Phonological System: Overlong Vowels and Consonants

I came up with a phonological system yesterday because we
happened to be studying case in Estonian (a Jim Blevins article).
Estonian has overlong consonants and vowels, so I decided to
come up with a system that would produce overlong vowels
and consonants.  I don't know if I'll ever put it to use, so I thought
I'd write it up here, just to keep a record of it.

[Below, I'm going to use quote marks ("...") when talking about
orthography, and the regular brackets and slashes when transcribing.
Also, the asterisk * is going to mark historical forms, not
ungrammatical ones.]

Older Form of the Language (I'll call it Es):

Consonants:
Stops: p/b, t/d, k/g
Fricatives: f, T, s/z, S/Z, x, h
Nasals: n, m
Approximants: r, l

In this form of the language, all consonants can be geminate except
for voiced obstruents /b, d, g, z, Z/ and /h/.

Vowels:
Short: a, e, i, o, u
Long: a:, e:, i:, o:, u:

There's a general rule that a long vowel can't precede a geminate.

At this point, the writing system is established.  It's an alphabetic
writing system, and there's a character for every phoneme listed
above, including totally separate characters for the long vowels.
Geminate consonants are written as two version of the same consonant
glyph.  Thus, you can take all the phonemes above and put them
in quotes and you'll get the "orthography".

------------------------------------------------------------------------
-----------
Modern Form of Es:

Sound changes:
Consonants:
(1) All singletons remains singletons (e.g., *p > p).
(2) Voiced obstruents > geminate voiceless obstruents (e.g., *b > pp;
*z > ss).
(3) Obstruent geminates > overlong: (e.g., *pp > ppp; *SS > SSS).
(4) Geminate weak fric. > affricates (i.e., *ff > pf; *TT > tT; *xx >
kx).
(5) Sonorant geminates remain the same (e.g., *mm > mm).
(6) *CC > C / _#

[Note: Nothing happens to lonely /h/.]

Vowels:
(1) V > central / _C:: (or C:C, CCC) (e.g., *a/e/o > @; *i > 1; *u >
u-).
(2) V > V / _C:V (or CCV).
(3) V > V: / _CV (e.g., *a > a:).
(4) V: > V / _CCCV (e.g., *a: > a).
(5) V: > V: /_CCV
(6) V: > V:: / _CV
(7) V > central /_C# (e.g., *a/e/o > @; *i > 1; *u > u-).
(8) V: > V / _C# (e.g., *a: > a).

Thus, the new phoneme inventory of the language is as follows:

Consonants:
Stops Reg.: p, t, k
Stops Long: pp, tt, kk
Stops Overlong: ppp, ttt, kkk
Fricatives Reg.: f, T, s, S, x, h
Fricatives Long: ss, SS
Fricatives Overlong: sss, SSS
Affricates: pf, tT, kx
Nasals: n, m
Approximants: r, l

Vowels:
Central: @, 1, u-
Short: a, e, i, o, u
Long: a:, e:, i:, o:, u:
Overlong: a::, e::, i::, o::, u::

Here's how they would be spelled:

Stops Reg.: p, t, k
Stops Long: b, d, g
Stops Overlong: pp, tt, kk
Fricatives Reg.: f, T, s, S, x, h
Fricatives Long: z, Z
Fricatives Overlong: ss, SS
Affricates: ff, TT, xx
Nasals: n, m
Approximants: r, l

Vowels:
Central: a/e/o, i, u
Short: a, e, i, o, u
Long: a:, e:, i:, o:, u:
Overlong: a:, e:, i:, o:, u:

And the vowels, then, would be distinguishable by the number
and quality of consonants that follow them.

One bothersome bit about this system is that it's unbalanced, it
would seem, in that there are tons of voiceless consonants, and
the only voiced consonants are sonorants and vowels.  To counterbalance
this, you could have voiceless singleton obstruents becoming
voiced intervocalically (and, say, after nasals).  This'd give you
back /b/, /d/, /g/, /z/, and /Z/ (now spelled "p", "t", "k", "s" and
"S")
in specific contexts.

Now here's some things that could be done with this system.

------------------------------------------------------------------------
-----------
So, first, now you can have the a bunch of distinctions that didn't
exist previously merely in vowel quality and or length (plus
consonant length, I suppose):

-"kapa" > [ka:p@]
-"kaba" > [kapp@]
-"kappa" > [k@ppp@]
-"ka:pa" > [ka::p@]
-"ka:ba" > [ka:pp@]
-"kapa:" > [ka:pa]
-"kaba:" > [kappa]
-"kappa:" > [k@pppa]
-"ka:pa:" > [ka::pa]
-"ka:ba:" > [ka:ppa]

For added fun, you could throw in a rule where you lose [h] in
coda position, but it still counts as a C for the purposes of
determining
the length of a vowel.  Thus, you could have "kappa" [k@ppp@] forming
a minimal pair with "kahba" [k@pp@].

In a morphological system (let's say, a case system), you could
then make use of the differing vowel and consonant qualities:

"kaff" [k@f] (= N, "dog", let's say)
Sg.
Nom: "kaff" [k@f]
Acc: "ka:f" [kaf]
Gen: "ka:fad" [ka::f@t]
All: "ka:fadam" [ka::fatt@m]
Ade: "ka:fadal" [ka::fatt@l]
Abl: "ka:fadas" [ka::fatt@s]

Plu.
Nom: "keffi" [kepfi]
Acc: "ke:fi" [ke::fi]
Gen: "ke:fid" [ke::f1t]
All: "ke:fidim" [ke::fitt1m]
Ade: "ke:fidil" [ke::fitt1l]
Abl: "ke:fidis" [ke::fitt1s]

I added a little umlaut for fun (and, then, you could have a whole
set of new orthographical forms for fronted back vowels, even
though their qualities are identical to the front ones).  So now
you have the same word with three different vowel lengths, and
three different qualities.  Additionally, you have one form where
the affricate appears.  Looks totally bizarre, and almost as if it
comes out of the middle of nowhere if you look only at the
transcriptions, but, in fact, it's regular and predicted if you know
the last sound is a geminate.  Anyway, so that could be one class,
that has a vowel length change.  You could have another class that
has a consonant length change:

"nesi" [ne:si] (= N, "cat", let's say)
Sg.
Nom: "nesi" [ne:si]
Acc: "nessi" [n@sssi]
Gen: "nessid" [n@sss1t]
All: "nessidim" [n@sssitt1m]
Ade: "nessidil" [n@sssitt1l]
Abl: "nessidis" [n@sssitt1s]

Plu.
Nom: "nesi:" [ne:si]
Acc: "nessi:" [n@sssi]
Gen: "nessi:d" [n@sssit]
All: "nessi:dim" [n@sssi:tt1m]
Ade: "nessi:dil" [n@sssi:tt1l]
Abl: "nessi:dis" [n@sssi:tt1s]

In addition to increasing the length of a segment to form the
accusative, you could also decrease the length.  Additionally,
if you didn't want syncretism for the nominative and accusative
in this class you could make it so that even word-final vowels
became centralized.  Or you could have some different way of
forming a plural with V-final words.

Anyway, what you get is a system that looks regular in the
orthography, but which is really rather mischievous.

-David
*******************************************************************
"sunly eleSkarez ygralleryf ydZZixelje je ox2mejze."
"No eternal reward will forgive us now for wasting the dawn."

-Jim Morrison

http://dedalvs.free.fr/