Anthologica Universe Atlas / Forums / Department of Creativity / dhoklang Scratchpad (NP: Algonquitut?)
VAIs
We are now beginning to get into the meat of the verbal system, with real person marking. Because these verbs are intransitive, there's still no polypersonal or direct/inverse marking, but we will see some incorporation. The template is pretty much the same as that for VIIs:
[personal marking/order]-(TAM)-(preverbs)-verb root-(voice/number marking)-(medials)-(more TAM)-(finals)
but we'll actually see some medials in play.
Personal Marking/Order
Here, the order system of independent/conjunct I/conjunct II that we've already seen is now crossbred with a personal marking system: first person, second person, third person animate proximate, and third person animate obviative. Order affixes usually go after the personal markers, but occasionally there are syncretic forms, so we're treating them at the same time.
The personal markers are similar in the first and third person proximate to possessive prefixes on nouns, but the second and third person obviative are not the same. Motivations for this are unclear.
The order markers are similar to those in VIIs:
Order prefixes usually appear after personal markers. However, there are a couple of syncretic forms in the conjunct orders: the third person proximate will combine with Conjunct II to produce the prefix hō(h)-, and the third person obviative will combine with Conjunct II to produce kʷ'ē(h)-. The second person will combine with Conjunct I to produce šɨ:(h)-. Additionally, the first and second person prefixes nɨ- and čɨ- will back to no- and čo- when followed by Conjunct II.
So, to recap:
The TAM prefixes and preverbs are pretty much the same system as in VIIs, so there's no need to go over them again.
Voice/Number marking is pretty simple. VIAs may have, in addition to the passive -(h)iknɨ- that we've already seen in VIIs, an antipassive -(y)ekʷ(a)- suffix which can be added to transitive verbs to eliminate their object. Neither of these are particularly common, however, and more interesting is the number marking suffix, which appears after any voice suffix. No suffix appears in the singular; in the plural, there are two possible affixes:
———————————————-
OK, onto medials. Here's where we're going to see noun incorporation. (Arguably the more nouny preverbs also count as noun incorporation, but they're not quite as obvious.)
Generally speaking, the only nouns that get incorporated are objects and instruments. Pretty much any noun can be incorporated, but it has to be stripped of any affixes, including prenouns, number-marking, obviation and case. If an object is incorporated, you can stop there; if an instrument was incorporated, you'll additionally want the instrumental suffix -cē- appearing before the incorporated noun. Usually, object medials are attached to VTA or VTI stems, which turns them into VIA stems. An example (skipping ahead to flesh out a bit of VTA morphology, which I'll get to later...)
If an incorporated noun has a final vowel, this is deleted if a following, vowel-initial affix follows. If it reappears and was long in the independent noun, it'll always be short in the incorporated form.
nismā nɨ-Ø-p'i:nɨ-Ø-Ø
potato 1sg-ind-eat smth (VTI)-sg subj/sg obj-nonpast
'I am eating a potato.'
nɨ-Ø-p'i:nɨ-Ø-nisma-Ø
1st-ind-eat smth (VTI -> VIA)-sg subj-potato-nonpast
'I am eating a potato, I am potato-eating.'
Only one noun can be incorporated into a verb.
TAM suffixes and finals are the same as in VIIs, so I'm not going to go over them again.
Another glossed sentence:
łɨ-hoteʔ-āt-et hā-Ø-kōhθ-ēkma-cik, nɨ-hoteʔ-ohta-het ke-Ø-skēł-ēkma-cik.
2sg-sheep-pl.prox-sg.poss 3PROX-ind-be.black (VAI)-pl.-μέν, 1sg-sheep-pl.obv-sg.poss 3OBV-ind-be.white (VAI)-pl.-δέ
'Your sheep are black, on the one hand, but my sheep, on the other, are white.'
We are now beginning to get into the meat of the verbal system, with real person marking. Because these verbs are intransitive, there's still no polypersonal or direct/inverse marking, but we will see some incorporation. The template is pretty much the same as that for VIIs:
[personal marking/order]-(TAM)-(preverbs)-verb root-(voice/number marking)-(medials)-(more TAM)-(finals)
but we'll actually see some medials in play.
Personal Marking/Order
Here, the order system of independent/conjunct I/conjunct II that we've already seen is now crossbred with a personal marking system: first person, second person, third person animate proximate, and third person animate obviative. Order affixes usually go after the personal markers, but occasionally there are syncretic forms, so we're treating them at the same time.
The personal markers are similar in the first and third person proximate to possessive prefixes on nouns, but the second and third person obviative are not the same. Motivations for this are unclear.
| 1st | nɨ- |
| 2nd | čɨ- |
| 3AN.Prox | hā- |
| 3AN.Obv | ke[ʔ]- |
The order markers are similar to those in VIIs:
| Independent | -Ø- |
| Conjunct I | -se(h)- |
| Conjunct II | -we(h)- |
Order prefixes usually appear after personal markers. However, there are a couple of syncretic forms in the conjunct orders: the third person proximate will combine with Conjunct II to produce the prefix hō(h)-, and the third person obviative will combine with Conjunct II to produce kʷ'ē(h)-. The second person will combine with Conjunct I to produce šɨ:(h)-. Additionally, the first and second person prefixes nɨ- and čɨ- will back to no- and čo- when followed by Conjunct II.
So, to recap:
| Independent | Conjunct I | Conjunct II | |
|---|---|---|---|
| 1st | nɨ- | nɨse(h)- | nowe(h)- |
| 2nd | čɨ- | šɨ:(h)- | čowe(h)- |
| 3rd Prox | hā- | hāse(h)- | hō(h)- |
| 3rd Obv | ke[ʔ]- | keʔse(h)- | kʷ'ē(h)- |
The TAM prefixes and preverbs are pretty much the same system as in VIIs, so there's no need to go over them again.
Voice/Number marking is pretty simple. VIAs may have, in addition to the passive -(h)iknɨ- that we've already seen in VIIs, an antipassive -(y)ekʷ(a)- suffix which can be added to transitive verbs to eliminate their object. Neither of these are particularly common, however, and more interesting is the number marking suffix, which appears after any voice suffix. No suffix appears in the singular; in the plural, there are two possible affixes:
-ahw(ɨ)- indicates the presence of a first-person argument. This is always the pluralizing suffix used when the subject is first person plural; however, the prefix will make a clusivity distinction. If the subject is first person exclusive, use a first-person prefix; if the subject is first person inclusive, use a second-person prefix.
-(y)ēkm(a)- indicates the lack of a first-person argument. Thus it's used in the second person plural and with third-person subjects.
———————————————-
OK, onto medials. Here's where we're going to see noun incorporation. (Arguably the more nouny preverbs also count as noun incorporation, but they're not quite as obvious.)
Generally speaking, the only nouns that get incorporated are objects and instruments. Pretty much any noun can be incorporated, but it has to be stripped of any affixes, including prenouns, number-marking, obviation and case. If an object is incorporated, you can stop there; if an instrument was incorporated, you'll additionally want the instrumental suffix -cē- appearing before the incorporated noun. Usually, object medials are attached to VTA or VTI stems, which turns them into VIA stems. An example (skipping ahead to flesh out a bit of VTA morphology, which I'll get to later...)
If an incorporated noun has a final vowel, this is deleted if a following, vowel-initial affix follows. If it reappears and was long in the independent noun, it'll always be short in the incorporated form.
nismā nɨ-Ø-p'i:nɨ-Ø-Ø
potato 1sg-ind-eat smth (VTI)-sg subj/sg obj-nonpast
'I am eating a potato.'
nɨ-Ø-p'i:nɨ-Ø-nisma-Ø
1st-ind-eat smth (VTI -> VIA)-sg subj-potato-nonpast
'I am eating a potato, I am potato-eating.'
Only one noun can be incorporated into a verb.
TAM suffixes and finals are the same as in VIIs, so I'm not going to go over them again.
Another glossed sentence:
łɨ-hoteʔ-āt-et hā-Ø-kōhθ-ēkma-cik, nɨ-hoteʔ-ohta-het ke-Ø-skēł-ēkma-cik.
2sg-sheep-pl.prox-sg.poss 3PROX-ind-be.black (VAI)-pl.-μέν, 1sg-sheep-pl.obv-sg.poss 3OBV-ind-be.white (VAI)-pl.-δέ
'Your sheep are black, on the one hand, but my sheep, on the other, are white.'
Taking a break from verbs to start work on the first daughter, Old Quomonak (kʷomonak- I reserve the right to change the name later on, since I'm not sure what this will mean), which will be the ancestor of the language of my main conculture. We'll start with some sound changes, and then move onto nouns and then eventually verbs, which I expect to be a bit of a clusterfuck, but hopefully analogy will present some solutions to straighten it all out.
The calendar of my main conculture starts at year 0, which represents the mythical founding of its mother city (somewhat like the founding of Rome). The current year is 2673, and the proto-language is thought to date to around -3000.
Pre-Quomonakian Changes, c.-3000 to c.-2500
Many of these changes can be found in other branches as well.
1. Initial glide-vowel reduction
The weak vowels /e ɨ/ merge with a preceding glide into the vocalic version of that glide when word-initial. Schematically:
we wɨ -> o / #_
ye yɨ -> i / #_
So, *weyɨ̄ *yente -> owi īte.
2. Loss of *ʔ
The glottal stop *ʔ is lost before a vowel:
ʔ -> Ø / _V
Examples include *nɨʔo -> niyo and *ʔočīča -> ohceyahca.
3. Merger of *r and *l.
r -> l / _
Common to almost all branches of the Nahtak family. Example: *k'ōri -> kōli.
4a. Vowel lowering before coda *h.
Before coda /h/, the vowel /i/ lowers to /e/ (whether short or long), and /ɨ e o/ lower to /a/, again, whether short or long.
i -> e / _h[C#]
ɨ, e, o -> a / _h[C#]
Example: *kʷēh -> kʷē. (This doesn't look like lowering, but there's an intermediate step *kʷā.)
4b. Merger of coda *h and *ʔ.
*h becomes *ʔ when a coda. Even though neither *h nor *ʔ survived in coda position to the extant language, effects on surrounding consonants and vowels make it clear that this (or something similar) must have occured. Examples aren't going to illustrate the point very well, so I'll just move on.
