zompist bboard

vecfaranti wrote:[tl̥] is just traditionalist transcription. It goes with all the other devoiced signs Icelandic has, I guess. Sometimes people pronounce it as an unvoiced approximant, sometimes as an unvoiced fricative. Depends on mood. I'm pretty sure that's how it is in most languages. The phonetic difference between the two is subjective at best.

well, exactly, there you go.

The way I was taught it was that there was no basic difference per se but that [ɬ] generally indicates something more forceful (or something). Then the teacher got a Welsh speaker to demonstrate [ɬ] followed by his impression of [l̥].

finlay wrote:

vecfaranti wrote:[tl̥] is just traditionalist transcription. It goes with all the other devoiced signs Icelandic has, I guess. Sometimes people pronounce it as an unvoiced approximant, sometimes as an unvoiced fricative. Depends on mood. I'm pretty sure that's how it is in most languages. The phonetic difference between the two is subjective at best.

well, exactly, there you go.

The way I was taught it was that there was no basic difference per se but that [ɬ] generally indicates something more forceful (or something). Then the teacher got a Welsh speaker to demonstrate [ɬ] followed by his impression of [l̥].

WELSHMAN DOES SOMETHING USEFUL

MORE AT 11

finlay wrote:Also, a lot of non-native speakers of Welsh pronounce the Welsh LL as /kl/ in English, as in for instance the towns Llandudno or Llanelli, which end up as /klan/-dudno and /klanekli/.

Note: Most people seem to have /l/; I don't often hear /klan/-dudno and /klaneli/ (never ever heard /klanekli/ and it sounds so wrong I doubt I'd recognize it straightaway). But Llangollen is /(k)laŋgɒθlən/ (never /(k)laŋgɒ(k)lən/), and Pwllheli is /pwəθɛli/ (never /pwə(k)lɛli/). So... It varies; it's not just that a lot of people pronounce it as /kl/ - different words do have their own random pronunciations of the <ll>.

Anyway, it does become /kl/ sometimes. Carry on.

On the subject of /ɬ/, would a sound change /ɬ/ > /ʃ/ be believable? I want to get rid of it, but not turn it into /s/. Note that my conlang already has /s ts ʂ tʂ ɕ tɕ/

Astraios wrote:and Pwllheli is /pwəθɛli/ (never /pwə(k)lɛli/)

None of my (English-speaking) friends even bother to pronounce Pwllheli, and I, being the token Welshman in my group, always have to say something like 'no, you don't have to tear out your throat and slice your tongue in two to pronounce Welsh' whenever someone mentions the holiday camp there and asks me to pronounce the name of the place.

Lordshrew wrote:On the subject of /ɬ/, would a sound change /ɬ/ > /ʃ/ be believable? I want to get rid of it, but not turn it into /s/. Note that my conlang already has /s ts ʂ tʂ ɕ tɕ/

I see no problem with that.

I wonder how stable /s ʃ ʂ ɕ/ would be

In one of my langs, l+V sort of becomes pronounced simultaneously, with the openness of the vowel being indicated by the openness of the mouth and the frontness or backness being indicated by roundness. The lang has [i y ʊ u e ə o æ ɐ]. li→ʎ, ly→lʎʷ, lʊ lə → l, lu→lʷ, all the rest as described above. Does this make any sense, does it happen in any natlangs, and is there any way to represent this in IPA?

Mbwa wrote:I wonder how stable /s ʃ ʂ ɕ/ would be

One would expect some mergers; though I think some languages in the Caucasus have all four.

ISTR a system with a four-type coronal sibilant harmony... it'll take me a while to remember where though. In your system the weakest contrast pair would be /ɕ ʃ/ I think.. the rest are fine.

Ubykh had all four of /s ʃ ʂ ɕ/ - as well as /ɬ/. According to Wikipedia, though, some of these contrasts were beginning to be lost in one particular dialect, so maybe the contrast isn't very stable. In any case, I'd be happy to merge /ʃ/ with /ɕ/.

