Mari
My ad hoc reconstruction of Mari vocalism (note: read literature):
- *e *i *o *u *ü *ö = branches agree
- Hill Mari /ä/ ← *e / _{ŋ, r}_
- reduced high vowels: *ɪ *ʏ *ʊ = Hill (Western) Mari /ə ə ɤ/, Meadow (Eastern) Mari /i y u/
- reduced mid vowels: *ə *ɤ = Hill Mari /ə ɤ/, Meadow Mari /ɤ o/ (no front rounded member?)
- NB: /ə ɤ/ are more accurately [ə̟ ə̠], or perhaps [e̽ o̽].
- *a *oo = in Hill Mari /a o/, Meadow Mari /o u/
- /a/ also exists sporadically in Meadow Mari.
The vowel inventories are:
- Me /a e i o u ü ö/
- Hi /a e i o u ü ö ä ə ɤ/
- NW dialects also /ʊ ʏ/
Etymology notes
The development *e_i → *ə appears to be limited to loanwords (so not common inheritance after all?):
- *əšte- "to be in time" ← *eštə-; in Permic may be a BF loan
- *ləštaš "leaf" ← *leštə ← IE
- *šəm "7" ← #śeććəmə ← IE? NB irregular medial loss (but could be secondary haplology)
(more frequent in *e_ä-roots, but note *rəwəž "fox" ← *repä ← II)
Palatalization
- *kɪćke- "to harness" ← *ki/ütkə- (Ms. *kät- ~ H. köt!)
- *šʏć "rust" ← #sïta
- *wɪć "5" ← *wɪt ← #wittə
- *tić "full" ← (??) *tejət ← *täwðətä
but not in:
- *jʏt "night" ← *jEtɜ
- *kit "hand" ← *kätə
- *wʏt "water" ← *wetə
- *tütka- "to spread" ← *tütkə-
Denasalization
Apparently *mp, *nč → *p, *č (despite *nt, *ŋk → *nd, *ŋg):
- *mp:
- *lop "depression" ← *lëmpɜ
- *leewə "warm" ← *lämpə
- *kowə "wave" ← *kompa
- *tup "back" ~ U tɨbɨr
- *mt → *md:
- *komdəš "lid" ← *kamtə
- *kʊmda "wide" ← *kumtə
- *nt → *ð (← Permic?):
- *oðar "shoot" ← II *andʰas
- *lʊðə "goose" ← #lunta
- *šeðäŋ "wheat" ← #šäŋtɜ
- *nt → *nd:
- *kande- "to carry" ← *kanə-ta-
- *kində "grain" ~ P *kɨd
- *pandə "stick" ← *pontə
- *pʊndaš "bottom" ~ P *pɨdes
- *šandə "dung" ~ F *sonta
- *šɪnde- "to stretch" ~ S *sɤntē-
- *nč → *č:
- *nś → *ž:
- *ŋt → *nd:
- undə "sting" ← *oŋtə
- *ŋk → *ŋg:
- *oŋgə "hook" ← *oŋkə
- *moŋgər "body" ← *moŋkɜ
- *poŋgə "mushroom" ← *pëŋka
- *šeŋgə "dry wood" ~ F. säŋki?
- *šeŋgel "behind" ← *śäŋkɜ
- *šoŋgə "old" ← *šoŋka
- *šʏŋgä "hill" ← *ćüŋkä
No reflexes: *ëmta-, *omtə, *tumtə-; *jäntə, *kënta, *konta, *küntə, *piNtä-; *jäŋtä-; *jäŋkä, *këŋkar
The distribution of *a vs. *o
These two vowels appear basically in the same contexts with not much respect for etymological origin; regularly from *a and *ë_a, slightly less regularly from *o_ə, sporadically from *o_a and *ë_ə. Let's see if the distribution has any relation to the consonant environment:
*a | *o | |
---|---|---|
*a-heavy | ||
*p- | default *a
|
*o / _[+BACK]
|
*w- | default *a
|
(no examples!) |
*j- |
|
(*joɣe- "to flow" ← Chuv. jox-) |
*o-heavy | ||
∅- | *a / _Ć
|
default *o
|
unclear | ||
*ć- |
| |
*č- |
|
|
*l- |
|
|
*m- |
|
|
*n- |
|
|
*r- |
|
*o in hiatus?
|
*š- | chiefly in loans + original CVCCV roots?
|
chiefly in original CVCV roots?
|
*t- |
|
|
*k- |
Labial coloring
In theory this could be a West Mari innovation from PMa *u, but this makes better phonetic sense as an earlier change *o → *u.
o → ʊ / _m_
|
exceptions etc.
|
o → ʊ / _p_
|
exceptions etc.
|
o → ʊ / w_
|
exceptions etc. |
The development of *ð, *ðʲ, *t
*ð- → l-
|
*ðʲ- → l-
|
*t- → t-
[abundant examples, not included] |
*-ð- → ∅
|
-ðʲə → ∅
|
*-t- → ∅
|
(no examples of *ð → ð/t!) | *-ðʲa → -ð-
possibly also:
|
*-ta → -ð-
|
*-ðʲ- → -t
possibly:
|
*-tə → -t
| |
*-ta → -t
|
Exceptions: *kəl "handle" ← *käwðə "rope", probably from Permic (note also irregular *ə)