5. Backing of *i.
*i becomes /ɨ/ when preceded by a velar or labiovelar, that is, one of /k k' kʷ kʷ' ŋ w/.
Example: *mɨkʷik -> mohkʷoh.
6. Loss of *h.
*h is lost in all positions, everywhere.
h -> Ø / _
Pre-Quomonakian Changes Cont., c.-2500 to c. -2000.
Past this point we start to see differentiation of the Quomonakian branch.
7a. Palatalization of alveolar sibilants.
*s *c *c' change to *š *č *č' when followed by *i.
s -> š / _[i ī]
c -> č / _[i ī]
c' -> č' / _[i ī]
7b. Debuccalization of *s.
Remaining *s changes to *h in all positions.
s -> h / _
8a. Simplification of affricates.
*c changes to *s, while *c' and *č change to *c. *č' becomes *č. Schematically:
c -> s / _
c', č -> c / _
č' -> č / _
8b. Palatalization of *t
*t and *t' become *č before *i. This must have occurred after change 7a, since we see č as the modern reflex, not **c.
t, t' -> č / _[i ī]
9. Loss of *θ
*θ becomes /s/ intervocalically and /t/ otherwise.
θ -> s / V_V
θ -> t / otherwise
10. Initial spirantization of stops.
*p *t *k *kʷ become *ɸ *θ *h *w when word-initial and followed by a short vowel.
p -> ɸ / #_V[+short]
t -> θ / #_V[+short]
k -> h / #_V[+short]
kʷ -> w / #_V[+short]
11. Final vowel shortening.
Long vowels become short when word-final.
V: -> V / _#
Pre-Quomonakian to Proto-Quomonakian, c. -2000 to c. -1000
The Quomonakian branch comes into its own. I am considering breaking off a closely-related sister branch (think Baltic to Slavic or Indic to Iranian) somewhere in the middle here.
12. Insertion of /h/ before internal stops and affricates.
An epenthetic /h/ is inserted before intervocalic *p *t *c *č *k *kʷ.
Ø -> h / V_C[+stop/affricate]V
13. Loss of coda *ʔ.
Coda *ʔ, which by now is the only instance of *ʔ, is deleted.
ʔ -> Ø / _
14. Ejectives merge with plain stops.
Remaining ejectives merge into their plain counterparts:
C' -> C / _
15. Palatalization of *k *kʷ
Yet another palatalization rule- here, *k becomes *c, and *kʷ becomes *k, before *i.
k kʷ -> c k / _[i ī]
16a. Shortening of homoorganic vowels.
A long vowel, followed in the next syllable by the same vowel (whether short or long), is shortened. So, *īkʷi -> *ikʷi, and *ōpō -> *opō, but *āpi does not change.
V₁: -> V₁ / _$V₁
16b. Shortening of long vowels between long syllables.
A long vowel, if both the preceding and following syllables contain long vowels, is shortened. This remains a conditioned rule in the language.
V: -> V / V:$_$V:
17. Long Vowel Raising
Long vowels, except *ɨ̄ *ō, raise, with *ī breaking:
ā -> ē / _
ē -> ī / _
ī -> ya / #_
ī -> eya / otherwise
18. Merger of *ɨ
*ɨ merges into *o after labials, and into *a otherwise.
ɨ ɨ̄ -> o ō / [p m w kʷ]_
ɨ ɨ̄ -> a ā / otherwise
19a. Loss of glides before similar vowels
*w is lost before the back vowels *a *o, and *y before the front vowels *e *i:
w -> Ø / _[a ā o ō]
y -> Ø / _[e ē i ī]
19b. Glide reinsertion
Quomonak doesn't like vowel hiatuses, so new glides pretty much immediately get reinserted into the language:
Ø -> w / B_V
Ø -> y / F_V
Proto-Quomonak to Old Quomonak, c. -1000 to c. 300
By this point we're dealing with a branch at about the scale of Italic or Celtic, and changes from here on in detail the more recent development of Old Quomonak itself.
20. *h deletion in double clusters
When a coda *h is followed by a syllable with another coda *h, it gets deleted.
h -> Ø / _CVh[C#]
21. *ŋ mergers
*ŋ merges with *m or *n depending on the following vowel:
*ŋ -> m / _B
*ŋ -> n / _F
22. *ł dentalization
*ł merges with those instances of /θ/ that have developed, generally from initial spirantization.
ł -> θ / _
22a. Coda *k spirantization
*k becomes /h/ in coda:
k -> h / _C, _#
22b. Coda *t to /k/ change
Coda *t becomes /k/ in a chain shift from the preceding change.
t -> k / _C, _#
22c. Coda glide mergers
The coda glides *y *w...oh boy. These merge and often change the preceding vowel in a variety of ways. They're too complicated to list in prose, so I'll just present the changes schematically.
Fy -> ī / _C, _#
iw, ew -> yo / #_[C#]
īw, ēw -> īyo, ēyo / _[C#]
ay, oy -> wi / #_[C#]
ay, oy -> awi, owi / _[C#]
āy, ōy -> āwi, ōwi / _[C#]
Bw -> ō / _C, _#
22d. Loss of nasals in coda with lengthening.
Coda nasals delete, lengthening preceding vowels:
VN -> V: / _C, _#
22e. Approximation of coda *š *l
*š *l become /y w/ in coda:
š -> y / _C, _#
l -> w / _C, _#
22f. Dissimilation of *hh.
Double *h isn't allowed, so new examples of *hh become *hk.
h -> k / h_
23a. Loss of initial *w.
Initial *w is lost.
w -> Ø / #_
23b. Approximation of *ɸ.
*ɸ becomes *w in all positions, likely in a chain shift (since it's basically only found initially.)
ɸ -> w / _
The calendar of my main conculture starts at year 0, which represents the mythical founding of its mother city (somewhat like the founding of Rome). The current year is 2673, and the proto-language is thought to date to around -3000.
Pre-Quomonakian Changes, c.-3000 to c.-2500
Many of these changes can be found in other branches as well.
1. Initial glide-vowel reduction
The weak vowels /e ɨ/ merge with a preceding glide into the vocalic version of that glide when word-initial. Schematically:
we wɨ -> o / #_
ye yɨ -> i / #_
So, *weyɨ̄ *yente -> owi īte.
2. Loss of *ʔ
The glottal stop *ʔ is lost before a vowel:
ʔ -> Ø / _V
Examples include *nɨʔo -> niyo and *ʔočīča -> ohceyahca.
3. Merger of *r and *l.
r -> l / _
Common to almost all branches of the Nahtak family. Example: *k'ōri -> kōli.
4a. Vowel lowering before coda *h.
Before coda /h/, the vowel /i/ lowers to /e/ (whether short or long), and /ɨ e o/ lower to /a/, again, whether short or long.
i -> e / _h[C#]
ɨ, e, o -> a / _h[C#]
Example: *kʷēh -> kʷē. (This doesn't look like lowering, but there's an intermediate step *kʷā.)
4b. Merger of coda *h and *ʔ.
*h becomes *ʔ when a coda. Even though neither *h nor *ʔ survived in coda position to the extant language, effects on surrounding consonants and vowels make it clear that this (or something similar) must have occured. Examples aren't going to illustrate the point very well, so I'll just move on.
5. Backing of *i.
*i becomes /ɨ/ when preceded by a velar or labiovelar, that is, one of /k k' kʷ kʷ' ŋ w/.
Example: *mɨkʷik -> mohkʷoh.
6. Loss of *h.
*h is lost in all positions, everywhere.
h -> Ø / _
Pre-Quomonakian Changes Cont., c.-2500 to c. -2000.
Past this point we start to see differentiation of the Quomonakian branch.
7a. Palatalization of alveolar sibilants.
*s *c *c' change to *š *č *č' when followed by *i.
s -> š / _[i ī]
c -> č / _[i ī]
c' -> č' / _[i ī]
7b. Debuccalization of *s.
Remaining *s changes to *h in all positions.
s -> h / _
8a. Simplification of affricates.
*c changes to *s, while *c' and *č change to *c. *č' becomes *č. Schematically:
c -> s / _
c', č -> c / _
č' -> č / _
8b. Palatalization of *t
*t and *t' become *č before *i. This must have occurred after change 7a, since we see č as the modern reflex, not **c.
t, t' -> č / _[i ī]
9. Loss of *θ
*θ becomes /s/ intervocalically and /t/ otherwise.
θ -> s / V_V
θ -> t / otherwise
10. Initial spirantization of stops.
*p *t *k *kʷ become *ɸ *θ *h *w when word-initial and followed by a short vowel.
p -> ɸ / #_V[+short]
t -> θ / #_V[+short]
k -> h / #_V[+short]
kʷ -> w / #_V[+short]
11. Final vowel shortening.
Long vowels become short when word-final.
V: -> V / _#
Pre-Quomonakian to Proto-Quomonakian, c. -2000 to c. -1000
The Quomonakian branch comes into its own. I am considering breaking off a closely-related sister branch (think Baltic to Slavic or Indic to Iranian) somewhere in the middle here.
12. Insertion of /h/ before internal stops and affricates.
An epenthetic /h/ is inserted before intervocalic *p *t *c *č *k *kʷ.
Ø -> h / V_C[+stop/affricate]V
13. Loss of coda *ʔ.
Coda *ʔ, which by now is the only instance of *ʔ, is deleted.
ʔ -> Ø / _
14. Ejectives merge with plain stops.
Remaining ejectives merge into their plain counterparts:
C' -> C / _
15. Palatalization of *k *kʷ
Yet another palatalization rule- here, *k becomes *c, and *kʷ becomes *k, before *i.
k kʷ -> c k / _[i ī]
16a. Shortening of homoorganic vowels.
A long vowel, followed in the next syllable by the same vowel (whether short or long), is shortened. So, *īkʷi -> *ikʷi, and *ōpō -> *opō, but *āpi does not change.
V₁: -> V₁ / _$V₁
16b. Shortening of long vowels between long syllables.
A long vowel, if both the preceding and following syllables contain long vowels, is shortened. This remains a conditioned rule in the language.