Astraios wrote:

finlay wrote:Also, a lot of non-native speakers of Welsh pronounce the Welsh LL as /kl/ in English, as in for instance the towns Llandudno or Llanelli, which end up as /klan/-dudno and /klanekli/.

Note: Most people seem to have /l/; I don't often hear /klan/-dudno and /klaneli/ (never ever heard /klanekli/ and it sounds so wrong I doubt I'd recognize it straightaway). But Llangollen is /(k)laŋgɒθlən/ (never /(k)laŋgɒ(k)lən/), and Pwllheli is /pwəθɛli/ (never /pwə(k)lɛli/). So... It varies; it's not just that a lot of people pronounce it as /kl/ - different words do have their own random pronunciations of the <ll>.

Anyway, it does become /kl/ sometimes. Carry on.

From your description, it seems as if ɬ becomes kl only word-initially.

Astraios wrote:Most people seem to have /l/; I don't often hear /klan/-dudno and /klaneli/ (never ever heard /klanekli/ and it sounds so wrong I doubt I'd recognize it straightaway). But Llangollen is /(k)laŋgɒθlən/ (never /(k)laŋgɒ(k)lən/), and Pwllheli is /pwəθɛli/ (never /pwə(k)lɛli/). So... It varies; it's not just that a lot of people pronounce it as /kl/ - different words do have their own random pronunciations of the <ll>.

I just did a straw poll of my mother the people currently in the house. I elicited <Llanelli> and was given [xɫə'nɛθli], with an assertion that while she's heard [kɫə'nɛθli] it is definitely incorrect.

<Llangollen> was [xɫæŋgɒxɫən], <Pwllheli> was [pwᵻθɛli].

I figure that something like /c ɟ > tʃ dʒ/ and vice-versa seems plausible (it happens to me in my own idiolect of Vietnamese and other Vietnamese I know, probably because of English influence), but does anyone know if it is actually attested in any natlangs?

Let's say I have a language that has three syllabic nasals *m̩ *n̩ *ŋ̩ at some point...how plausible would the collapse of said nasals to say only phonemic /n̩/ be? With allophonic assimilation to the following consonant be?

cybrxkhan wrote:I figure that something like /c ɟ > tʃ dʒ/ and vice-versa seems plausible (it happens to me in my own idiolect of Vietnamese and other Vietnamese I know, probably because of English influence), but does anyone know if it is actually attested in any natlangs?

this sound change happens literally all the time and can't be reliably ascribed to "English influence".

roninbodhisattva wrote:Let's say I have a language that has three syllabic nasals *m̩ *n̩ *ŋ̩ at some point...how plausible would the collapse of said nasals to say only phonemic /n̩/ be?

Happened in Cantonese: 唔 [m̩˧˩] 'not', 吳 [ŋ̩˧˩] 'Ng (last name), Wu state' > [m̩˧˩].

With allophonic assimilation to the following consonant be?

Cantonese didn't gain assimilation for place of articulation, but I don't see why it couldn't happen too.

Serafín wrote:Happened in Cantonese: 唔 [m̩˧˩] 'not', 吳 [ŋ̩˧˩] 'Ng (last name), 'Wu state' > [m̩˧˩]

Awesome. Thank you.

finlay wrote:

cybrxkhan wrote:I figure that something like /c ɟ > tʃ dʒ/ and vice-versa seems plausible (it happens to me in my own idiolect of Vietnamese and other Vietnamese I know, probably because of English influence), but does anyone know if it is actually attested in any natlangs?

this sound change happens literally all the time and can't be reliably ascribed to "English influence".

Good point. The two sound so similar to me, I don't think I can actually tell them apart.

cybrxkhan wrote:I figure that something like /c ɟ > tʃ dʒ/ and vice-versa seems plausible

Well, as finlay said, palatal stops becoming post-alveolar affricates is ridiculously common; however, I think the reverse, /tʃ dʒ/ > /c ɟ/, is uncommon; sounds tend to move forward in the mouth much more than backwards.

