Proto-Serazi

General Information
Proto-Serazi is the most recent common ancestor of all Serazi languages, which includes both Alemarese and Barejine. It was most likely spoken in the center of the Chevin peninsula, spreading inland, to the coasts, and to the surrounding islands (displacing other languages minus Varrit).

Consonants
The traditional classification and symbols are shown below.
 * The column labeled 'alveolar' was most likely laminal (denti)alveolar.
 * /f/ could have been either bilabial [ɸ] or labiodental [f].
 * /r/ could have been nearly any rhotic consonant, but it is currently constructed as an alveolar approximant.
 * /h/ and /ɣ/ were most likely farther back in the mouth than velar, possibly pharyngeal/epiglottal.
 * /ɣ/ could have been a stop, fricative, approximant, or even a trill. There is simply no way to know given its diverse reflexes in descendant languages.

Vowels

 * The vowel length distinction may have been accompanied by a tense/lax distinction.
 * The only reconstructed diphthongs are /ey/, /oy/, /ay/, and /ew/, /ow/, /aw/. It seems that combinations of /u/ and /i/ with /y/ and /w/ caused the vowels to lower.

Therefore, corēmu* "guideline, rule" may have been pronounced [kɔˈɹeˑmʊ].

Stress
Stress was fairly regular, appearing on the final syllable of the root when no long vowels were present. Long vowels took much more stress than short vowels, even when they weren't the last root vowel. fetútya* "tree" vs. bā́rumu* "mountain"

Gender
Proto-Serazi is believed to have had three gender classes easily identifiable by their final vowel: animate (-a), lustrous (-i), and dull (-u). The historical diminutive (-ya) was exclusive to the animate class and widely used for plants, therefore some postulate it to be in its own class, but it did not trigger any unique agreement.

Examples
 * Animate: luɣa* "spinnerfly"
 * Diminutive: irdya* "flower"
 * Lustrous: dōhi* "gold"
 * Dull: ɣwertu* "house, home"

Number
Number was fairly simply expressed in Proto-Serazi, by adding a final -y. This results in /u/ and /i/ lowering to /o/ and /e/.
 * -a->-ay
 * -i->-iy->-ey
 * -u->-uy->-oy