Print

Print


Happy St Mesrob's!  Seems a fine day to show something off.

This isn't a script I've just now whipped up for the day, but one
about a week old from the latest of the succession of sketches which
happen to me on my commute.  One source of its inspiration was the
"Conscript naturalism" thread, where I said this, rashly:

Den 2018-01-26 kl. 13:28, skrev Alex Fink:
> The alphabet has a nice feature which I see too little of in conscripts: it has digraphs!

In response many folks pointed out conscripts _in the Roman alphabet_
with digraphs, and I saw I should have made that proviso.  With the
Roman alphabet we are used not only to digraphs but to other
non-straightforward uses, other violations of the "one phoneme = one
glyph" principle, so people emulate these features.  But outside Roman
these complexities are much scarcer.  Benct's Sohlob example is the
kind of thing I like here:
  https://listserv.brown.edu/archives/cgi-bin/wa?A2=conlang;794cb1a0.1801D

So I decided to give my sketch a non-boring script by first inventing
a script for a different phonology entirely and then trying to adapt
it.  Works a treat!  Approach heartily recommended.


= Phonology in brief =

In Nov06' the basic phonological unit above the segment is the foot.
Including the concept of syllable doesn't add much descriptive value.
(This is inspired by something I had been reading about Khoisan.)

The template for a foot is (C1) (L) V1 (C2) (V2), where each slot has
its own inventory.  Not both C2 and V2 can be missing.  There are also
unfooted sequences of form (C1) V, where V is like V2 but with only
the height contrast present, which I romanise with a grave accent.
Here are the phonemic inventories.  The romanisation at right I put
between _underscores_ below.

C1:

. p   t   s  ts̻  tʃ  k   kʷ  . p  t  s  c  č  k  kw
. b   d   z  dz̻  dʒ  g   gʷ  . b  d  z  j  ǰ  g  gw
. mb  nd     ndz̻ ndʒ ŋg  ŋgʷ . mb nd    nj nǰ ng ngw
. m   n          ɲ   ŋ   ŋʷ  . m  n        ň  ŋ  ŋw
.     l   r      j           .    l  r     y

L:

. l   r                      . l  r

V1:

. i   ɯ                      . i  u
. æ   ɑ                      . e  a

times tone

. ˩   ˥   ˥˩                 . V  V́  V̂

times phonation

. V   V̤   V̰                  . V  Vh 'V

C2:

. b   l   g                  . b  l  g  (or if V2 is nasal:) m  n  ŋ
.     ɾ                      .    r

V2:

. ɨ   ʉ                      . i  u
. ɜ   ɞ                      . e  o
. a   ɒ                      . a  å

times nasality

. V   Ṽ                      . V  V̨  (but using a nasal letter for C2
takes precedence)

There is lots of feature spread in the V1-C2-V2 sequence, especially
if C2 is absent.  I won't spell out the details.  The combinatorics
justifies which features are allocated to which slot: for instance,
the first V in a CV(C)V foot is often rounded, but the V in a CVC foot
cannot be.

= Graphemes =

I decided to adapt to this an abugida, with a final consonant
indicable as a subscript, with the borrowing made at a diachronic time
a little before the one presented above.  Here are the forms.  Nothing
special, a bit of a Ge'ez pastiche.

https://drive.google.com/open?id=11nG1VGm9b1MMnAHaHZMz2dUwrbMVDDSM
(sorry, I'm temporarily without access to my domain)

The consonant inventory is ‹0 p t c k b d j g s m n r l› plus possibly
some others that weren't borrowed so I didn't make them ;), the vowel
inventory ‹a e i o u›, and there's one diacritic ‹:› that will have
had some effect of making rising diphthongs in the source script.  I
will refer to these graphemic entities in ‹single guillemets› below.

= Use of graphemes =

In C1 position, ‹p t c k b d j g s m n r l› straightforwardly spell _p
t c k b d j g s m n r l_; a vowel spelling V1 will be part of the same
glyph, of course.  Another set of consonants are spelled with a whole
glyph, followed by an onsetless glyph for V1: _kw gw č ǰ ň y_ are ‹ku
gu cu ju ni i›.  (At time of borrowing _č ǰ_ differed from _c j_ in
rounding more than in place.)  The spelling for _ŋ_ is generalised
from C2 position: it is ‹gVn› with a coda ‹-n›, where ‹V› is the
spelling of the V1.  _ŋw_, naturally, is ‹gun› plus an onsetless
glyph.  _z_ is spelled as ‹-d.s›, i.e. with a ‹-d› coda on the
previous glyph, which may need a dead vowel to host it word-initially
(_z_ used to be an affricate too).

