Re: Nasalized fricatives ...

Quoting Benct Philip Jonsson <bpj@...>:

> Andreas Johansson wrote:
>
> > I'm considering /s/>/h/ mutation. On the plus side, it would yield some
> totally
> > hysterical mutations like _seomas_ ['Somas] "creation" > _sheomas_
> ['hjomas]
> > "the creation" and _sem_ [Sem] "tree" > _seshem_ [Se'hem]. On the minus
> side,
> > it might make [s] a rarer and [h] a commoner sound than I really want.
>
> You could have a distinction between formerly geminate and former
> simple *s as you apparently have for nasals.
I already do - that's why the final -s of _seomas_ stays - but the /s/ of _sem_,
_seshem_ can't be from *ss, since it's stem-initial (proto-Meghean only allowed
geminates in post-vocalic position).

> > It occurs to me right know that it would imply the existence of words like
> > _tash_ [tah], which I'd have every excuse to pluralize as _tans_ [tans].
>
> So make it *tass@!  IMO it can be plural _tans_ anyway!
> 'Tis called analogy, y'know! :)
Nah, I prefer _tash_->_tans_ and _tas_->_tasan_ (where the -s of _tas_ is from
*ss).

(There can still be individual exceptions, of course; as mentioned in an earlier
mail, _guthu_->_gunt_, not ->**_guthun_ as expected.)

> > That's
> > probably to much to pass up along with _seshem_ [Se'hem].
>
> What would thàt mean?
That s>h mutation is in.

> > I've considered having [r\] as the mutated ("fricativized") version of [r],
> and
> > even of "anti-mutating" [l] in non-mutating positions to [K\], but I do not
> see
> > any reason the definite form _shouldn't_ be dysfunctional in some cases. In
> any
> > case, it's hard to see what it could _do_ to [i e j w], at least as long as
> I
> > don't allow [j] before front vowels, which isn't gonna happen.
>
> If /r/ is trilled, you can have both /r/ and /l/ become [4] and [l\]
> when lenited -- which IMO would be much cooler.  If you can't pronounce
> a proper [l\] you can always cheat with good ol' Swedish /l\`]! ;)
That's a possibility. I'll consider it.

(Incidentally, I think my 'rl' is simply [l`] - it doesn't seem any less
approximanty than plain 'l' to me.)

> OTOH I agree about the non-lenition of [i e j w] (unless [j w]
> go to zero!)
As said, that would simply wreck to much havoc.

> I'd also have to convince myself of a more intricate reason not
> to
> > have pl an>aD~, which I don't want to happen. Deriving it from a geminate
> is
> > the obvious solution, but, since it's supposed to be "the same" element as
> the
> > infixed -n- in plurals like _ñoch_>_ñoñc_, that's perhaps rather odd.
>
> Why?  The infixed -nn- may have simplified before the following
> obstruent prior to lenition.
I don't know, infixing a geminate just seems weird to me. Any natlang precedent?
In any case, nasals in NC clusters don't lenite, so there's no saying it
wouldn't *work* as intended.

> > The really bright side of having them finally is that it would suggest the
> > existence of a plural formation class of the type _gamh_>_gam_. I like the
> > thought of orthographic plural by truncation.
>
> Brilliant!
I rather thought so myself. :)

> > It's worth noting that even entirely regular inflections can get pretty
> hairy by
> > now; eg _ñoch_ [Nox] "day", _ñhoiñc_ [G~ojNk] "of the days". Neither onset
> not
> > nucleus nor coda is left intact!
>
> Groovy!  Similarity of historically or systemically
> related forms is of course no imperative unless yóu
> make it one, and la langue mechante doesn't seem the
> right place for that!
Had I wanted to keep stems easily recognizeable I'd never touched infixes and
mutations in the first place. I need something aggressively fusional as
counterweight to the agglutinating structure of most of my other projects.

                                                    Andreas

Reply