Hallo conlangers!

On 21.06.2016 14:02, BPJ wrote:

> A case in point regarding doubtful naturalistic realism in Sohlob: given
> the following sound changes (where <y> is phonetically and perhaps also
> phonemically palatalization)
>
> ty > ts > s
> dy > dz > z
> ky > tś (CXS [ts\])
> gy > dź
> ry > l (ell)

Fair.

> how likely is it that different branches of the family would have
>
> try > tl
> dry > dl
> kry > kl
> gry > gl
>
> or
>
> try > tsr > sr > str (and similar for dy)
> kry > tśr > ts` (retroflex) (idem gy)

Palatal > retroflex seems a bit like a stretch, but not too far out, and 
you have /r/ here, which acts as a retroflexing force in quite a few 
languages (whom do I tell that? You have it in your L1!).

> or that some branch have the lateral outcome with ty, dy and the affricate
> + r outcome with ky, gy, or for that matter an affeicate + lateral outcome?
>
> Also how likely is it that a /N_j/ (probably [J]) and /n_j/ stay distinct
> any length of time, or that the former merges with /j/ rather than with
> /n_j/?

The two palatalized nasals are indeed likely to merge, but I don't see 
how the change you have was impossible.

> Tertio: the likelihood that a mid front unrounded vowel merge with a schwa
> and then the schwa merging with /i\/, rather than /e/ merging with /&/ in
> the course of the development of a high vs. low vowel harmony?

I cannot give a definite without knowing about the rest of your vowel 
inventory, but I see no problem with this. I have seen weirder changes 
to vowels, such as the mess we see in Tocharian.

> Quarto: how likely is a height-based vowel harmony in the first place? I've
> read that Korean probably has had a height-based VH, but beyond that, esp.
> how one may realistically emerge from an /i a u/ or /i & a u/ vowel system
> with pervasive allophonic umlaut I'm merely speculating. It seemed
> naturalistic to me years ago but I'm long since not so sure anymore!

Not worse than what happened in Chukotko-Kamchatkan. That family has a 
height-based vowel harmony /i æ u/ ~ /e a o/ plus a neutral schwa.

> These questions have literally bugged me for years! Currently I'm trying to
> model several outcomes as a dialect continuum, but I have my doubts that
> the more 'exotic' options are naturalistic at all!
>
> PS I hope there aren't too many typos. I'm typing this in gmail on my
> iphone. It has a very small unenlargable font and my eyesight has taken a
> turn for the worse recently!

On 21.06.2016 15:06, Melroch 'Aestan wrote:

> I should have mentioned that *sr [hr] and *sry [hr_j/hl] become [r_0] 
> and [K] respectively or both ]K] depending on dialect. *sty and *zdy 
> become śt and źd The same queries apply to *xy, mutatis mutandis. 

Fine.

> I long ago nicked from Middle Chinese a scenario where
>
> py > f
> by > v
> my > v
> wy > v
>
>   but how certain/common is this scenario?

It may be uncommon, but I see no reason why it couldn't have happened.

> Moreover I long ago assumed [D] > [z] based on the Persian/Turkish/Indic
> adaptation of Arabic. I have since learnt that such a change actually
> occurred between vowels in Occitan (where however [Dr] > [jr]) but how
> common is this change? In particular did it actually happen in dialects of
> Arabic. I guess the answer is yes since */D_e/ is commonly transliterated
> <z_.> (dot below) or is this a notational Persianism or something which
> didn't apply similarly to the unemphatic counterpart? Conversely, might [z]
> become [D]?

/T/ > /s/ and /D/ > /z/ are fair. Happened in Persian as Alex has 
pointed out. Also, Germans do it all the time when trying to speak 
English! /z/ > /D/ may be more of a stretch, but I wouldn't say it 
couldn't occur.

> Further: how likely is it that the outcome of coronal + /j/ and coronal
> before /i/ is different, i.e. that they palatalized to [ts] or [ts\]
> etc. at different times or the latter not at all, there also being an /i\/
> in the older form of the language, which in some dialects merged with /i/
> late enough to not affect preceding consonants or not at all?

While there are many examples of languages where coronals palatalized 
alike before both /i/ and /j/, I wouldn't say that's an unbreakable law! 
And if you have an /i\/ merging into /i/, that may have merged after the 
palatalization having taken its course, of course.

> This coupled
> with *i possibly becoming /i\/ before an /u/ and/or /i\/ in the following
> syllable and *u becoming /i\/ or /y/ before an *i in the following
> syllable. By itself each change is not implausible, but how would/could
> they plausibly interact? There are clearly several possibilities.

This is a nice way of getting palatalized consonants before back vowels, 
and unpalatalized consonants before front vowels!

> Lastly: how might [l] arise other than from [r_j]. I've seen all of [r D
> rd/rD n] become [l] in different languages, Iranian languages being
> especially fertile.

Yep. There are plenty of sources for laterals. Tolkien had /d/ go to /l/ 
in many contexts in Quenya (which you certainly are familiar with), and 
he knew very well what he was doing.

> How do/would/might [rr DD nn] avoid or not avoid
> becoming [l] if the corresponding singletons do?

In some languages, geminates are generally more resistant to change than 
singletons, I think.

> Well, I'm afraid that was almost the whole bunch of major Sohlob
> uncertainties in one go.
>
> I'll try to wait and see what responses come in until tonight/tomorrow and
> answer as much as I can in one go.

Fair.

--
... brought to you by the Weeping Elf
http://www.joerg-rhiemeier.de/Conlang/index.html
"Bêsel asa Éam, a Éam atha cvanthal a cvanth atha Éamal." - SiM 1:1