V: -> V / V:$_$V:
17. Long Vowel Raising
Long vowels, except *ɨ̄ *ō, raise, with *ī breaking:
ā -> ē / _
ē -> ī / _
ī -> ya / #_
ī -> eya / otherwise
18. Merger of *ɨ
*ɨ merges into *o after labials, and into *a otherwise.
ɨ ɨ̄ -> o ō / [p m w kʷ]_
ɨ ɨ̄ -> a ā / otherwise
19a. Loss of glides before similar vowels
*w is lost before the back vowels *a *o, and *y before the front vowels *e *i:
w -> Ø / _[a ā o ō]
y -> Ø / _[e ē i ī]
19b. Glide reinsertion
Quomonak doesn't like vowel hiatuses, so new glides pretty much immediately get reinserted into the language:
Ø -> w / B_V
Ø -> y / F_V
Proto-Quomonak to Old Quomonak, c. -1000 to c. 300
By this point we're dealing with a branch at about the scale of Italic or Celtic, and changes from here on in detail the more recent development of Old Quomonak itself.
20. *h deletion in double clusters
When a coda *h is followed by a syllable with another coda *h, it gets deleted.
h -> Ø / _CVh[C#]
21. *ŋ mergers
*ŋ merges with *m or *n depending on the following vowel:
*ŋ -> m / _B
*ŋ -> n / _F
22. *ł dentalization
*ł merges with those instances of /θ/ that have developed, generally from initial spirantization.
ł -> θ / _
22a. Coda *k spirantization
*k becomes /h/ in coda:
k -> h / _C, _#
22b. Coda *t to /k/ change
Coda *t becomes /k/ in a chain shift from the preceding change.
t -> k / _C, _#
22c. Coda glide mergers
The coda glides *y *w...oh boy. These merge and often change the preceding vowel in a variety of ways. They're too complicated to list in prose, so I'll just present the changes schematically.
Fy -> ī / _C, _#
iw, ew -> yo / #_[C#]
īw, ēw -> īyo, ēyo / _[C#]
ay, oy -> wi / #_[C#]
ay, oy -> awi, owi / _[C#]
āy, ōy -> āwi, ōwi / _[C#]
Bw -> ō / _C, _#
22d. Loss of nasals in coda with lengthening.
Coda nasals delete, lengthening preceding vowels:
VN -> V: / _C, _#
22e. Approximation of coda *š *l
*š *l become /y w/ in coda:
š -> y / _C, _#
l -> w / _C, _#
22f. Dissimilation of *hh.
Double *h isn't allowed, so new examples of *hh become *hk.
h -> k / h_
23a. Loss of initial *w.
Initial *w is lost.
w -> Ø / #_
23b. Approximation of *ɸ.
*ɸ becomes *w in all positions, likely in a chain shift (since it's basically only found initially.)
ɸ -> w / _
Old Quomonak Nouns
Old Quomonak is generally fairly conservative grammatically (and phonologically): it retains the noun system of PN pretty much intact. Again, this is the template:
[possession]-[prenoun]-[stem]-[obviation/plurality]-[locative, vocative, instrumental]-[possession].
Widespread coda and occasionally initial changes have led to the development of two stems for many nouns. The citation form is usually standalone noun, noun possessed by 1sg argument:
kʷē, na-hkʷīy-ek 'woman'
ese, naw-esey-ek 'dog'
mohkʷoh, na-mokʷohk-et 'cloud'
The possessive prefixes are straightforward. Use the consonants in parentheses before a following vowel.
Prenouns work pretty much the same way as in Proto-Nahtak: ese 'dog' -> ōšiy-ese 'large dog', θalahka 'hill' -> sōwa-salahka 'tall hill'.
Obviation and plurality markers differ in animate and inanimate nouns (most notably in that inanimates can't take obviation marking).
For inanimates, the plural suffix is -ka, attached to the possessed stem. If the possessed stem ends in /k/, just add -a. If it ends in another consonant, the ending is -ihka.
For animates, we have this nice little table:
Note that there are separate endings when the obviation/plurality suffix is word-final as opposed to when it isn't.
Case Markers
It feels kind of like cheating to call these "case markers", since there are only three of them and they don't cover core case functions, just a few oblique functions. However.
Finally, possessive suffixes appear:
Old Quomonak is generally fairly conservative grammatically (and phonologically): it retains the noun system of PN pretty much intact. Again, this is the template:
[possession]-[prenoun]-[stem]-[obviation/plurality]-[locative, vocative, instrumental]-[possession].
Widespread coda and occasionally initial changes have led to the development of two stems for many nouns. The citation form is usually standalone noun, noun possessed by 1sg argument:
kʷē, na-hkʷīy-ek 'woman'
ese, naw-esey-ek 'dog'
mohkʷoh, na-mokʷohk-et 'cloud'
The possessive prefixes are straightforward. Use the consonants in parentheses before a following vowel.
| 1 | na(w)- |
| 2 | θa(w)- |
| 3 PROX | ē(n)- |
| 3 OBV | ō(w)- |
Prenouns work pretty much the same way as in Proto-Nahtak: ese 'dog' -> ōšiy-ese 'large dog', θalahka 'hill' -> sōwa-salahka 'tall hill'.
Obviation and plurality markers differ in animate and inanimate nouns (most notably in that inanimates can't take obviation marking).
For inanimates, the plural suffix is -ka, attached to the possessed stem. If the possessed stem ends in /k/, just add -a. If it ends in another consonant, the ending is -ihka.
For animates, we have this nice little table:
| Singular | Plural | |
|---|---|---|
| Proximate | -Ø | -ēk#/-ēht- |
| Obviative | -ā/-āw- | -(a)ta/-(a)ta- |
Note that there are separate endings when the obviation/plurality suffix is word-final as opposed to when it isn't.
Case Markers
It feels kind of like cheating to call these "case markers", since there are only three of them and they don't cover core case functions, just a few oblique functions. However.
The locative ending is -(i)he in words whose stem's first vowel is /i ī e ē/. If the stem's first vowel is /a ā o ō/, use -(a)ho instead.
The instrumental ending is -eši/-oši, again with the same vowel-harmony rule.
The vocative ending only appears on animates. Word-finally it's -a; when not word-final, -ē-.
Finally, possessive suffixes appear:
When the possessor is singular, the suffix is -ek.
When the possessor is plural and includes a first-person argument, the suffix is -nah. This takes the first-person prefix na(y)- if the possessor is first-person exclusive, and the second-person prefix θa(y)- if the possessor is first-person inclusive.
Other plural possessors take the suffix -(e)h.
OK, gonna give the verbal system a good redo, and the nominal system a bit of a redo. I'm keeping the gender, obviation and direct/inverse distinctions but allowing inanimates to be subjects and going for Inuktitut-style incorporation rather than Algonquian-style. I'm not in the mood to redo the morphosyntax at the moment, though.
OK, a redo on the nouns. The main categories are basically the same as previously, except that inanimates can be marked for obviation and are allowed to possess things.
Also a minor phonological change: I'm no longer going to mark word-initial glottal stop since it isn't really phonemic. I should also probably flesh out word-medial clusters better too but eh.
So, the template for nouns is:
[possessive prefix]-[prenoun]-stem-[obviation/plurality]-[locative/vocative/instrumental]-[possessive suffix]
Possessive affix table:
Though this looks somewhat hairy, the underlying system is fairly simple:
Note additionally that a first person plural inclusive possessor is marked with a second-person prefix but a first-person suffix.
Prenouns are the same as they were during the first draft of noun inflection, so I'm not going over them again.
Obviation and Plurality marking has changed a bit, especially because inanimate nouns can now take obviation marking.
For animates, the obviation/plurality suffixes are as follows:
And for inanimates, these suffixes are used. Note that a large number of inanimate nouns ending in short *e, *a or *ɨ will drop this vowel before endings where possible, even when the first vowel of that ending is marked as dropping after vowels:
doot doot doot
"Case" endings have the last slot (before possessive suffixes, but we've covered those). There are instrumental, locative and vocative endings- only a few oblique functions are marked, not core ones.
That should be it for nouns.
Also a minor phonological change: I'm no longer going to mark word-initial glottal stop since it isn't really phonemic. I should also probably flesh out word-medial clusters better too but eh.
So, the template for nouns is:
[possessive prefix]-[prenoun]-stem-[obviation/plurality]-[locative/vocative/instrumental]-[possessive suffix]
Possessive affix table:
| Sg | Pl | |
|---|---|---|
| 1 | nɨ- | nɨ-...-(e)nēk |
| 21 | N/A | sɨ-...-(e)nēk |
| 2 | sɨ- | sɨ-...-(y)ēwat |
| 3AN.PROX | hā- | hā...-(y)ēwat |
| 3AN.OBV | hā-...-(w)ɨ:na | hā-...-(y)āθat |
| 3IN.PROX | ke- | ke-...(a)htō |
| 3IN.OBV | ke-...(y)isi | ke-...(i)yahtɨ |
Though this looks somewhat hairy, the underlying system is fairly simple:
Prefixes distinguish 1st, 2nd, 3rd animate, and 3rd inanimate persons.
Suffixes distinguish plurality, exclusivity and obviation, with separate suffixes for animates and inanimates. In the singular, the only suffixes that appear are third-person obviative suffixes, with an animate -(w)ɨ:na and inanimate -(y)isi marking obviative possessors. All other suffixes appear with plural possessors, with distinctions in the inanimate between plural proximate (-(a)htō) and obviative -(i)yahtɨ possessors. In the animate, there are three plural possessive suffixes: -(e)nēk, appearing with first person plural possessors, both inclusive and exclusive (in other words, -(e)nēk signals a first-person argument in the possessive); -(y)ēwat marks non-obviative (that is, 2nd person and 3rd person animate proximate) arguments; and -(y)āθat marks an obviative 3rd person plural possessor.
Note additionally that a first person plural inclusive possessor is marked with a second-person prefix but a first-person suffix.
Prenouns are the same as they were during the first draft of noun inflection, so I'm not going over them again.
Obviation and Plurality marking has changed a bit, especially because inanimate nouns can now take obviation marking.
