Kash phonology II

DISTRIBUTION AND SYLLABLE STRUCTURE

1. Aside from the prenasalized stops (if they are clusters), the only
permitted clusters are / p t k/ + r, initial and medial; and /nd mb/ + r,
medial only. In cases where a C-initial suffix joins a C-final base, some of
the resulting clusters, such as -r + t, nasal  + nasal,  are subject to
sandhi rules (of which more later); the remainder, such as -k + t or -s + k,
are broken up by inserting a weak schwa.

2. Given the predominant free morpheme structure CVCVC, these restrictions
apply:
a. only /p t k m n ñ r l s/ or zero occur finally.  /-sh/ also occurs, as
the neut. plural marker, and in the single free form _kash_ '(n) person;
(adj.) pertaining to the Kash people'.  A few interjections have final -f
and -h.
b. all C and clusters may occur medially.  -NC(r)- occur only medially.
(According to the writing system, zero does not occur here).
c. "transitional w" and the sequence yi- do not occur initially; otherwise,
all C and permitted Cr clusters, or zero,  may occur.
d.  two /r/s may not occur in sequence i.e. no #rV(NC)r... or  ...(NC)rVr#,
Various strategies are employed to avoid such sequences in speech,
principally dissimilation of the first /r/ > /l/, or pausing/inserting a
hesitation syllable between the two words.
e. vowels may occur freely in either position, BUT: there are no sequences
of -uwu-, -owo-, -iyi-, -eye-.  (In effect, sequences of identical V are not
permitted within a base; the writing system obscures this.)  The latter two
can occur across a morpheme boundary (e.g. she'njiyi /'Senji+i/ genitive of
the name Shenji, fa'leye /'fale+e/ dative of fale 'market') and are
generally pronounced long [SE'nji:], [fa'le:], or more correctly, doubled
[SE'njii] [fa'lee] with falling intonation on the unstressed final segment.
Also across a morpheme boundary, two /a/'s my occur-- these coalesce to a
single /a/, as in ma- '1st sing.' + 'ale 'to be' > 'male 'I am'.

3. In addition to the predominant two-syllable bases, free morphemes may
also be:
a. monosyllabic V:  i 'and', e 'def. article'.  CV:  ta 'not', ro 'two' (VC
forms are not attested, but probably possible.)  C(r)VC: leñ 'good', prum
'belly'; tup '(imit. of bouncing)'-- one of many onomatopoetic/imitative
forms.
b. trisyllabic:  usually inflected forms, derivations or compounds, though
some are not.  makaya 'I know', angaya 'knowledge'; cindero 'noon' (cini
'middle' + lero 'day'); onoshi gen.pl. of ono 'egg'.  lipembet 'to wrap'
(perhaps an old compd. based on limbe 'to wear'?), pilimen 'to consider as,
view as...', shoteru 'fog'.
c.  words of 4 syllables or more are invariably compounds, usually quite
transparent-- like apambraka 'billion (1000 million)' < amba reduced to apa
'l000' + ambraka 'million' itself < amba + raka 'big'; though some have
undergone minor phonological "adjustments", like latondrele 'planet' < lato
ri nele 'wander in sky'.  Really long compounds are likely to be archaic,
scientific, or just spur-of-the-moment.  If they were to enter common usage,
they would tend over time to be contracted to 4 or 3 syllables.

Next:  some morphology