Hallo conlangers!

On Sunday 06 April 2014 13:30:17 BPJ wrote:

> (I'm writing this on my phone which is keyboard challenged. Rather than
> continously switching apps and copy-pasting characters I use CXS and ñ for
> /N/ in romanization.)
> 
> I've been thinking about the Proto-Sohlob 'deictic vowels' I mentioned
> yesterday, as they may be too engelangy. When I first conceived of them I
> thought PS had only the three vowels *i *a *u and ideophonic deictics where
> *i was proximal, *a medial and *u distal wasn't too much of a stretch,

Not worse than the Proto-Semitic case system, preserved in
Akkadian and Classical Arabic: -a absolutive, -u nominative,
-i genitive.

>       but
> now that I know that there were also front *ä and central *ï I'm divided.
> OT1H it may seem less engelangy when not every vowel in the lang is neatly
> slotted into the system,

Yes.

>       but OTOH it also means that the selection of
> exactly these three vowel qualities seems too neat and artificial.

I don't really see the problem here.  But if it feels too neat
to you, change it!

>       I also
> note that the vowels *i *a *u seem too close for comfort to the vowels of
> the English _this, that, yon_.

At least English has _yon_ and not _thut_!

>       I don't know if it's a good or bad thing
> that ideophony may actually be involved in the English forms: the Germanic
> antecedents of _this_ and _yon_ are *þes- and *jain-, and at least the
> latter ending up with a _short_ /o/ already in Old English _geon_ is at
> least not wholly regular. If it weren't for the _æ_ of OE _þæt_ I could let
> the PS medial deictic vowel be *ä and become _ya-_ in Kijeb, so _yana_ pl.
> _yanda_ > _yan, yandar_ in Classical Sohlob and the inanimate _yaru, yadru
> > yor, yodor_ instead of _an~andar, or~odor_. That would be nice as it
> would introduce some more irregularity.

Fine.

>      OT1H Sohlob already contains some
> winks and outright jokes 'connecting' it to various natlangs like _on_ 'is'
> ~ _of_ 'is not' (cf. Finnish _on~ei_ -- originally because of a joke my
> brother-in-law made), but OTOH I'm not sure it needs more of that stuff.

It is you who decides; if I were you, I would not overdo such
in-jokes.

> Incidentally Kijeb _i-_ must be < *ï since *i- would probably > *yi- >
> _gi-_ since /ji ij wu uw/ aren't allowed in Kijeb, becoming /gi igj gu
> ugw/. I'm not sure Proto-Central Sohlob which is the ancestor of Classical
> Sohlob, so that if PS had *in(d)a Kijeb would get _gin(d)a_ and Cidilib
> _jin(dir)_ while PCS would get *yin(da) and CS _yender_ /ji\ndi\r/. I'm not
> sure if that would be the better way to go.

Go inside you, and do what feels best to you.  There is no
definite answer.
 
> The demonstrative stems with which *i- *a- *u- combine to form the actual
> demonstratives are also problematic since they are *-nya/-ni, *-na, *-ru,
> along with interrogative *añya and relative *imi -- because vocab building
> is hard for me and I wanted to be able to do pluralization by reduplication
> like *a-na-na > *anna > Kj. _anda_ and *aruru > *arru > Kj. _adru_ they are
> all *VRV, with a preponderance of nasals because I wanted to
> extrafictionally avoid the 'iyi-uwu' dissimilation described above since it
> would obliterate the effects of reduplication. Things are a bit better now
> that I've definitely concluded that PS had geminate obstruents. A stem
> could be e.g. *-ga giving prox.sg. *aga prox.pl. *agga. Now these __would__
> regularly merge in Kijeb but stay distinct in PCS becoming _aq_ /AR/ ~
> _ag_. Kijeb might make the pl. _añga_ by analogy with _anda_. A possibility
> I have also overlooked would be if one root was *wV, say *wu; an
> interesting pattern would emerge:
> 
> PS      awu     awwu
> Kj.     agu     abu
> Cdl.    oq      of
> PCS     awu     ag(w)u
> CS      ow      oq

Nice!

> Thus a gratis interdialectal complication! ;-) It could also be that
> Kijeb's lack of gemination was only graphical, so that they resorted to
> spellings like __an-ka__ or __at-ka__ as devices to distinguish _agga_ from
> _aga_ or conversely that they wrote _aga_ as __aa__ assuming it was already
> [ARA]! I'm thinking that maybe Kijeb writers had a diacritic (possibly a
> line under the syllable since __-n__ is already an overdot, altho the
> underline is currently an abbreviation sign) to indicate gemination, but
> then they might as well have diacritics for voicing, fricativeness and
> nasal release, and then the basic underspecificness holding on into the two
> later scripts would be hard to explain.

This is fine, too.

>       IIRC Japanese writers in older
> times used kana diacritics only very sporadically although they were in
> primciple available.

And considered the Iroha a perfect pangram, though it uses only
the basic characters, not the diacriticized ones.

>       Perhaps Kijeb writers only ever came up with two
> diacritics which they used randomly with different functions as they saw a
> need for disambiguation, and this bidiacriticness hung on to the later
> scripts. Also they perhaps not got the idea to e.g. write [R] as __a i u__
> with an underline because it was anyway only an allophone, or perhaps it
> wasn't yet fricative between vowels in slow/careful speech, much like in
> modern Swedish. Perhaps they used diacritics more sparingly in inscriptions
> because of conservatism and because the texts were formulaic, as IIRC there
> were more matres lectionis in Phoenician ostraka than in inscriptions.

Yep.  Diacritics are only used when needed, and there non-use
in formal contexts makes perfect sense.

--
... 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