Re: Consonant clusters

Christopher Bates wrote:
> nasal + stop (at same place of articulation) eg went, wand, jump, mb
> (can't think of a word offhand with that one in)
nasal-stop can only be syllable final if the stop is voiceless (like
camp, cant, bank) or /d/ (land, wand, etc. - altho those often come out
as just /n/ in my experience).  Of course, nasal plus voiced stop can
occur word-medially, when they're in different syllables, as in "symbol"
or "finger", "candy".

> For some reason we don't have ngk or ngg...
Not written that way, but we do.  "Bank" has /Nk/, and /Ng/ can be found
word-medially, as in "finger".

Uatakassi is pretty simple.  Syllable initially, you can have only a
consonant followed by a glide or l (which is realized as [r\] - alveolar
approximate - after dentals).  A moraic [C] may also precede voiceless
consonants, but this is underlyingly /Ci/.  Syllable-finally one can
have a fricative, nasal, l, or gemination.  Labial consonants cannot be
followed by /w/, and t, d, and n were originally not permitted to be
followed by /w/, but recent sound changes have introduced a few [tw],
[dw], and [nw] clusters.  /tS/, /dZ/, and /C/ can only be followed by
vowels, and cannot be geminated unless /i/ follows.  (Due to their
origin as /tj/, /dj/, and /kj/ respectively)

There are some restrictions:

Geminated stops cannot be followed by glides or l.  (Thus, appa is
   legal, but not appia or appla; but affla is legal)  Some
   conservative dialects apply this restriction to *all* geminated
   consonants.
Sequences of stops and fricatives must agree in voicing, so, not *azka,
   but you can have azga or aska.  Assimilation is progressive.
/s/ or /z/ cannot be followed by /tS/, /dZ/, or /C/.  These
   assimilate to /SS/ or /ZZ/.  Thus, the suffixes -tas plus
   -ki (/Ci/) become -tassi (/taSSi/)
[S], [Z], [tS], [dZ], [C] may be followed by /wi/, but not /j/ or /l/
   This, incidentally, creates a problem in analysis.  Normally,
   [S] and [Z] behave as tho they are underlyingly /sj/ and /zj/,
   which is their origin, so that, for example, *[Sl] is illegal,
   because /j/ must be followed by a vowel.  However, /jw/ is
   an illegal, so why does [Swi] exist?  Historically, this was [Suj],
   but it's still difficult to explain synchronically.
   [tS], [dZ], and [C] sometimes behave as underlying /tj/, /dj/, and
   /kj/, however, [tj], [dj], and [kj] (the first two sometimes
   realized as [ts] and [dz] instead) do exist due to subsequent
   sound changes, making /tS/, /dZ/, and /C/ marginal phonemes.
   Perhaps /S/ and /Z/ should likewise be analyzed as marginal
   phonemes.
Moraic /C/ may not occur between two consonants.  [CtawtaSSi] is legal,
   for example (and means "(s)he is feared"), but something like
*[AfCta]
   wouldn't be, it would be pronounced [AfCita].  Moraic [C] is simply
   a realization of /Ci/.
Moraic /C/ may NOT be followed by a vowel.  It loses its moraicity if
   a vowel follows (underlyingly /Ci/ -> /Cj/ - realized as simply [C])
/g/ may not be followed by /j/, I think.  Originally it could not be
   followed by /i/ at all, but sound changes and borrowings have
   introduced /gi/ back to the language.

Hmm ... I think that's it.  It does allow some fairly surprising
clusters like /ml/ or /nl/ (realized as [nr\])

--
"There's no such thing as 'cool'.  Everyone's just a big dork or nerd,
you just have to find people who are dorky the same way you are." -
overheard
ICQ: 18656696
AIM Screen-Name: NikTaylor42