For animates, the obviation/plurality suffixes are as follows:
| Sg | Pl | |
|---|---|---|
| Prox | -Ø | -(e)yon |
| Obv | -(n)āt | -(n)ačo- |
And for inanimates, these suffixes are used. Note that a large number of inanimate nouns ending in short *e, *a or *ɨ will drop this vowel before endings where possible, even when the first vowel of that ending is marked as dropping after vowels:
| Sg | Pl | |
|---|---|---|
| Prox | -Ø | -(ɨ)kʷi |
| Obv | -(e)θak | -(e)θākʷi |
doot doot doot
"Case" endings have the last slot (before possessive suffixes, but we've covered those). There are instrumental, locative and vocative endings- only a few oblique functions are marked, not core ones.
The instrumental ending is -ehpa/-ohpa, with the former ending appearing on nouns whose first syllable contains a front vowel (*i *ī *e *ē), and the latter appearing on those whose first syllable contains a back vowel (*a *ā *o *ō). Whether *ɨ *ɨ: are considered front or back depends on the branch.
The locative ending is -(e)se/-(a)so, also varying by vowel harmony.
The vocative ending, appearing only on animate nouns, is -(h)ā.
That should be it for nouns.
OK, onto verbs. Verbs in Proto-Nahtak still feature noun incorporation, but it's Inuktitut-style rather than Algonquian-style, which means that any noun- once stripped of any inflectional morphology, except prenouns which may stay- can take one of a couple hundred "postbases" which turns it into the object of a transitive verb.
For example, the noun nismā 'potato' can take a verbal postbase -yakʷ- 'eat' to create nismā-yakʷ- 'to eat a potato'. Postbases are recursive: thus -ōθ- 'want' can also be attached to create the stem nismā-yakʷ-ōθ- 'want to eat a potato'. Or a prenoun can be added: šō-nismā-yakʷ-ōθ- 'to want to eat a large potato'.
A verb stem in Proto-Nahtak is considered to be:
- a verbal root, such as toŋ 'drink (smth)', OR
- an incorporative morpheme complex consisting of a noun with at least one postbase and possibly a prenoun, as we've seen above.
Verbal morphology proper can't get inside the verb stem, so for the following verbal templates we will consider it to have one and only one slot. Transitive and intransitive verbs have different templates; when a noun-postbase combination is used with transitive morphology, the noun is usually some sort of oblique argument like an instrumental.
Intransitive Template
[personal prefix]-[verb stem]-["voice"]-[TAM 1]-[personal suffix]-[TAM 2]-(clitics)
Transitive Template
[personal prefix, "primary argument"]-[personal prefix, "secondary argument"]-[verb stem]-["voice"]-[TAM 1]-[number/personal suffix]-[direct/inverse marking]-[TAM 2]-(clitics)
I'd elaborate on this, but it's half past midnight and I have baking to do tomorrow...
For example, the noun nismā 'potato' can take a verbal postbase -yakʷ- 'eat' to create nismā-yakʷ- 'to eat a potato'. Postbases are recursive: thus -ōθ- 'want' can also be attached to create the stem nismā-yakʷ-ōθ- 'want to eat a potato'. Or a prenoun can be added: šō-nismā-yakʷ-ōθ- 'to want to eat a large potato'.
A verb stem in Proto-Nahtak is considered to be:
- a verbal root, such as toŋ 'drink (smth)', OR
- an incorporative morpheme complex consisting of a noun with at least one postbase and possibly a prenoun, as we've seen above.
Verbal morphology proper can't get inside the verb stem, so for the following verbal templates we will consider it to have one and only one slot. Transitive and intransitive verbs have different templates; when a noun-postbase combination is used with transitive morphology, the noun is usually some sort of oblique argument like an instrumental.
Intransitive Template
[personal prefix]-[verb stem]-["voice"]-[TAM 1]-[personal suffix]-[TAM 2]-(clitics)
Transitive Template
[personal prefix, "primary argument"]-[personal prefix, "secondary argument"]-[verb stem]-["voice"]-[TAM 1]-[number/personal suffix]-[direct/inverse marking]-[TAM 2]-(clitics)
I'd elaborate on this, but it's half past midnight and I have baking to do tomorrow...
Couple executive decisions on the phonology: I'm going to kill the ejective series because I hardly ever use it, it feels weird, and most of the wacky sound changes are deriving from clusters and the like anyways (I've started sketching out sound changes for branch II).
An updated cluster table:
[s]Additionally, clusters beginning with /N θ s š/ are allowed word-initially.[/s] No initial clusters.
Allowed word-final consonants are /t k ʔ n s h/.
An updated cluster table:
[s]Additionally, clusters beginning with /N θ s š/ are allowed word-initially.[/s] No initial clusters.
Allowed word-final consonants are /t k ʔ n s h/.
Aaaaand since I'm not in the mood to work on a redo of Old Quomonak's diachronics I'll move onto fucking around with one of the other branches, tentatively referred to as Western Steppe Nahtak. This branch shares a few early changes with Old Quomonakian:
a) the reduction of initial *we/*wɨ, *ye/*yɨ segments to *o *i;
b) the merger of *r and *l into a single phoneme *l;
c) merger of *θC clusters into *sC clusters;
d) loss of *ʔ in *ʔR clusters (where *R = /w r l y/).
Very Early Changes, c. -3000 to c. -2500
Again, many of these changes are shared with other branches, particularly Quomonakian.
1. Initial glide-vowel reduction
*we, *wɨ -> o / #_
*ye, *yɨ -> i / #_
2. *r and *l merger
*r -> l / _
3. Merger of *θ and *s in clusters
*θ -> s / _C
Resulting *sč becomes *št; *sc becomes *st.
4. Loss of *ʔ in clusters with glides
ʔ -> Ø / _R (R=w, y, l)
5. Rhinoglottophilia
In an initial syllable of the sort
#hV₁CV₂
where C is not part of a cluster and isn't a nasal, a nasal is inserted before C:
hV₁CV₂ -> hV₁NCV₂ / #_
This gives rise to new nasal-fricative clusters *nθ *ns *nš *nh *nł. *ns and *nš merge into *nc and *nč immediately; *nł merges with *nl, leaving *nθ and *nh as the only nasal-fricative clusters going forward. The cluster *nʔ immediately metathesizes to *ʔn.
In initial syllables of the sort
#hV₁NV₂
where N = *m *n *ŋ, no additional nasal is inserted, but V₁ is lengthened.
6. Loss of initial *h.
h -> Ø / #_
Proto-Nahtak to Pre-Proto-Western Steppe, c. -2500 to c. -1800
Branch differentiation becomes more concrete at this stage.
7. Loss of *š.
*š becomes *x before back vowels (*a *o *ɨ) and in clusters; otherwise it merges with *s:
*š -> x / _B, _C
*š -> s / otherwise
8. *a to *o shift after labials
After labial and labialized consonants- that is, any of *p *m *w *kʷ- *a *ā become *o *ō:
*a *ā -> o ō / [p m w kʷ]_
9. Gemination in clusters with *ʔ
In any cluster *ʔC, *ʔ is lost and the following consonant is geminated:
ʔC -> C: / _
10. Loss of *h with compensatory lengthening; loss of *ʔ
*h is lost in all remaining positions. When a coda, it lengthens any preceding vowel:
Vh -> V: / _C, _#
h -> Ø / _
Remaining *ʔ is lost as well:
ʔ -> Ø / _
11. Reappearence of /š/; loss of *č and *ł
/š/ reappears from two sources: *č, except after a nasal, and *ł initially:
č -> š / _ ! N_
ł -> š / #_
*ł merges into /l/ otherwise:
ł -> l / _
Pre-Proto-Western Steppe to Early Proto-Western Steppe, c. -1800 to c. -750
Sometime in the early-to-mid negative second millennium, two major changes occurred which mark the emergence of a defined Western Steppe branch: "Western Steppe Umlaut", a vowel assimilation process, and the "Western Steppe Velar Shift". There are also some hard-to-date cluster changes stuck in here, too.
12. Western Steppe Umlaut
Word-final vowels cause vowels in the preceding syllable to change. A large number of words are affected, making it difficult to identify vowel correspondences with other branches.
Length was conserved but played no role in the details of the shift (that is, e. g., where *a shifted to *e, *ā likewise shifted to *ē), and will not be marked below. <$> marks a syllable boundary.
a) Changes before final *i
Before a word-final *i, the following changes occurred:
*ɨ ->i / _$i#
*e -> i / _$i#
*a -> e / _$i#
*o -> ø / _$i# (*ø is almost immediately broken into /(w)e/, as we shall see, but is reconstructed as a necessary intermediate)
b) Changes before final *o
Before a word-final *o, the following changes occurred:
*i -> ɨ / _$o#
*ɨ -> a / _$o#
*a -> o / _$o#
c) Change before final *a
Only *i was affected by a word-final *a:
*i -> e / _$a#
13. Breaking of *ø
New umlauted *ø doesn't survive long, and quickly unpacks as follows:
ø ø: -> we wē / V[t c k n θ s š h]_, #_, #k_
ø ø: -> e ē / otherwise
Resulting -kw- became kʷ, and -cw- became -sw-.
14. Western Steppe Velar Shift
Briefly speaking, *k shifted to either /ʔ/ or /x/ except as the second member of a cluster, and was replaced by *kʷ (medially) and *ŋ (initially); initial *kʷ became /p/ and medial *ŋ became *n.
*k -> ʔ / V_V
*k -> x / #_, _C, _#
*k maintained / C_ or as a geminate
*ŋ -> k / #_
*ŋ -> n / otherwise
*kʷ -> p / # _
*kʷ -> k / otherwise
14. Merger of *c and *č
*c becomes *č in all positions:
*c -> č / _
15. Loss of glides in double-sonorant clusters
The glides *w, *l and *y are lost when they are the second member of a cluster and follow a sonorant:
R -> Ø / N_, R_
(But they are reinserted, with epenthetic ă, to analogize some paradigms, such as in the 3rd animate possessive of nouns.)
16. Spirantization of *t in clusters
*t becomes /θ/ when it is the first member of a cluster.
t -> θ / _C
17. Frication of *l in clusters.
*l becomes *h when it is the first member of a cluster. It's likely that this took the form *l -> r -> R -> h or something of that nature; it's difficult to say.
