Weird sound changes

[Otto Kretschmer]

Do you guys know examples of sounds changing into something not close phonemically in a single step?

Like, it is common to see [s] change to [z] or [h] or [ʃ] but can it change into something much further with even a slight probability like [n]?

[alice]

A change from [s] to [n] is only really likely to happen by assimilation, such as something like [ns] -> [nn]. Otherwise, every sound change ultimately proceeds in small steps.

[zompist]

Like, it is common to see [s] change to [z] or [h] or [ʃ] but can it change into something much further with even a slight probability like [n]?

Let's see:
s > r - attested in Latin
r > l - attested in Caribbean Spanish
l > n - attested in Cantonese

As alice says, sound changes are usually small steps. But you have to take timelines into account. If you have 3000 years to play with and no real attestations in between, than s > n could happen in what looks like a single step but isn't (because the intermediate stages are unknown).

[linguistcat]

I wonder if any speech games (like people using UWU talk online for fun, only irl) has ever caused any weird sound changes that stuck around for more than a generation. I think those could cause some interesting sound changes, but I also don't know how plausible it would be.

[KathTheDragon]

s > r - attested in Latin

But didn't this have z as an intermediate?

[zompist]

s > r - attested in Latin

But didn't this have z as an intermediate?

This is commonly assumed; is there evidence for it? (I'm not doubting it; I just wonder how certain it is.)

But that's the point of the rest of my post: the number of steps depends on how closely you zoom in.

[KathTheDragon]

s > r - attested in Latin

But didn't this have z as an intermediate?

This is commonly assumed; is there evidence for it? (I'm not doubting it; I just wonder how certain it is.)

It only happened intervocalically, so s > z seems pretty likely.

But that's the point of the rest of my post: the number of steps depends on how closely you zoom in.

Quite

[Zju]

I can't recall any sources or languages, but there were those two random sound changes in two different oceanic languages, without any apparent motivation or intermediate steps:

1. b > c / #_, nc / V_V
2. gr > kʰ

[Skookum]

Apparently *s > n is attested in Arapaho, but I don't know any theories about what the intermediate steps, if any, were. Plains Algonquian languages in general have a ton of really bizarre sound changes.

[Nortaneous]

Apparently *s > n is attested in Arapaho, but I don't know any theories about what the intermediate steps, if any, were. Plains Algonquian languages in general have a ton of really bizarre sound changes.

https://www.academia.edu/2107195/The_so ... in_Arapaho

I can't recall any sources or languages, but there were those two random sound changes in two different oceanic languages, without any apparent motivation or intermediate steps:

1. b > c / #_, nc / V_V
2. gr > kʰ

1 is Sundanese. I'm not sure what 2 is but there's Drehet *dr > kʰ, Levei *dr > c; also *dr > c in some other languages like Likum and Nauna, and *dr > k~x in Seimat. Maybe an intermediate *tɹ̝̥ˠ as in English tr-? Or just *tʂ. The devoicing is weird but maybe it was +spirant.

[Travis B.]

2. gr > kʰ

My random uneducated guess here would be ɡr > ɡʁ > ɡχ > kχ > kʰ.

[Travis B.]

Of course the above implies kr > kʰ as well.

[Travis B.]

dr > dʐ > dʒ > tʃ > tɕ > c does not seem strange to me.

Hell, I devoice /dr/ even where I wouldn't devoice /d/...

[alice]

This is beginning to resemble one of those word puzzles where you have to get from PHONEME to CONLANG by changing one letter at a time or something.

[Linguoboy]

gr > kʰ

This is reminiscent of Osage /kð/ > /l/. (No intermediate steps attested and the original cluster persists in Omaha-Ponca.)

[Travis B.]

gr > kʰ

This is reminiscent of Osage /kð/ > /l/. (No intermediate steps attested and the original cluster persists in Omaha-Ponca.)

When one breaks things down into a sequence of steps, though, these sorts of things start to seem more plausible. For instance this could very well have been simply kð > kɬ > tɬ > ɬ > l or kð > ɡɮ > dɮ > ɮ > l or like.

[Vijay]

I can't recall any sources or languages, but there were those two random sound changes in two different oceanic languages, without any apparent motivation or intermediate steps:

1. b > c / #_, nc / V_V
2. gr > kʰ

1 is Sundanese.

Huh? Sundanese isn't Oceanic.

[Nortaneous]

Of course the above implies kr > kʰ as well.

*dr was a unit *ⁿdʳ, the prenasalized equivalent of /r/.

I can't recall any sources or languages, but there were those two random sound changes in two different oceanic languages, without any apparent motivation or intermediate steps:

1. b > c / #_, nc / V_V
2. gr > kʰ

1 is Sundanese.

Huh? Sundanese isn't Oceanic.

Sundanese isn't Oceanic, but it's the language mentioned as having that exact sound change in Blust's famous paper on weird sound changes in Austronesian.

(Some of that paper is overstated - Rennellese *l > ⁿg has to be seen in the context of extensive Melanesian influence in the language, so probably we have something like *l > *gʟ followed by contact-induced loss of the lateral component and introduction of prenasalization.)

[Linguoboy]

When one breaks things down into a sequence of steps, though, these sorts of things start to seem more plausible. For instance this could very well have been simply kð > kɬ > tɬ > ɬ > l or kð > ɡɮ > dɮ > ɮ > l or like.

But the point is that this is a recent change and none of those intermediate steps are attested. (*kɬ in particular seems unlikely given that there is a corresponding change of /khð/ to [hl̥] ~ [l].)

[Skookum]

Apparently *s > n is attested in Arapaho, but I don't know any theories about what the intermediate steps, if any, were. Plains Algonquian languages in general have a ton of really bizarre sound changes.

https://www.academia.edu/2107195/The_so ... in_Arapaho

Guess I should have just googled. Interesting, the intermediate steps seem well motivated given the other shifts in the language. Thanks!

The fact that Arapaho has the sound change *p > k / tʃ before front vowels reminds me of a similar well-known "weird" shift: that of *p p' m m' > tʃ tʃ’ ŋ ŋ' in Straits Salish. Unlike Arapaho, this shift occurs in all environments. Its phonological implausibility leads some to suggest an intermediate shift of *p p' > *k k'. As far as I know, the only direct evidence of this is the suffix for "ground": Klallam -nukʷ, -ənəkʷ, Saanich -ənəkʷ, which come from Proto-Salish *-nup. In this case the intermediate velar has been preserved due to irregular labialization after *u.