bradrn wrote: ↑Wed Nov 08, 2023 3:24 am
Man in Space wrote: ↑Wed Nov 08, 2023 2:57 am
Travis B. wrote: ↑Mon Nov 06, 2023 12:33 pm
You need your language's orthography to be all uppercase and to use overstruck slashes and dashes as its primary diacritics.
There's a specific language you're referring to and I cannot find it on Google for some reason.
Saanich, natively SENĆOŦEN.
Thanks!
Redoing Wǫkratąk…because I can't leave
well enough alone with this POS language family…
Proto-Wǫkratąk (hereinafter
PW because that is a pain to have to type or paste repeatedly)
/m n ŋ/ <m n ŋ>
/mb nd ŋg/ <ṃb ṇd ṇg>
/p b t d k g/ <p b t d k g>
/h/ <h>
/w r (j)/ <w r (y)>
*y tends not to show up in roots but does in some affixes.
/ɔ ɔ̃ a ã ɛ ɛ̃ i/ <o ǫ a ą e ę i>
C₁C₂C₃
Syllables could be of the form C(w/y)V(w/y/N)(C)
V > Ø / #CVC_C
Final high vowels lost
Final *e > i
boda > boda
boda-g > bodg
CǫCraCąk
*√wkt/o 'to order, to command'
*-ar- NMLZ
CVCVCV
nasal theme vowel marks passive voice
C₁enC₂VC₃i > C₁ęC₂VC₃i > C₁ęC₂C₃i > C₁ęC₂C₃
bendVhi > będVhi > będhi > będh > bęęd
wo<n>ko<ra>t-anki > wonkratanki > wõkratãki > wõkratãk
he-wo<n>ko<ra>ta<n>ki > hewnkoratanki > he~wkoratnki > he~wkorantki > he~wkorãtki
C₁C₂C₃
<ra> dative, <-anki> collectivizer
wokot-e > wokti
wokot-i > wokt
dlt
dalat-
dalat-i > dalt
dalat-e > dalti
da<n>lat-e > danlati > dalati
garat-e > gart-i
garat-anki > gartank
ga<ra>rat-i > garrat
hęwkorątki ~ wǫkratąk
*C₁ęC₂C₃i > C₁C₂C₃
he-wenket-i > hew~keti > hęwket >> huket
There seems to have been a profound and progressive vowel harmony at some point because the stems generally carry a
theme vowel with them. Over time, the development of various affixes screwed this up to a fantastic degree, but the person-marking suffixes of the verb often (
q.v.) echoed the theme vowel. The basic
stem in PW was of the shape
*C₁TC₂TC₃-. The person suffixes worked thus:
1
SG *C₁TC₂TC₃-Ty
1
PL *C₁TC₂TC₃-Tn
2
*C₁TC₂TC₃-T
3
*C₁TC₂TC₃-wT
Great! The exception to this was the past-tense marker, which was the vowel
*-i- irrespective of any theme vowels. But that's it. So, given a form like (say)
*√wkt/o, the root is
*w-k-t and the theme vowel is
*o, and it looks like:
Nonpast
1
SG *wokotoy
1
PL *wokoton
2
*wokoto
3
*wokotwo
Past
1
SG *wokotiy
1
PL *wokotin
2
*wokoti
3
*wokotwi
On the way from PW to Archaic W (AW), a few developments of note occurred.
1. V > Ø / #CVC_CV
This change right here did massive damage.
Nonpast
1
SG *woktoy
1
PL *wokton
2
*wokto
3
*wokotwo
Past
1
SG *woktiy
1
PL *woktin
2
*wokti
3
*wokotwi
2. Final high vowels lost
Confounding Doc Brown, this somehow affects the past alone rather than the present.
Nonpast
1
SG *woktoy
1
PL *wokton
2
*wokto
3
*wokotwo
Past
1
SG *woktiy
1
PL *woktin
2
*wokt
3
*wokotw >
*wokoto
3. Final *e > i
…Actually, not all that relevant here. Moving on.
4. Nasal vowel collapse
…and here we go.
Nonpast
1
SG *woktoy
1
PL *woktǫ
2
*wokto
3
*wokotwo
Past
1
SG *woktiy
1
PL *woktį > becomes
*woktę because there is no
Xį in Ba Sing Se
2
*wokt
3
*wokotw >
*wokoto
5. Glides off into the distance
Right, so we have these
*w and
*y in the ultima that got there because of these verb endings. How do we get rid of them while leaving a tantalizing clue as to their presence?
Like this:
Nonpast
1
SG *woktô
1
PL *woktǫ
2
*wokto
3
*wokôto
Past
1
SG *woktî
1
PL *woktę
2
*wokt
3
*wokoto
6. The Best or Worst Syncope Rule
Before long vowels, a short vowel dropped. (This appears to include nasal vowels, which appear to have been pronounced phonetically long.)
Nonpast
1
SG *wktô
1
PL *wktǫ
2
*wokto
3
*wkôto
Past
1
SG *wktî
1
PL *wktę
2
*wokt
3
*wokoto
7. Length-be-gone
And then the phonemic length distinction was toast.
Nonpast
1
SG *wkto
1
PL *wktǫ
2
*wokto
3
*wkoto
Past
1
SG *wkti
1
PL *wktę
2
*wokt
3
*wokoto
This…
This…this is…
I
love this.
This, theoretically, means the only "hash collisions" would occur in verbs with
*i as a theme vowel (in the 1
SG the past and nonpast are the same) or
*e as a theme vowel (the past and nonpast of the 1
PL are the same). We get by in English with "I cut it"…and probably a few other verbs I cannot identify offhand because I
want to…I think they'll do.
So the citation form of the root could well just be the 1
SG nonpast:
√wkto.