*l -> h / _C
18. Initial *y split.
*y is lost initially before front vowels *e *i, and becomes /h/ initially otherwise:
y -> Ø / #_F
y -> h / #_ otherwise
Early Proto-Western-Steppe to Late Proto-Western-Steppe, c. -750 to -250
Most Western Steppe languages split off towards the end of this list of final sound changes, but at least one will be split off before the short vowel collapse and subsequent restructuring of the vowel system.
19. Long Vowel Shortening
Long vowels shorten before geminates:
V: -> V / _C₁C₁
20. Short Vowel Lengthening
...just to confuse you. The second of a sequence of three consecutive short vowels, if it is not followed by a cluster, lengthens:
V -> V: / V[+short]$_$V[+short] ! _CC
21. Short Vowel Collapse
The short vowels undergo mergers, trimming their numbers from five to three:
*ŏ-> a / _
*ă *ĕ -> ə / _
*ĭ ɨ -> i /_
21. Various Long Vowel Changes
a) *ē and *ī merge as ī, except before *h, *w or *x, where they merge as ē:
*ī *ē -> ē / _[h, w, x]
*ī *ē -> ī / otherwise
b) *ɨ: merges with *ī after *č *š and with *ē otherwise:
ɨ: -> ī / [č, š]_
ɨ: -> ē / otherwise
22a. Syncope
If a short vowel- that's /a ə i/, now-
it will delete. Any resulting cluster must:
If it fails to meet one of these criteria, it's broken up: CV[+short]C -> !CC -> CəC.
In some words the resulting epenthesis of /ə/ makes it hard to tell that there was any syncope at all; however, because clusters are not allowed word-finally, a word-final cluster (including geminates) resulting from the deletion of a word-final short vowel will ALWAYS be epenthesized.
22b. Word-final consonant changes resulting from syncope
Syncope leaves a lot of new word-final consonants, which undergo a few changes:
- final *č becomes /š/:
č -> š / _#
- final *t becomes /ʔ/; final *θ becomes /t/ in its place:
t -> ʔ / _#
θ -> t / _#
- final *l merges with *w if it's preceded by /ə/:
l -> w / ə_#
- the segments *əy/*ĭy and *əw/*ăw become /ī/ and /ō/:
əy, iy -> ī / _#
əw, aw -> ō / _#
- *w and *y otherwise delete at the end of a word.
23. Final vowel shift.
A few Western Steppe languages do not exhibit this change, which destroyed the last vestiges of a length distinction.
*ĭ -> e
*ī -> i
*ē -> ɛ: -> ɛ
*ā -> ɔ: -> o: -> o
*ō -> u: -> u
*ă -> ɑ -> ɔ
*ə -> a
That should be it for the sound changes. I'll try and get the revamped Quomonak changes done this evening if I can.
The resulting Proto-Western-Steppe inventory should be:
Stops/Affricates: *p *t *č *k *ʔ
Nasals: *m *n
Fricatives: *θ *s *š *x *h
Liquids: *w *l *y
Vowels: *i *e *ɛ *a *ɔ *o *u
a) the reduction of initial *we/*wɨ, *ye/*yɨ segments to *o *i;
b) the merger of *r and *l into a single phoneme *l;
c) merger of *θC clusters into *sC clusters;
d) loss of *ʔ in *ʔR clusters (where *R = /w r l y/).
Very Early Changes, c. -3000 to c. -2500
Again, many of these changes are shared with other branches, particularly Quomonakian.
1. Initial glide-vowel reduction
*we, *wɨ -> o / #_
*ye, *yɨ -> i / #_
2. *r and *l merger
*r -> l / _
3. Merger of *θ and *s in clusters
*θ -> s / _C
Resulting *sč becomes *št; *sc becomes *st.
4. Loss of *ʔ in clusters with glides
ʔ -> Ø / _R (R=w, y, l)
5. Rhinoglottophilia
In an initial syllable of the sort
#hV₁CV₂
where C is not part of a cluster and isn't a nasal, a nasal is inserted before C:
hV₁CV₂ -> hV₁NCV₂ / #_
This gives rise to new nasal-fricative clusters *nθ *ns *nš *nh *nł. *ns and *nš merge into *nc and *nč immediately; *nł merges with *nl, leaving *nθ and *nh as the only nasal-fricative clusters going forward. The cluster *nʔ immediately metathesizes to *ʔn.
In initial syllables of the sort
#hV₁NV₂
where N = *m *n *ŋ, no additional nasal is inserted, but V₁ is lengthened.
6. Loss of initial *h.
h -> Ø / #_
Proto-Nahtak to Pre-Proto-Western Steppe, c. -2500 to c. -1800
Branch differentiation becomes more concrete at this stage.
7. Loss of *š.
*š becomes *x before back vowels (*a *o *ɨ) and in clusters; otherwise it merges with *s:
*š -> x / _B, _C
*š -> s / otherwise
8. *a to *o shift after labials
After labial and labialized consonants- that is, any of *p *m *w *kʷ- *a *ā become *o *ō:
*a *ā -> o ō / [p m w kʷ]_
9. Gemination in clusters with *ʔ
In any cluster *ʔC, *ʔ is lost and the following consonant is geminated:
ʔC -> C: / _
10. Loss of *h with compensatory lengthening; loss of *ʔ
*h is lost in all remaining positions. When a coda, it lengthens any preceding vowel:
Vh -> V: / _C, _#
h -> Ø / _
Remaining *ʔ is lost as well:
ʔ -> Ø / _
11. Reappearence of /š/; loss of *č and *ł
/š/ reappears from two sources: *č, except after a nasal, and *ł initially:
č -> š / _ ! N_
ł -> š / #_
*ł merges into /l/ otherwise:
ł -> l / _
Pre-Proto-Western Steppe to Early Proto-Western Steppe, c. -1800 to c. -750
Sometime in the early-to-mid negative second millennium, two major changes occurred which mark the emergence of a defined Western Steppe branch: "Western Steppe Umlaut", a vowel assimilation process, and the "Western Steppe Velar Shift". There are also some hard-to-date cluster changes stuck in here, too.
12. Western Steppe Umlaut
Word-final vowels cause vowels in the preceding syllable to change. A large number of words are affected, making it difficult to identify vowel correspondences with other branches.
Length was conserved but played no role in the details of the shift (that is, e. g., where *a shifted to *e, *ā likewise shifted to *ē), and will not be marked below. <$> marks a syllable boundary.
a) Changes before final *i
Before a word-final *i, the following changes occurred:
*ɨ ->i / _$i#
*e -> i / _$i#
*a -> e / _$i#
*o -> ø / _$i# (*ø is almost immediately broken into /(w)e/, as we shall see, but is reconstructed as a necessary intermediate)
b) Changes before final *o
Before a word-final *o, the following changes occurred:
*i -> ɨ / _$o#
*ɨ -> a / _$o#
*a -> o / _$o#
c) Change before final *a
Only *i was affected by a word-final *a:
*i -> e / _$a#
13. Breaking of *ø
New umlauted *ø doesn't survive long, and quickly unpacks as follows:
ø ø: -> we wē / V[t c k n θ s š h]_, #_, #k_
ø ø: -> e ē / otherwise
Resulting -kw- became kʷ, and -cw- became -sw-.
14. Western Steppe Velar Shift
Briefly speaking, *k shifted to either /ʔ/ or /x/ except as the second member of a cluster, and was replaced by *kʷ (medially) and *ŋ (initially); initial *kʷ became /p/ and medial *ŋ became *n.
*k -> ʔ / V_V
*k -> x / #_, _C, _#
*k maintained / C_ or as a geminate
*ŋ -> k / #_
*ŋ -> n / otherwise
*kʷ -> p / # _
*kʷ -> k / otherwise
14. Merger of *c and *č
*c becomes *č in all positions:
*c -> č / _
15. Loss of glides in double-sonorant clusters
The glides *w, *l and *y are lost when they are the second member of a cluster and follow a sonorant:
R -> Ø / N_, R_
(But they are reinserted, with epenthetic ă, to analogize some paradigms, such as in the 3rd animate possessive of nouns.)
16. Spirantization of *t in clusters
*t becomes /θ/ when it is the first member of a cluster.
t -> θ / _C
17. Frication of *l in clusters.
*l becomes *h when it is the first member of a cluster. It's likely that this took the form *l -> r -> R -> h or something of that nature; it's difficult to say.
*l -> h / _C
18. Initial *y split.
*y is lost initially before front vowels *e *i, and becomes /h/ initially otherwise:
y -> Ø / #_F
y -> h / #_ otherwise
Early Proto-Western-Steppe to Late Proto-Western-Steppe, c. -750 to -250
Most Western Steppe languages split off towards the end of this list of final sound changes, but at least one will be split off before the short vowel collapse and subsequent restructuring of the vowel system.
19. Long Vowel Shortening
Long vowels shorten before geminates:
V: -> V / _C₁C₁
20. Short Vowel Lengthening
...just to confuse you. The second of a sequence of three consecutive short vowels, if it is not followed by a cluster, lengthens:
V -> V: / V[+short]$_$V[+short] ! _CC
21. Short Vowel Collapse
The short vowels undergo mergers, trimming their numbers from five to three:
*ŏ-> a / _
*ă *ĕ -> ə / _
*ĭ ɨ -> i /_
21. Various Long Vowel Changes
a) *ē and *ī merge as ī, except before *h, *w or *x, where they merge as ē:
*ī *ē -> ē / _[h, w, x]
*ī *ē -> ī / otherwise
b) *ɨ: merges with *ī after *č *š and with *ē otherwise:
ɨ: -> ī / [č, š]_
ɨ: -> ē / otherwise
22a. Syncope
If a short vowel- that's /a ə i/, now-
is in the second syllable of a word AND
is either word-final, OR
is within a word but is neither preceded nor followed by a cluster
it will delete. Any resulting cluster must:
consist of a nasal plus a stop or one of /θ h/ OR
consist of a fricative (/θ s š x h/, just to refresh your memory) plus a sound that isn't a fricative OR
be a geminate consonant
If it fails to meet one of these criteria, it's broken up: CV[+short]C -> !CC -> CəC.
In some words the resulting epenthesis of /ə/ makes it hard to tell that there was any syncope at all; however, because clusters are not allowed word-finally, a word-final cluster (including geminates) resulting from the deletion of a word-final short vowel will ALWAYS be epenthesized.
