| Name: Kel
Type: Agglutinating, partly fusional head-initial test-drive language
Alignment: Accusative, with topic prominence
Head Direction: Initial
Number of genders: 4
|Flaps / taps|
Aspirated plosives: /ph/, /th/ and /kh/ are the aspirated counterparts of /p/, /t/ and /k/ rsp. You can experience the difference if you pronounce "blackhat" vs. "blackadder". Most important, /th/ has nothing to do with the (English) th in "thief" or "the", and /kh/ is not the Scottish, German or Russian "ch"-sound.
On the other hand, the non-aspirated plosives are to be pronounced with less aspiration than the average English or German speaker would do.
Liquids: /r/ is pronounced with a single flap between vowels (i. e. [ɾ], and as a trill (i. e. multiple flaps) elsewhere. More or less like in Italian or Spanish. /l/ has a "darker" (more velarized) sound in the absolute ending of a word (i. e. , like in English "well" or (European) Portugese "Brazil" (not quite as dark as in Polish [ł]).
/j/ is always the y of "yes", never the j of "judge" (EN) or "jacques" (FR). It counts as a consonant, so although /paj/ seems to have a diphthong (and sounds exactly the same), it doesn't. In phonotactics, it's explained why.
There are five phonemes in Kel: /a/ /e/ /i/ /o/ /u/. The tables shows the pronunciation in accented syllables. In unstressed ones, /a/, /i/ and /u/ may be pronounced a bit laxer, but they do not neutralize like in English or Russian.
|Front||Near front||Central||Near back||Back|
There is no length/quantity difference (long or short vowels).
To make things easier, I will use /e/ for /ɛ/ and /o/ for /ɔ/ from now on.
Syllable structure is (C)V(C), while the first syllable allows (C1(C2))V(C). Only /s/, /n/, /l/ or /r/ may appear as C2.
Phonotactically spoken, the semivowel /j/ counts as a consonant. That means that a syllable like /khaj/ or /suj/ cannot take another consonant on the coda; /khajd/ or /sujl/ are not possible.
If by derivation or flexion two equal consonants meet at the syllable boundary (e. g. /fog-go/), they are pronunciated as if there was only one ([fogo]). There are no geminates.
On syllable boundary, voice (where there is an opposition) assimilates regressively (backwards), thus /apda/ will be pronounced ['abda], but /abta/ will result in ['apta]. This may lead to more geminate neutralizations (e. g. /tak-go/ > */tag-go/ > [tago]).
This does not apply to consonants who haven't got a voiced/unvoiced opposition, i. e. /f/, /s/, /m/, /n/, /r/, /l/ and /j/ do not affect their neighbours.
Prosody and SuprasegmentalsEdit
Accent is marked by pitch and (moderate) stress and regularly falls on the penultima (last-but-one syllable), like in Polish or Esperanto. The language is syllable-timed (like Spanish), as opposed to stress-timed (like English) or mora-timed (like Latin or Japanese).
Phrase melody is normally falling, with the beginning (normally the topic) marked with a higher pitch. In yes/no-questions without a special question marker, the voice rises at the end of the phrase again.ci
Stiip 01:58, June 5, 2011 (UTC)