^^Wow, really? Could kt>xt>s also work? I don't know which I prefer.

I see no problems with that.

What about kt > th > ts > s ?

th = aspirated t

[Whimemsz]

that's fine too

[tezcatlip0ca]

Is /d/ > /ð/ [ð̞ˤ] > /ʕ/ > vowel retraction plausible? I was thinking that for a descendant of Dŕbatkovi I could have /d/ and syllable-final /t/ and /θ/ pharyngealize and cause all kinds of damage on the surrounding vowels, to eventually have a 3-vowel system with length and an overload of glottal stops in the daughterlang...

[----]

That's a weird mechanism for the overall change, but cross-linguistically voiced fricatives do crazy things. Perhaps they reduce to glottals but then undergo fortition to pharyngeals? I don't know if such a change is attested but it makes sense to me.

[ná'oolkiłí]

I don't think d > ð > ʕ is too weird. In one of my langs I have *d > ʕ: *d lenites to *ð and then merges with *ɣ, which has an allophone of [ʕ] before back vowels; vowels then move around to make /ʕ/. I kinda find it weird that *t and *θ become pharyngeal(ized), though. What are the exact details you're envisioning? Is it to make them more contrastive with another coronal series or something?

[tezcatlip0ca]

I don't think d > ð > ʕ is too weird. In one of my langs I have *d > ʕ: *d lenites to *ð and then merges with *ɣ, which has an allophone of [ʕ] before back vowels; vowels then move around to make /ʕ/. I kinda find it weird that *t and *θ become pharyngeal(ized), though. What are the exact details you're envisioning? Is it to make them more contrastive with another coronal series or something?

The details are that syllable final plosives are already lenited, so the /t~θ/ distinction is already neutralized in that environment and only surfaces in morphological forms where the consonant occurs before a vowel. Then, /θ/ and /ð/ pharyngealize to maximize contrast with /s z/. After that, /θ/ voices unconditionally, and the resulting [ð̞ˤ] debuccalizes into [ʕ]. The vowel system at this stage of the language is /aː ɛi e iː ɔu o uː/, and the front vowels centralize adjacent to /ʕ/: [ʕɒː ʕɜɨ ʕɘ ʕɨː, ɜɨʕ ɘʕ ɨʕ]. A few hundred years later, the pharyngeal is lost, and the central vowels are phonemes in their own right. Then, the vowel sequences develop into a rather irregular diphthong system (which I have yet to derive), which then simplifies into a Sanskrit-like /aː ə eː ɪ iː oː ʊ uː aɛ əu/, with glottal stops breaking up the remaining hiatuses.

[sangi39]

Could these sets of vowel changes occur within the same language?

*i *i: * u *u:
*e *e: *o *o:
*æ *æ: *ɑ *ɑ:

(1a) Raising of long vowels followed by (2a) diphthongisation in open syllables

*i: > ai
*e: > i: ( > ie)
*æ: > e: ( > ei)
*u: > au
*o: > u: ( > uo)
*ɑ: > o: ( > ou)

(1b) “Circular” shift of short vowels followed by (2b) lengthening in open syllables

*u > i ( > i:)
*i > e ( > e:)
*e > æ ( > æ:)
*æ > ɑ ( > ɑ:)
*ɑ > o ( > o:)
*o > u ( > u:)

(3) Phonemicisation of long vowels through consonant cluster simplification and shortening of geminate consonants

While I’m fairly certain, 1a, 2a, 2b and 3 are attested shifts, it’s 2a I’m more concerned about. I vaguely recall reading about a circular vowel shift but I can’t remember the language or the details. I’m also not sure how likely 1a and 1b are to happen in the same language. Perhaps it would be more realistic for 1b to occur after 1a rather than the two occurring simultaneously.

[Tropylium]

The difthongization pattern seems odd… you have (using the front vowels):
iː ai
eː ie
æː ei
with the ends ending up closing but the middle ending up opening. I'd swap the last two around. This would add up to a chronology such as:
1) Non-open long vowels develop an off-glide: iː eː ij ej
2) The close member dissimilates əj aj
3) In closed syllables ej iː
4) Raising æː eː
5) Lengthening of all monofthongs in open syllables
5) Development of an on-glide in overlong mid vowels: eːː:> ie

I'm not sure about the circular short vowel shift either. All the parts are attested, but in particular I can't recall a precedent for u i pushing i e. Usually the latter implies the existence of u o, and as you have long vowels around, that would then make it unlikely to end up with e o i u or anything of the sort in addition.

An occasionally attested circular shift is the following:
i ɪ̆ e
iː iˑ i
eː iː
e eˑ eː (or: æː eː, æ æː, ɪ̆ ɛ̆ æ)
But tying anything like this into the long vowel scheme you've go would take more time than I have right now…

[Gagylpus]

Hey folks, I have a question...

