Hallo conlangers!

On Saturday 02 July 2011 15:09:08, BPJ wrote:

> Does anyone know of other sources of distinctive vowel
> quantity than contraction, open syllable lengthening,
> monophthongization, loss of sonorants or lengthening
> under stress?

Loss of obstruents, of course, or lengthening before a voiced
consonant (while vowels before voiceless consonants remain
short).  There is also "Winter's Law", a (controversial)
sound change in Balto-Slavic wherein a vowel is lengthened
before a voiced stop in a closed syllable but not before a
voiced aspirated stop (some Indo-Europeanists claim that as
evidence for the glottalic theory).  Indo-Iranian has
"Brugmann's Law" in which *o is lengthened in non-final open
syllables.
 
> It feels like Linjeb (a.k.a Western Sohlob) should have
> contrastive vowel length, but I don't know what to
> derive it from. The ancestral language lacked or at
> least lost vowel length[^1], had fixed word-initial
> stress and only two falling diphthongs _*ai, *au_, and
> I already used contraction with a vengance in another
> Sohlob language.

Which makes it reasonable to tread a different road in Linjeb.
 
> It now seems an even earlier form of the language had
> vowel length but lost it by either *ā or *ă becoming
> *æ, and then variously becoming *ya or merging into the
> other *a, *ī and *ū diphthongizing alla
> English/Dutch/German and *ē and *ō either merging with
> *ĭ and *ŭ or diphthongizing or even triphthongizing.

Diphthongs can be used to implement "leap-frogging" sound
changes, as I do in Roman Germanech, where Vulgar Latin
/E/ and /O/ leap-frog /e/ and /o/ by first diphthongizing
to /iE/ and /uO/ (a change that has happened in several
Romance languages) and later monophthongizing to /i/ and
/u/, respectively (a change that has happened in German).

> In fact a whole new image of Proto-Sohlob is emerging,
> partly as a result of working out a related language
> family, partly as a result of an attempt to find a
> suitable intermediate stage between Classical Sohlob
> which is different from Kijeb (the conhistorically
> attested ancestor of the Northern and Western
> varieties) and yet not identical or almost identical to
> Proto-Sohlob.
> 
> In the table below the leftmost column is the old view
> of Proto-Sohlob (which I guess is **Proto-Sohlob now!)
> and the second is the new view of PS. The third column
> is Kijeb, whose three-vowel system without quantity or
> falling diphthongs is more or less set in stone by now.
> The fourth column is Proto-South-Eastern, which used to
> be almost identical to Proto-Sohlob, with a few
> possible alternatives occasionad by the new
> Proto-Sohlob after the commas.
> 
>    Proto-Sohlob   Altern. PS      Kijeb (NW)   Proto-SE
>    -------------- --------------- ------------ -----------
>    *a             *a/*aː [ɑ(ː)]   a            *a
>    *ai            *iː             i            *ai
>    *i             *i              i            *i
>    *ia            *yaː/*ya        ya           *ya
>    *ya            *aː/*a [æ(ː)]   a            *ya
>    *yai           *eː             í, ya        *yai, *æ
>    *yau           *yuː            yu           *yau
>    *iu            *yoː            yu           *yau, *yɔ
>    *yu            *yu             yú, u        *yu
>    *au            *uː             u            *au
>    *u             *u              u            *u
>    *ua            *wa/*waː        wa           *wa
>    *wa            *waː/*wa [ɔː]   a            *wa
>    *wai           *wiː            wi           *wai
>    *ui            *weː            wi           *wai, *wæ
>    *wi            *wi             wí, i        *wi
>    *wau           *oː             ú, wa        *wau, *ɔ
> 
> An obvious advantage of the new PS is that innovation
> and archaism ar more evenly spread between Kijeb and
> PSE.

Yes.  The old Proto-Sohlob is almost identical, vowel-wise,
with Proto-South-Eastern.  The new Proto-Sohlob makes for
more interesting correspondences.  The diachronic developments
underlying your chart make sense and are natural and attractive.

>        I still have to sort out whether it's the short or
> the long *a(ː) which becomes *æ, and whether PSE really
> had a three- vowel system with a lot of diphthongs and
> triphthongs, or if it had in fact a slightly odd
> five-vowel system:
> 
>      i       u
> 
>          ə
> 
>      æ       ɔ/ɑ

This quincunx-shaped vowel inventory is nice and interesting and
not overused.  A good idea, if you ask me.  I cannot recall a
natlang precedent for now, but /i u æ ɑ/ occurs, and you just
add a schwa to that, thus there is nothing wrong about it!

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