Harmony: Dominance Patterns in Seinundjé (Comment?)

So, I was thinking about consonant and vowel harmony, and I realized that
there is a certain limited consonant harmony in Seinundjé.

Seinundjé has the following consonant inventory:

velar: k g x|kh| G|gh| N|ng|
"palatal": tS|c| dZ|j| S|ch| Z|jh| (J) l
retroflex: t` d` s` z` n` (r)
alveolar: t d s z n 4
dental:   T|th| D|dh|
labial: p b f v m w
glottal:   h

Retroflexes have a strange orthography: they are spelled as alveolars, but
at the end of a retroflex cluster we write a |j|.  SeinuNDJé is an
occurrence of this.

The consonants in parentheses only occur due to harmony processes.
The palatal, retroflex, and alveolar series all participate in a harmony
process, where the place of articulation propagates to the right from an
initial consonant.  Among all the consonants, each series is opaque to the
harmony of series below it and vulnerable to harmony of series above it.
So, velars always halt harmony, while dentals, labials, and /h/ are
transparent to it.  This harmony occurs over the span of phrases.  In
addition, back vowels not only halt palatal harmony, but prevent it in
their onset consonants, except for nasals, which create a /j/ onglide
(offglide?).

This has the effect that within words, the orthography need only specify
the POA of the first harmony-triggering consonant before harmony is
halted; the rest can be written as alveolars (which have the least
cumbersome orthography.)

Example:

Árichesa inu harén  feledhla.
/a:4iSeSa iJju ha4e:n feleDla/
Árichesa PERF sun-ACC swallow-3s
Árichesa swallowed the sun.

Comments?

 ---
Shreyas Sampat