I'm working on a phonology right now with a four-way phonotion contrast in plosives (aspirated, voiceless, breathy voiced, modal voiced) and a two-way phonation contrast in nasals (breathy voiced, modal voiced) at the bilabial, dental, and velar places of articulation. The breathy vs modal distinction also exists for the liquid consonants (l, r). I'm planning on having the aspirated and breathy plosives change to voiceless and modal voiced fricative, respectively. What are some likely changes that can occur to the breathy sonorants to eliminate phonemic breathy voice, while still having an influence on the daughter languange? (i.e. not simply losing the distinction between breathy and modal voice in all contexts).

So far my only ideas are changing the breathy sonorants to voiceless sonorants, or introducing lengthening of the following vowel. I fear I don't know enough about the effects of different phonation types to come up with something more interesting.

[Nortaneous]

tone

see: punjabi

[Lyra]

Is this feasible?

ʔ/ŋ/_#

~Lyra

[Thry]

Is this feasible?

ʔ/ŋ/_#

~Lyra

Maybe including a nasalization as a middle step?

-aʔ > -ã > -aŋ

I'm not sure though, as I'm unfamiliar with glottal stops. Second opinions are welcome.

[sangi39]

Would these changes make sense between vowels?


pʰr > ɸr > hr > r˳ > r
pr > br > br > :br > :r
br > βr > βʙ > β˕: > v:
tʰr > θr > hr > r˳ > r
tr > dr > dr > :dr > :r
dr > ðr > ðr > ɹ: > r:
kʰr > xr > hr > r˳ > r
kr > gr > gr > :gr > :r
gr > ɣr > ɣʀ > ɰ: > j:

[Cedh]

Would these changes make sense between vowels?

pʰr > ɸr > hr > r˳ > r
pr > br > br > :br > :r
br > βr > βʙ > β˕: > v:
tʰr > θr > hr > r˳ > r
tr > dr > dr > :dr > :r
dr > ðr > ðr > ɹ: > r:
kʰr > xr > hr > r˳ > r
kr > gr > gr > :gr > :r
gr > ɣr > ɣʀ > ɰ: > j:

The sequences for original aspirated and voiceless stops followed by *r definitely make sense, not just individually but also as a system.

I'm not fully convinced by the sequences for the clusters with original voiced stops though: [ɣr] > [ɣʀ] and [ɹː] > [rː] are unusual but possible (but of course you don't need a special justification for [dr] > [rː], only perhaps for keeping this [rː] distinct when you also have a later [dr] > [ːr] change), and [βr] > [βʙ] is extremely unlikely IMO, firstly because [ʙ] itself is already very rare, and secondly because the sequence [βʙ] in particular is quite difficult to pronounce (while still strange, [bʙ] would be a much more likely intermediate - but of course that can eventually go to [vː] too). My biggest problem here is not with any individual shift though, but with the system: these changes are not at all parallel to those of clusters with a voiceless or aspirated stop, so they're unlikely to happen in the same language. A more plausible way to get *br *gr > [vː jː] might be [br ɡr] > [bð ɡð] > [bv ɡj] > [vː jː].

[Haplogy]

Does ɹj > ɻj > ɽʲ > ɽ > ɾ > r make sense?
And ks > q͡χ / __(back vowels) ?

[Colz]

Is this feasible?

ʔ/ŋ/_#

~Lyra

I've seen at least one natlang example of ʔ > ŋ syllable-final, but conditioned by a following consonant. This doesn't sound too weird to me though, depending on what other finals are like.

[Whimemsz]

You can probably appeal to rhinoglottophilia if you need to justify it.

Esmelthien: yes, though you can easily do away with a number of the intermediate steps (ɹj > ɽʲ > r is fine on its own, really)

[Qwynegold]

Does ɹj > ɻj > ɽʲ > ɽ > ɾ > r make sense?

I've had ɹj > rj > r in one conlang. The [ɹ] turned into [r] directly to contrast more with [j] because they were both approximants.

[GrinningManiac]

I'm working on a Lang that I want to have descended from Hindi but have a particular sound.
The Lang would distinguish sounds down Voiced, Unvoiced and Aspirated. Thus:
G K Kh
There are also labalised (gw/kw/khw) and palatalised (gj/kj/khj) distinctions

Hindi however does not have labalised or palatalised consonants and it also has Retroflex and Dental stops (t,th,d,dh) whereas I only retain the dental. Hindi also has aspirated voiced consonants and two vowel lengths. I lose the voiced aspirates and gain a third extra-long length which is distinguished with a slight tone along the lines of Mandarin's third tone.

What sound changes do you think would have occurred to make any or all of these things occur?

[WeepingElf]

I'm working on a Lang that I want to have descended from Hindi but have a particular sound.
The Lang would distinguish sounds down Voiced, Unvoiced and Aspirated. Thus:
G K Kh
There are also labalised (gw/kw/khw) and palatalised (gj/kj/khj) distinctions

Hindi however does not have labalised or palatalised consonants and it also has Retroflex and Dental stops (t,th,d,dh) whereas I only retain the dental. Hindi also has aspirated voiced consonants and two vowel lengths. I lose the voiced aspirates and gain a third extra-long length which is distinguished with a slight tone along the lines of Mandarin's third tone.