Jetboy wrote:sounds tend to move forward in the mouth much more than backwards.

Really? That seems like a fishy generalization to me. What about /sʲ/ → /ʃ/, /q/ → /ʔ/, /ʃ/ → /x/ → /χ/, /θ/ → /s/, /ɸ/ → /f/?

ná'oolkiłí wrote:

Jetboy wrote:sounds tend to move forward in the mouth much more than backwards.

Really? That seems like a fishy generalization to me. What about /sʲ/ → /ʃ/, /q/ → /ʔ/, /ʃ/ → /x/ → /χ/, /θ/ → /s/, /ɸ/ → /f/?

You're right. If that generalisation was true, we'd have no sounds but [p] by now. I think [tʃ] is just an easier sound in general to pronounce than [c].

True, that is a rather broad generalization; I think I recall Bricka saying something along that line. It does at least hold true for vowels, which are much more likely to front than to back.

I'm writing notes on ideas for one branch of a language family, which will have rampant palatalization, labialization and labio-palatalization. And I have some questions about what some of these non-plain consonants may turn into.

1. If labialized consonants turn into labials, e.g. kʷ > p, what should lʷ turn into? If ɹʷ turns into ʋ, should lʷ do the same and merge?
2. If palatalized alveolars and velars both turn into palatals, e.g. tʲ, kʲ > c, what should become of palatalized labials (p, b, m, f, v, ʋ)? One idea I have is that the palatalization is simply lost, another idea is Cʲ > Cj. But could any prettier things happen?
3. When the things in questions nr 2 happen, what should happen to θʲ? Should it turn into f or s, or something else?
4. Are these phonemes even possible: hʲ, hʷ, hᶣ?
5. Would it be plausible to have the remaining velars merge with the palatals after all these sound changes, simply because they are so uncommon now. I did a quick test on a sample of 55 words, and got these results:

Labials: 19,17% > 43,52%
Alveolars: 65,28% > 21,24%
Palatals: 0% > 33,16%
Velars: 15,03% > 2,07%
Glottals: 0,52% > 0%

This means that when for example labials made 19,17% of all consonants in the protolanguage, in the daughterlang they have increased to 43,52% of all consonants.

1. Any of these seem pretty likely: *lʷ → {w, ʋ, ɫ}
2. Maybe you could have a variety of things happen?
*pʲ *bʲ *mʲ → p b m
*fʲ → ç
(*v) *vʲ → (w) v
(*ʋ) *ʋʲ → (w) j
3. That's kind of a strange sound, so yeah, you could do a bunch of things with it: *θʲ → {ç, s, ʃ, c, j, Ø}
4. Sure. They're basically the same phonetically as [ç ɸ çʷ].
5. You could, but I imagine having a palatal series with no velar series is very, very rare. Sound frequency doesn't necessarily correlate with sound change—a very frequent sound could simply disappear, and an incredibly rare one could stay for thousands of years. Also, after the sound changes your daughters are likely to borrow words or invent that contain velars or what have you, so you won't be stuck with those percentages.

I'm curious: what are you doing with your labio-palatalized consonants?

I'm trying to think of things I can do to my Vr.tV clusters. In one related language it became /ʈ/, in another it became /ʃ/, and I believe there is at least one language where it stayed /rt/. What other possibilities do I have? Is it possible to have a palatal-alveolar thibilant?* I was also thinking of possibly using /ʂ/ or /ɕ/ or maybe even both with the latter being an allophone before /i/ (my only front vowel). As my language already has /ʃ/ and /tʃ/, are there languages that distinguish both /ʃ/ and /ʂ/?

*I know an alveolar thibilant is possible as I have it as a speech impediment for /s/ in some environments (a near merger of Cathy and Cassie).