The prenasalised C1s, together with the C1-L clusters, are spelled
using dead vowels, e.g. _nd_ as ‹nV.dV›, _gr_ as ‹gV.rV›.  There
originally were articulated vowels in these dead positions, and so a
spelling appropriate to the vowel quality was originally used, but
synchronically there's probably some homogeneisation.

The spellings of the qualities of V1 and V2 are fused.
_ii iu ui uu_ ‹i   iu  u:  u ›
_ei eu ai au_ ‹e:  e:u a:  o:›
_ie io ue uo_ ‹ie  ieu u:e ue›
_ee eo ae ao_ ‹e   eu  ae  oe›
_ia iå ua uå_ ‹ia  io  ua  uo›
_ea eå aa aå_ ‹ea  eo  a   o ›

How are these lists of vowel letters to be interpreted?  When there
are two vowels, the first makes a glyph together with C1, and the
second makes a glyph together with C2, or an onsetless glyph if there
is no C2, e.g. _piru_ ‹pi.ru›, _piu_ ‹pi.u›.  A third vowel is just
tacked on the end, _piro_ ‹pi.re.u›.  When there is one vowel, it is
copied after C2, and if there is no C2 the whole foot gets by with one
glyph, e.g. _piri_ ‹pi.ri›, _pii_ ‹pi›.

If there is no V2, then V1 _i u e a_ are ‹i  u: e  a›.

The funny spellings here are due to adoption rather than sound change:
in particular I imagine ‹u:› as having been a faute de mieux reuse of
a diacritic that happened to be around.  And the distinction between
_ee_ and _ea_ might not be as clean cut as this.

The unmarked spellings of /b l g r/ as C2 are ‹b d g r›.  /l/ can also
be ‹l›, for etymological reasons, though never when V2 is absent.
Some other features are also encoded here.  Nasality of V2 can be
shown by an ‹-n› coda on the last glyph of the foot, but can also be
shown by changing ‹b d› to ‹m n›; again the distribution is
etymological.  There are also etymologies that would have called for
‹ŋ› had there been such a glyph series, but as there isn't, /gV~/ in
C2V2 can only be ‹gVn›.

Non-modal phonation on V1 is marked by the use of ‹p t k› instead of
‹b d g› in C2 position.  If there is no C2 then it is marked by the
provision of a silent ‹-p -t -k› coda homorganic with the following
C1.  This rule was only a fallible correlation etymologically, but the
remaining cases have been analogically brought into line.

And, traditionally, that's it; the script is defective with respect to
the remaining features.  This includes the creaky/breathy distinction,
tone, and sundry minor ambiguities in the above, but perhaps most
significantly the foot boundaries: ‹pi.ru› can be dipedal _piiruu_
just as well as unipedal _piru_.

That's a fairly big ambiguity.  So I posit some sort of scholarly or
clerical circle (you can see the conculturing has not been fleshed
out) that would be discontent with it and innovate the following
refinements:

- The first glyph of a foot is written taller, though still with the
same baseline.  (Something like Japanese large vs. small kana.)

- Later, high tone, the marked and less common level tone, gets
written by taking this first glyph and lifting the baseline as well --
a natural type of thing to exploit given the previous practice, and in
this language the metaphor for pitch is high-pitched = small.  Falling
tone, which is very rare, gets tacked on to the system by a diacritic
in the space above the baseline, originally cartoonesque "motion
lines" conveying 'this glyph is small = high but growing bigger =
lower'.

- Breathy feet get an overdot.  This even works in cases like _pihru_
‹pi.ru̇› which the traditional system doesn't distinguish from modal.
Happily C2 _r_ and creakiness are incompatible so this is a complete
solution.

Here's an example of the names of numbers 1 to 9 written in the
refined system.  Sorry, Janko, I haven't coined 10 yet...
https://drive.google.com/open?id=1LyUk_lSv1z4haUveHJ7lbLOdLPnL3jvP

I'll use capitals to transcribe tall glyphs, and use the same
diacritics as my romanisation for tone indication:
'1' ‹GIti› _g'ili_ [gḭli˩]
'2' ‹JUe:tė› _ǰehli_ [dʒæ̤lɨ˩]
'3' ‹LÚmu› _lúmu_ [lumũ˥]
'4' ‹GOre› _garo_ [gɔɾɞ˩]
'5' ‹SEto› _s'elå_ [sʼæ̰lɒ˩]
'6' ‹PÉaNID› _péanil_ [pʰæː˥nil˩]
'7' ‹KÉ:me› _kémi_ [kʰæmɨ̃˥]
'8' ‹PÁdanMÚ› _pánamúu_ [pʰɑnɑ̃˥muː˥]
'9' ‹GÎNre› _ŋîre_ [ŋiɾɜ˥˩]

(Bonus points if you spotted that the parent language's system was quinary.)


Alex