What sound changes do you think would have occurred to make any or all of these things occur?

Some ideas:

You could merge the voiced aspirates with the voiced unaspirated stops with vowel lengthening, e. g. /bha/ > /ba:/, /bha:/ > /ba::/. You thus also get the third degree of vowel length you wanted.

Palatalizations and labializations could be induced by front and rounded vowels (or semivowels), respectively.

Getting rid of the retroflexes is easy: merge them with the dentals, or split them up to /r/ + dental. The palatals could become palatalized velars or palatalized dentals.

[Tropylium]

Does ɹj > ɻj > ɽʲ > ɽ > ɾ > r make sense?

The first step— palatalization into a retroflex — seems odd; a typical feature of retroflexes is that they resist the effects of palatalization.

Something like a postalveolar approximant [ʒ˕] would do for an additional intermediate (dunno if that's directly attested, but r >> j is a known sound change from languages like Burmese) but is there any particular reason for you to route this via retroflexes? Perhaps starting as ɹj > ɾj would be the simplest, by dissimilation of the approximant+approximant cluster.

And ks > q͡χ / __(back vowels) ?

Do you also have s > χ and k > q elsewhere before back vowels? Would seem unexpected if not, but even in that case something like this would work:
1) ks > kʃ
2) ʃ > x before back vowels, ʃ > s elsewhere
(if this doesn't occur syllable-final and you have no frontness-neutral vowels, you could even say that this is ks > kx unconditionally, then [x] palatalizes back to [s] before front vowels)
3) x > χ
4) kχ > qχ

[sangi39]

The sequences for original aspirated and voiceless stops followed by *r definitely make sense, not just individually but also as a system.

I'm not fully convinced by the sequences for the clusters with original voiced stops though: [ɣr] > [ɣʀ] and [ɹː] > [rː] are unusual but possible (but of course you don't need a special justification for [dr] > [rː], only perhaps for keeping this [rː] distinct when you also have a later [dr] > [ːr] change), and [βr] > [βʙ] is extremely unlikely IMO, firstly because [ʙ] itself is already very rare, and secondly because the sequence [βʙ] in particular is quite difficult to pronounce (while still strange, [bʙ] would be a much more likely intermediate - but of course that can eventually go to [vː] too). My biggest problem here is not with any individual shift though, but with the system: these changes are not at all parallel to those of clusters with a voiceless or aspirated stop, so they're unlikely to happen in the same language. A more plausible way to get *br *gr > [vː jː] might be [br ɡr] > [bð ɡð] > [bv ɡj] > [vː jː].

Hmmm, got some stuff to think about there...

I ask because I currently have a language which allows the following consonant clusters:

Initial:
pʰr pr br tʰr tr dr kʰr kr gr
pʰl pl bl tʰl tl dl kʰl kl gl

Intervocalic:
pʰr pr br tʰr tr dr kʰr kr gr*
pʰl pl bl tʰl tl dl kʰl kl gl*
rpʰ rp rb rtʰ rt rd rkʰ rk rg
lpʰ lp lb ltʰ lt ld lkʰ lk lg
mpʰ mp mb ntʰ nt nd nkʰ nk ng
pʰm pm bm tʰn tn dn

*These can also be preceded by /m, n, r, l/

Final:
rpʰ rp rb rtʰ rt rd rkʰ rk rg
lpʰ lp lb ltʰ lt ld lkʰ lk lg
mpʰ mp mb ntʰ nt nd nkʰ nk ng

... so far anyway (it's an IE language so I'm still checking the results of previous sound changes to see what turns up and where).

Anyway, I want to get rid of the plosive-initial clusters of the language and I think I've come up with 2 solutions (other than the one I posted before):

Pre-Solution: Add epenthetic schwa after word-final clusters then either,
1) Simply remove plosives in coda position and when followed by /r/ or /l/ word initially, or
2) Delete word-final plosives and then analyse plosive-initial clusters as syllable onsets both word-initially and word-internally (with other clusters being divided between syllables with the second plosive element being the onset of the second syllable). Plosive-initial clusters then undergo changes similar to written (classical?) to modern Tibetan (as seen here), e.g. *pr > tr (then to tʂ in this case), *kl > l and so on.

I like the idea of 2 (that way I could have fun treating i- and u- initial diphthongs as j- and w-final consonant clusters with similar changes) but I'm note sure about "analyse plosive-initial clusters as syllable onsets" having only come across it briefly here:

Most commonly, a single consonant between vowels is grouped with the following syllable (i.e. /CV.CV/), while two consonants between vowels are split between syllables (i.e. /CVC.CV/). In some languages, however, such as Old Church Slavonic, any group of consonants that can occur at the beginning of a word is grouped with the following syllable; hence, a word such as pazdva would be syllabified /pa.zdva/.

I don't know if this kind of syllabification can actually have any bearing on future sound changes or not.

Really, I suppose, I'm asking whether to try and stay away from the more complex (and probably unlikely) changes and go for something simple and well attested, or keep trying until I find a change that's both interesting and plausible. What do you think?