22b. Word-final consonant changes resulting from syncope
Syncope leaves a lot of new word-final consonants, which undergo a few changes:
- final *č becomes /š/:
č -> š / _#
- final *t becomes /ʔ/; final *θ becomes /t/ in its place:
t -> ʔ / _#
θ -> t / _#
- final *l merges with *w if it's preceded by /ə/:
l -> w / ə_#
- the segments *əy/*ĭy and *əw/*ăw become /ī/ and /ō/:
əy, iy -> ī / _#
əw, aw -> ō / _#
- *w and *y otherwise delete at the end of a word.
23. Final vowel shift.
A few Western Steppe languages do not exhibit this change, which destroyed the last vestiges of a length distinction.
*ĭ -> e
*ī -> i
*ē -> ɛ: -> ɛ
*ā -> ɔ: -> o: -> o
*ō -> u: -> u
*ă -> ɑ -> ɔ
*ə -> a
That should be it for the sound changes. I'll try and get the revamped Quomonak changes done this evening if I can.
The resulting Proto-Western-Steppe inventory should be:
Stops/Affricates: *p *t *č *k *ʔ
Nasals: *m *n
Fricatives: *θ *s *š *x *h
Liquids: *w *l *y
Vowels: *i *e *ɛ *a *ɔ *o *u
A crappy map of the major branches of Nahtak.
Tomorrow I'll probably make a diagram of the major branches with isoglosses. Thinner lines represent closer relationships, so that the Thanahapic languages are more closely related to the Western Steppe languages than they are to the Inland Quomonak languages. This doesn't show isoglosses that stretch across subfamilies or are broken by an intervening family. For example, while it's true that Quomonakian and the Western branch have several similarities and are more closely related to each other than either is to, say, Omashaloan, the Southern Coast family is a (divergent, but verifiable) member of the Central branch. Moreover, while the similarities within each larger group are usually tight enough to justify the classification, there is some cross-contamination: for example, while almost all of Quomonakian, Western and Central all share the change *r -> *l, Thanahapic instead merges *l with *ł and keeps *r separate, a state of affairs otherwise only found in Kothavasnic- but it is clearly close enough to the rest of Western to justify considering it as a member.
The Urheimat of the Nahtak languages is probably near the north coast of the Gulf of the East, roughly about the modern border between the Hovapaquequian and Soric languages on the map.
It is likely that my sense of scale is totally off with this map, but whatever. This map also does not represent the modern-day state of affairs, but rather the situation in the early centuries of the 1st millenium, about 0-250. (I am not entirely sure what a modern map would show. It's quite likely that Coastal Quomonak would encompass a far greater range than it currently does, that the Western Steppe languages would have descendents scattered as far as the Mediterranean climate south of the Insular Omashaloan territory on the map, and a couple branches- Thanhapic? North Coast?- would not appear at all.
Shitty map again with macro-branches labeled:
Tomorrow I'll probably make a diagram of the major branches with isoglosses. Thinner lines represent closer relationships, so that the Thanahapic languages are more closely related to the Western Steppe languages than they are to the Inland Quomonak languages. This doesn't show isoglosses that stretch across subfamilies or are broken by an intervening family. For example, while it's true that Quomonakian and the Western branch have several similarities and are more closely related to each other than either is to, say, Omashaloan, the Southern Coast family is a (divergent, but verifiable) member of the Central branch. Moreover, while the similarities within each larger group are usually tight enough to justify the classification, there is some cross-contamination: for example, while almost all of Quomonakian, Western and Central all share the change *r -> *l, Thanahapic instead merges *l with *ł and keeps *r separate, a state of affairs otherwise only found in Kothavasnic- but it is clearly close enough to the rest of Western to justify considering it as a member.
The Urheimat of the Nahtak languages is probably near the north coast of the Gulf of the East, roughly about the modern border between the Hovapaquequian and Soric languages on the map.
It is likely that my sense of scale is totally off with this map, but whatever. This map also does not represent the modern-day state of affairs, but rather the situation in the early centuries of the 1st millenium, about 0-250. (I am not entirely sure what a modern map would show. It's quite likely that Coastal Quomonak would encompass a far greater range than it currently does, that the Western Steppe languages would have descendents scattered as far as the Mediterranean climate south of the Insular Omashaloan territory on the map, and a couple branches- Thanhapic? North Coast?- would not appear at all.
Shitty map again with macro-branches labeled:
Sound changes for Quomonakian, revamped. Also I'm clearly going to need to redo the map, but never mind that...
Very Early Changes, c.-3000 to c.-2500
Many of these changes can be found in other branches as well.
1. Initial glide-vowel reduction
The weak vowels /e ɨ/ merge with a preceding glide into the vocalic version of that glide when word-initial. Schematically:
we wɨ -> o / #_
ye yɨ -> i / #_
So, *weyɨ̄ *yente -> owi īte.
2. Loss of *ʔ
The glottal stop *ʔ is lost before a vowel or a glide:
ʔ -> Ø / _V, _R
Examples include *nɨʔo -> niyo and *ʔočīča -> ohceyahca.
3. Merger of *r and *l.
r -> l / _
(However, [r] was probably retained as an allophone of /l/ when the first member of a cluster.)
4. Merger of *θ and *s in clusters.
When it's in a cluster (always as the first member), *θ becomes /s/.
θ -> s / _C
5a. Vowel lowering before coda *h.
Before coda /h/, the vowel /i/ lowers to /e/ (whether short or long), and /ɨ e o/ lower to /a/, again, whether short or long.
i -> e / _h[C#]
ɨ, e, o -> a / _h[C#]
Example: *kʷēh -> kʷē. (This doesn't look like lowering, but there's an intermediate step *kʷā.)
5b. Merger of coda *h and *ʔ.
*h becomes *ʔ when a coda. Even though neither *h nor *ʔ survived in coda position to the extant language, effects on surrounding consonants and vowels make it clear that this (or something similar) must have occured. Examples aren't going to illustrate the point very well, so I'll just move on.
6. Backing of *i.
*i becomes /ɨ/ when preceded by a velar or labiovelar, that is, one of /k kʷ ŋ w/.
Example: *mɨkʷik -> mohkʷoh.
7. Loss of *h.
*h is lost in all positions, everywhere.
h -> Ø / _
Pre-Quomonakian Changes Cont., c.-2500 to c. -2000.
Past this point we start to see differentiation of the Quomonakian branch.
8a. Palatalization of alveolar sibilants.
*s *c change to *š *č when followed by *i.
s -> š / _[i ī]
c -> č / _[i ī]
8b. Debuccalization of *s.
Remaining *s changes to *h in all positions.
s -> h / _
9. Simplification of affricates.
*c changes to *s, while *č becomes *c (except as the initial member of a cluster, where it merges with *š.).
c -> s / _
č -> c / _V
č -> š / _C
10. Loss of *θ
*θ merges with *s before front vowels and *t otherwise:
*θ -> s / _F
*θ -> t / otherwise
11. Initial spirantization of stops.
*p *t *k *kʷ become *ɸ *θ *h *w when word-initial and followed by a short vowel.
p -> ɸ / #_V[+short]
t -> θ / #_V[+short]
k -> h / #_V[+short]
kʷ -> w / #_V[+short]
12. Final vowel shortening.
Long vowels become short when word-final.
V: -> V / _#
Pre-Quomonakian to Proto-Quomonakian, c. -2000 to c. -700
The Quomonakian branch comes into its own. I am considering breaking off a closely-related sister branch (think Baltic to Slavic or Indic to Iranian) somewhere in the middle here.
13. Insertion of /h/ before internal stops.
An epenthetic /h/ is inserted before intervocalic *p *t *k *kʷ, if the preceding vowel is short.
Ø -> h / V[+short]_C[+stop]V
14. Loss of coda *ʔ.
Coda *ʔ, which by now is the only instance of *ʔ, is deleted.
ʔ -> Ø / _
15. Loss of initial *w.
Initial *w becomes *y before front vowels (*i *e) and is otherwise deleted.
w -> y / #_F
w -> Ø / #_B
16a. Shortening of homoorganic vowels.
A long vowel, followed in the next syllable by the same vowel (whether short or long), is shortened. So, *īkʷi -> *ikʷi, and *ōpō -> *opō, but *āpi does not change.
V₁: -> V₁ / _$V₁
16b. Shortening of long vowels between long syllables.
A long vowel, if both the preceding and following syllables contain long vowels, is shortened. This remains a conditioned rule in the language.
V: -> V / V:$_$V:
16c. Shortening of long vowels in cluster situations.
A long vowel is shortened if it's followed by a cluster which is itself followed by a long vowel.
V: -> V / _CCV:
17. Long Vowel Raising
Long vowels, except *ɨ̄ *ō, raise, with *ī breaking:
ā -> ē / _
ē -> ī / _
ī -> ya / #_
ī -> eya / otherwise
18. *š depalatalization in clusters.
When *š is the first member of a cluster, it becomes *s.
š -> s / _C
19. Merger of *ɨ
*ɨ merges into *o after labials, and into *a otherwise.
ɨ ɨ̄ -> o ō / [p m w kʷ ɸ]_
ɨ ɨ̄ -> a ā / otherwise
20. Coda *k spirantization
*k becomes /h/ in coda:
k -> h / _C, _#
21a. Loss of glides before similar vowels
*w is lost intervocalically before the back vowels *a *o, and *y before the front vowels *e *i:
w -> Ø / V_[a ā o ō]
y -> Ø / V_[e ē i ī]
21b. Glide reinsertion
Quomonak doesn't like vowel hiatuses, so new glides pretty much immediately get reinserted into the language:
Ø -> w / B_V
Ø -> y / F_V
22. Nasal deletion (with lengthening) in coda
When a nasal is the first member of a cluster or is word-final, it is deleted, with any preceding vowel lengthened:
Vn -> V: / _[C#]
Proto-Quomonak to Old Quomonak, c. -700 to c. 300
By this point we're dealing with a branch at about the scale of Italic or Celtic, and changes from here on in detail the more recent development of Old Quomonak itself.
