adapted from a 22 May 1997 post My fledgling conlang: phonology by Kenji Schwarz
I've got very little done on this thing so far, beyond what I'm posting here for comments & criticism. Since this is more of an exercise in "creative parallel evolution" than free-form invention, I have a fairly fixed idea of the overall nature of the beast, but haven't done enough research yet on the actual details.
At the moment I don't have a real name for this language. For now, it's called "AYUPT", from "As Yet Unnamed Pseudo-Tungusic." I'm considering it to be a Northern Tungusic language that was greatly influenced by a Southern Tungusic language (Khitan, maybe?) c. 12-13th centuries, and likewise by Nivkh (a "Paleosiberian" isolate). However, I haven't got ahold of much material on Nivkh yet, and am still reading up on North Tungusic, so any and all of this is likely to change before long. In the modern period, I'm thinking of a lot of loanwords from Russian and probably Japanese. Since the AYUPT-speakers are devout Commies, I'd like to eventually translate some bits of The Communist Manifesto into their language, rather than the Bible or Shakespeare.
Consonants | Vowels | Vowel harmony | Stress
AYUPT has 18 consonant phonemes. Nothing terribly exotic here:
| STOPS | FRICATIVES | NASALS | |||
| vls | vcd | ||||
| BILABIAL | p | b | v | m | |
| DENTAL/ALVEOLAR | t | d | s | l | n |
| ALVEOLAR | p | ||||
| PALATAL/ALVEOLAR | c | j | y | c^ | |
| VELAR | k | g | x | N | |
/b d g/ are technically murmured stops [b? d? g?], but have also been described as "aspirated voiced stops" by earlier Western linguists. They cannot occur word-finally.
/p t k/ are ejective [p` t` k`] in word-initial position, aspirated [p t k] when intervocalic, and unreleased when word-final.
/b/ becomes [B] before voiced consonants.
/k g x/ have fronted and retracted allophones to match the dimension of their surrounding vowels:
All consonants except the alveolar series /t d n/ palatalize before the front vowels /i e/: /p m l k/ -> [p; m; l; k;] (etc.)
/t d n/ are alveolar.
/c/ is a barely affricated, alveo-palatal stop. Originally a palatalized stop /t;/.
/j/ is also barely affricated. Especially in initial position, it can resemble [d;] - a strongly palatalized alveolar stop. In non-initial position, it can be a voiced palatal sibilant [Z].
Neither /c/ nor /j/ can occur word-finally (nor can /n^/).
/v/, normally a bilabial voiced fricative [B], appears as an approximant [w] in some dialects. In word-final position it devoices to [F]. The use of {v} to transcribe it phonemically is perhaps misleading, but for now the best option I can see.
/l/ is a dental laminal; it is pronounced with fairly spread lips.
/s/ has a number of allophones:
/r/ is an alveolar approximant. It cannot occur word-initially; word-finally it is muted and often alternates with /l/.
/N/ in word-final position shifts to a uvular nasal [n"].
/y/ is a strong fricative [j] word-initially, somewhat weaker intervocalically, and a weak approximant or off-glide word-finally.
(Sandhi in affixation is somewhat complicated and will be detailed later.)
AYUPT has a system of vowel harmony, an unusually regular one compared to other known Tungusic languages. Perhaps the system was not preserved from the proto-language, but rather artificially regularized in modern times?
Quantitative length of vowels is phonemic; all vowels, except /e/, can be either short or long.
| FRONT | BACK | |||
| HIGH | /i/ | /i"/ | /u/ | /u-/ |
| /e/ | /@/ | /o/ | ||
| LOW | /a/ | |||
All but one of these vowels have long counterparts: /i:, i":, u:, u-:, @:, o:, a:/. There is no corresponding length distinction for /e/. Unless stated otherwise, all comments below about short vowels apply equally to long ones.
/i"/ is farther back and opener than /i/. It does not palatalize preceding consonants.
/u-/ is unrounded, and perhaps slightly lower than /u/.
/i/ palatalizes preceding consonants (except /t d n/). Several allophones:
/e/ palatalizes preceding consonants (except /t d n/). Its length is not phonemic. Allophones:
[E] at the end of words is often further reduced to zero, though it 'reappears' when suffixes are added.
/@:/ is usually distinguished from /@/ by other features besides quantity. Often it is slightly rounded or has a slight labial offglide.Several allophones:
/o/ is slightly opener than /u/, but becomes more rounded word-finally and in closed syllables.
AYUPT shows fairly regular and consistent vowel harmony. The relevant distinction of vowels is between front/high and back/low vowels:
| Front/high: | /@, e, i, u/ | |
| Back/low: | /a, o, i", u-/ |
Each vowel is said to correspond to the one in the same place in the series of the other type: /a/:/@/ :: /o/:/e/ :: /i"/:/i/ :: /u-/:/u/
A given word may contain either only front or only back vowels.
The dimension of "roundedness" is also important in a few situations:
| Rounded: | /o, u/ | |
| Unrounded: | /a, e, @, i, i", u-/ |
Stress accent is free, and can fall on any syllable.
Pitch on stressed vowels is perceptibly higher than on unstressed ones.
Vowels that are both short and unstressed are often reduced in speech.