23. *h deletion in double clusters
When a coda *h is followed by a syllable with another coda *h, it gets deleted.
h -> Ø / _CVh[C#]
24. Approximation of *ɸ.
*ɸ becomes *w in all positions (it's basically only found initially.)
ɸ -> w / _
25. *ŋ mergers
*ŋ merges with *m or *n depending on the following vowel:
*ŋ -> m / _B
*ŋ -> n / _F
26. *l to N in clusters
*l (probably [r] or [ɾ] phonetically) becomes an assimilatory nasal when the first member of a cluster:
*l -> N / _C
27. *ł dentalization
*ł merges with those instances of /θ/ that have developed, generally from initial spirantization of *t.
ł -> θ / _
28. Coda *t to /k/ change
Coda *t becomes /k/.
t -> k / _C, _#
Very Early Changes, c.-3000 to c.-2500
Many of these changes can be found in other branches as well.
1. Initial glide-vowel reduction
The weak vowels /e ɨ/ merge with a preceding glide into the vocalic version of that glide when word-initial. Schematically:
we wɨ -> o / #_
ye yɨ -> i / #_
So, *weyɨ̄ *yente -> owi īte.
2. Loss of *ʔ
The glottal stop *ʔ is lost before a vowel or a glide:
ʔ -> Ø / _V, _R
Examples include *nɨʔo -> niyo and *ʔočīča -> ohceyahca.
3. Merger of *r and *l.
r -> l / _
(However, [r] was probably retained as an allophone of /l/ when the first member of a cluster.)
4. Merger of *θ and *s in clusters.
When it's in a cluster (always as the first member), *θ becomes /s/.
θ -> s / _C
5a. Vowel lowering before coda *h.
Before coda /h/, the vowel /i/ lowers to /e/ (whether short or long), and /ɨ e o/ lower to /a/, again, whether short or long.
i -> e / _h[C#]
ɨ, e, o -> a / _h[C#]
Example: *kʷēh -> kʷē. (This doesn't look like lowering, but there's an intermediate step *kʷā.)
5b. Merger of coda *h and *ʔ.
*h becomes *ʔ when a coda. Even though neither *h nor *ʔ survived in coda position to the extant language, effects on surrounding consonants and vowels make it clear that this (or something similar) must have occured. Examples aren't going to illustrate the point very well, so I'll just move on.
6. Backing of *i.
*i becomes /ɨ/ when preceded by a velar or labiovelar, that is, one of /k kʷ ŋ w/.
Example: *mɨkʷik -> mohkʷoh.
7. Loss of *h.
*h is lost in all positions, everywhere.
h -> Ø / _
Pre-Quomonakian Changes Cont., c.-2500 to c. -2000.
Past this point we start to see differentiation of the Quomonakian branch.
8a. Palatalization of alveolar sibilants.
*s *c change to *š *č when followed by *i.
s -> š / _[i ī]
c -> č / _[i ī]
8b. Debuccalization of *s.
Remaining *s changes to *h in all positions.
s -> h / _
9. Simplification of affricates.
*c changes to *s, while *č becomes *c (except as the initial member of a cluster, where it merges with *š.).
c -> s / _
č -> c / _V
č -> š / _C
10. Loss of *θ
*θ merges with *s before front vowels and *t otherwise:
*θ -> s / _F
*θ -> t / otherwise
11. Initial spirantization of stops.
*p *t *k *kʷ become *ɸ *θ *h *w when word-initial and followed by a short vowel.
p -> ɸ / #_V[+short]
t -> θ / #_V[+short]
k -> h / #_V[+short]
kʷ -> w / #_V[+short]
12. Final vowel shortening.
Long vowels become short when word-final.
V: -> V / _#
Pre-Quomonakian to Proto-Quomonakian, c. -2000 to c. -700
The Quomonakian branch comes into its own. I am considering breaking off a closely-related sister branch (think Baltic to Slavic or Indic to Iranian) somewhere in the middle here.
13. Insertion of /h/ before internal stops.
An epenthetic /h/ is inserted before intervocalic *p *t *k *kʷ, if the preceding vowel is short.
Ø -> h / V[+short]_C[+stop]V
14. Loss of coda *ʔ.
Coda *ʔ, which by now is the only instance of *ʔ, is deleted.
ʔ -> Ø / _
15. Loss of initial *w.
Initial *w becomes *y before front vowels (*i *e) and is otherwise deleted.
w -> y / #_F
w -> Ø / #_B
16a. Shortening of homoorganic vowels.
A long vowel, followed in the next syllable by the same vowel (whether short or long), is shortened. So, *īkʷi -> *ikʷi, and *ōpō -> *opō, but *āpi does not change.
V₁: -> V₁ / _$V₁
16b. Shortening of long vowels between long syllables.
A long vowel, if both the preceding and following syllables contain long vowels, is shortened. This remains a conditioned rule in the language.
V: -> V / V:$_$V:
16c. Shortening of long vowels in cluster situations.
A long vowel is shortened if it's followed by a cluster which is itself followed by a long vowel.
V: -> V / _CCV:
17. Long Vowel Raising
Long vowels, except *ɨ̄ *ō, raise, with *ī breaking:
ā -> ē / _
ē -> ī / _
ī -> ya / #_
ī -> eya / otherwise
18. *š depalatalization in clusters.
When *š is the first member of a cluster, it becomes *s.
š -> s / _C
19. Merger of *ɨ
*ɨ merges into *o after labials, and into *a otherwise.
ɨ ɨ̄ -> o ō / [p m w kʷ ɸ]_
ɨ ɨ̄ -> a ā / otherwise
20. Coda *k spirantization
*k becomes /h/ in coda:
k -> h / _C, _#
21a. Loss of glides before similar vowels
*w is lost intervocalically before the back vowels *a *o, and *y before the front vowels *e *i:
w -> Ø / V_[a ā o ō]
y -> Ø / V_[e ē i ī]
21b. Glide reinsertion
Quomonak doesn't like vowel hiatuses, so new glides pretty much immediately get reinserted into the language:
Ø -> w / B_V
Ø -> y / F_V
22. Nasal deletion (with lengthening) in coda
When a nasal is the first member of a cluster or is word-final, it is deleted, with any preceding vowel lengthened:
Vn -> V: / _[C#]
Proto-Quomonak to Old Quomonak, c. -700 to c. 300
By this point we're dealing with a branch at about the scale of Italic or Celtic, and changes from here on in detail the more recent development of Old Quomonak itself.
23. *h deletion in double clusters
When a coda *h is followed by a syllable with another coda *h, it gets deleted.
h -> Ø / _CVh[C#]
24. Approximation of *ɸ.
*ɸ becomes *w in all positions (it's basically only found initially.)
ɸ -> w / _
25. *ŋ mergers
*ŋ merges with *m or *n depending on the following vowel:
*ŋ -> m / _B
*ŋ -> n / _F
26. *l to N in clusters
*l (probably [r] or [ɾ] phonetically) becomes an assimilatory nasal when the first member of a cluster:
*l -> N / _C
27. *ł dentalization
*ł merges with those instances of /θ/ that have developed, generally from initial spirantization of *t.
ł -> θ / _
28. Coda *t to /k/ change
Coda *t becomes /k/.
t -> k / _C, _#
[EDIT: Killing the aspirates- the phonology now looks more like Salishan than Algonquian. Also, labialized consonants aren't allowed as the first member of clusters.]
OK, I'm increasing the time depth on this family to somewhere around 7000 years, which means that the Algonquian-esque family- Insular Nahtak- is just going to be a sub-branch, albeit a major one. The actual proto-language (which looks a lot more like Athabaskan than Algonquian) has the following phonology:
Notes on transcription: a comma above indicates an ejective or glottalized consonant; otherwise, a generally Americanist transcription is used, so <c č> are /ts tʃ/, macrons mark a long vowel, and nasal vowels are indicated with an ogonek.
The back vowels *o *u and their long/nasal versions were probably [ɤ ɯ] except after labialized consonants.
The syllable structure is (C)V(C), but this obscures the fact that the phonotactics are perhaps best understood on the level of the word. Words may begin with no more than one consonant (but are allowed to have no onset consonant at all); they must end with a vowel; and may contain intervocalic clusters of two consonants (no more). Vowel hiatuses are verboten.
Clusters
- The first member of such a cluster may not be labialized or glottalized.
- *t may only precede dorsals (no coronals.)
- A dorsal stop may not precede another dorsal.
- An affricate must precede a true stop (not an affricate or a fricative).
- The initial fricative may not be labialized;
- A sibilant (*s or*š) may not precede an affricate;
- The second member may not be a velar nasal;
- velars and uvulars may not form "mixed clusters" (thus, -xkʷ- and -χq̉- are allowed, but not -xqʷ- or -χk̉-).
OK, I'm increasing the time depth on this family to somewhere around 7000 years, which means that the Algonquian-esque family- Insular Nahtak- is just going to be a sub-branch, albeit a major one. The actual proto-language (which looks a lot more like Athabaskan than Algonquian) has the following phonology:
Notes on transcription: a comma above indicates an ejective or glottalized consonant; otherwise, a generally Americanist transcription is used, so <c č> are /ts tʃ/, macrons mark a long vowel, and nasal vowels are indicated with an ogonek.
Plain stops/affricates: *t *tʷ *c *cʷ *č *čʷ *k *kʷ *q *qʷ
Ejective stops/affricates: *t̉ *t̉ʷ *c̉ *c̉ʷ *č̉ *č̉ʷ *k̉ *k̉ʷ *q̉ *q̉ʷ
Nasals: *n *nʷ *ŋ *ŋʷ
Glottalized nasals: *n̉ *n̉ʷ *ŋ̉ *ŋ̉ʷ
Fricatives: *θ *θʷ *s *sʷ *š *šʷ *x *xʷ *χ *χʷ
Liquids: *r *l *y *w
Glottalized liquids: *r̉ *l̉ *ỷ *w̉
Short vowels: *a *e *i *o *u
Long vowels: *ā *ē *ī *ō *ū
Nasal vowels: *ą *ę *į *ǫ *ų
The back vowels *o *u and their long/nasal versions were probably [ɤ ɯ] except after labialized consonants.
The syllable structure is (C)V(C), but this obscures the fact that the phonotactics are perhaps best understood on the level of the word. Words may begin with no more than one consonant (but are allowed to have no onset consonant at all); they must end with a vowel; and may contain intervocalic clusters of two consonants (no more). Vowel hiatuses are verboten.
Clusters
Labialized consonants may never be the first member of a cluster. Glottalized consonants can only be the first member of a cluster in consonant-liquid clusters.
A nasal element N may precede any consonant except another nasal or a sibilant (*s *sʷ *š *šʷ). N is realized as *ŋ before dorsals and *n before coronals and does not distinguish glottalization except perhaps allophonically. The sibilants *s *sʷ *š *šʷ are not allowed as the second element of such a cluster and, when such an occasion would arise, they become affricates *c *cʷ *č *čʷ. Nasal vowels can't occur before these clusters (that is, Proto-Nahtak distinguishes VNC from ṼC, but not from *ṼNC.)
A plain stop or affricate may precede an obstruent (another stop- whether plain, aspirated or ejective- or a fricative) with some exceptions:
- The first member of such a cluster may not be labialized or glottalized.
- *t may only precede dorsals (no coronals.)
- A dorsal stop may not precede another dorsal.
- An affricate must precede a true stop (not an affricate or a fricative).
A fricative may precede a non-fricative consonant of any sort, with the following exceptions:
- The initial fricative may not be labialized;
- A sibilant (*s or*š) may not precede an affricate;
- The second member may not be a velar nasal;
- velars and uvulars may not form "mixed clusters" (thus, -xkʷ- and -χq̉- are allowed, but not -xqʷ- or -χk̉-).
*r or *l (not glottalized) may precede any other consonant other than another liquid.
A non-glottalized liquid other than *w (that is, *l *y *r) may follow any consonant other than another liquid.
OK, instead of going onto the morphology, we're going to look at roots. The prototypical root in Proto-Nahtak is (C)VC and is constructed from a reduced inventory where glottalization, nasalization and length are eliminated. Labialization may appear on either the first consonant or the second consonant, but not both (*w counts as a labialized consonant).
√iθʷ- 'to go, move' (activity)
√saŋ- 'to perceive, sense' (activity)
√qʷuš- 'to scratch, wound' (achievement)
√χet- 'to die, kill' (accomplishment)
√nel- 'to be born' (accomplishment)
√rox- 'to flow' (activity)
√yat- 'to copulate' (activity)
√aš- 'to bear, carry' (activity)
√loχʷ- 'to exist' (state)
√θekʷ- 'to know' (state)
Roots have an inherent argument structure and lexical aspect/Aktionsart (a four-way division between achievements/semelfactives, accomplishments, activities and states) but there are a number of valency or aspect-changing ablaut and similar processes that may take place within a root. For example, a state may be turned into an accomplishment through the transformation CVC > CVnC̉: θekʷ- 'to know', θeŋk̉ʷ- 'to realize'. Similarly, intransitive verbs can easily be causitivized with initial reduplication and lengthening: nel- 'to be born', nenēl- 'to give birth'.
A particular verb's aspect and transitivity properties can be indicated with a shorthand abbreviation, which we'll use from here on in. I marks an intransitive; T a transitive. A capital letter is used to indicate a telic aspect, while a lowercase letter is used for atelic aspect; <a> is used for durative aspects, while <x> is used for punctual ones.
√iθʷ- 'to go, move' (activity)
√saŋ- 'to perceive, sense' (activity)
√qʷuš- 'to scratch, wound' (achievement)
√χet- 'to die, kill' (accomplishment)
√nel- 'to be born' (accomplishment)
√rox- 'to flow' (activity)
√yat- 'to copulate' (activity)
√aš- 'to bear, carry' (activity)
√loχʷ- 'to exist' (state)
√θekʷ- 'to know' (state)
Roots have an inherent argument structure and lexical aspect/Aktionsart (a four-way division between achievements/semelfactives, accomplishments, activities and states) but there are a number of valency or aspect-changing ablaut and similar processes that may take place within a root. For example, a state may be turned into an accomplishment through the transformation CVC > CVnC̉: θekʷ- 'to know', θeŋk̉ʷ- 'to realize'. Similarly, intransitive verbs can easily be causitivized with initial reduplication and lengthening: nel- 'to be born', nenēl- 'to give birth'.
A particular verb's aspect and transitivity properties can be indicated with a shorthand abbreviation, which we'll use from here on in. I marks an intransitive; T a transitive. A capital letter is used to indicate a telic aspect, while a lowercase letter is used for atelic aspect; <a> is used for durative aspects, while <x> is used for punctual ones.
rebooting this because I actually have Proto-Nahtak probably figured out
Phonology:
Consonants: /p t ts tʃ k kʷ (ʔ) m n ŋ θ s ʃ h w l j/ <p t c č k kʷ ʔ m n ŋ θ s š h w l y>
Vowels: /a e i o ɨ a: e: i: o: ɨ: ã ẽ õ/ <a e i o ɨ ā ē ī ō ɨ̄ ą ę ǫ>
various clusters word-internally, /t k n s h w y/ word-finally
Some random morphophonology rules that I'll add to as I work things out:
I'm going to ignore nouns for the moment because lazy. The template for verbs differs between realis verbs (which take personal prefixes), and irrealis verbs (which don't). For realis verbs, the template looks something like this:
-5. Personal and gender prefixes for theme or superior argument.
-4. Adverbial prefixes.
-3. Evidential prefix.
-2. Incorporated noun.
-1. Incorporative instrumental prefix -mᵃ-.
0. STEM.
1. Instrumental suffix.
2. Transitive suffix -k-.
3a. Obviation marking for theme or superior argument.
3b. Plurality and clusivity marking for theme or superior argument.
4. Gender and person marking for inferior argument.
5a. Obviation marking for inferior argument.
5b. Plurality and clusivity marking for inferior argument.
6. Inverse marker -ʔwa-.
7. Gender and locality suffix for actor.
8. Tense marking.
9. Gender and locality suffix for goal or theme.
10. Discourse clitics (a large and somewhat mutable class.)
Irrealis verbs have the following template:
-4. Adverbial prefixes.
-3. Evidential prefix.
-2. Incorporated noun.
-1. Incorporative instrumental prefix -mᵃ-.
0. STEM.
1. Instrumental suffix.
2. Transitive suffix -k.
3. Irrealis marker -c-.
4. Gender/person marker for superior agent.
5a. Obviation marking for theme or superior argument.
5b. Plurality and clusivity marking for theme or superior argument.
6a. Gender and person marking for inferior argument.
6b. Obviation marking for inferior argument.
7. Plurality and clusivity marking for inferior argument.
8. Inverse marker -ʔwa-.
9. Gender and locality suffix for actor.
10. Tense marking.
11. Gender and locality suffix for goal or theme.
12. Discourse clitics.
Of note also is a special, "floating" affix for the habitual aspect, which reduplicates the affix to which it attaches. If tense marking (slots 8 and 10, respectively) are filled, it will attach to those; otherwise it will reduplicate the irrealis marker (3) in an irrealis verb; otherwise it will attach to an evidential prefix (-3); if all these are empty it will reduplicate the root (0).
(mostly just putting this here so I actually know what the fuck I'm doing when I fill out the paradigms; an actual article will be forthcoming)
Phonology:
Consonants: /p t ts tʃ k kʷ (ʔ) m n ŋ θ s ʃ h w l j/ <p t c č k kʷ ʔ m n ŋ θ s š h w l y>
Vowels: /a e i o ɨ a: e: i: o: ɨ: ã ẽ õ/ <a e i o ɨ ā ē ī ō ɨ̄ ą ę ǫ>
various clusters word-internally, /t k n s h w y/ word-finally
Some random morphophonology rules that I'll add to as I work things out:
If you get an inadmissable cluster, first turn w y > o i / {#C}_{C#}
Then look at the last two members of the super-cluster. Do they form a valid cluster by themselves? If yes, add an e directly before the second-to-last member of the super-cluster. If not, add an e before the last member.
Repeat as necessary.
Final -θ becomes -s.
Some affixes have special epenthetic vowels or consonants that only appear before or after a vowel. These are written with superscripts.
I'm going to ignore nouns for the moment because lazy. The template for verbs differs between realis verbs (which take personal prefixes), and irrealis verbs (which don't). For realis verbs, the template looks something like this:
-5. Personal and gender prefixes for theme or superior argument.
-4. Adverbial prefixes.
-3. Evidential prefix.
-2. Incorporated noun.
-1. Incorporative instrumental prefix -mᵃ-.
0. STEM.
1. Instrumental suffix.
2. Transitive suffix -k-.
3a. Obviation marking for theme or superior argument.
3b. Plurality and clusivity marking for theme or superior argument.
4. Gender and person marking for inferior argument.
5a. Obviation marking for inferior argument.
5b. Plurality and clusivity marking for inferior argument.
6. Inverse marker -ʔwa-.
7. Gender and locality suffix for actor.
8. Tense marking.
9. Gender and locality suffix for goal or theme.
10. Discourse clitics (a large and somewhat mutable class.)
Irrealis verbs have the following template:
-4. Adverbial prefixes.
-3. Evidential prefix.
-2. Incorporated noun.
-1. Incorporative instrumental prefix -mᵃ-.
0. STEM.
1. Instrumental suffix.
2. Transitive suffix -k.
3. Irrealis marker -c-.
4. Gender/person marker for superior agent.
5a. Obviation marking for theme or superior argument.
5b. Plurality and clusivity marking for theme or superior argument.
6a. Gender and person marking for inferior argument.
6b. Obviation marking for inferior argument.
7. Plurality and clusivity marking for inferior argument.
8. Inverse marker -ʔwa-.
9. Gender and locality suffix for actor.
10. Tense marking.
11. Gender and locality suffix for goal or theme.
12. Discourse clitics.
Of note also is a special, "floating" affix for the habitual aspect, which reduplicates the affix to which it attaches. If tense marking (slots 8 and 10, respectively) are filled, it will attach to those; otherwise it will reduplicate the irrealis marker (3) in an irrealis verb; otherwise it will attach to an evidential prefix (-3); if all these are empty it will reduplicate the root (0).
(mostly just putting this here so I actually know what the fuck I'm doing when I fill out the paradigms; an actual article will be forthcoming)