Re: DECAL: Examples #3: Phonological change rules

On Thu, Jan 13, 2005 at 01:18:20PM -0800, Sai Emrys wrote:
> Same deal.
>
> Try to use /UR/->[SR] / X_# type rules (e.g. /C/ -> [-vd C] / [-vd C]
> _ ) (# = word boundary).
>
> Q1: What are your *phonologically driven* sound change rules? I.e.,
> these will apply to *all* situations, once higher-level UR processes
> are done (e.g., morphology). Be sure to include the order of
> application, if it's relevant (e.g., you have feeding). This includes
> cases that are caused by word boundaries, syllable boundaries, cluster
> conflicts (e.g. VC + CV in a CVCV limit), etc.
There are no such processes in Ebisédian. Main reason: I had no idea
what I was doing when I designed Ebisédian phonology. :-) Well, there
is a complicated system of ablaut-like vowel shifts in nouns to
indicate case, but I believe that qualifies under morphology proper.

In Tatari Faran, the only truly general sound change rule is the
pronunciation of the phoneme /r/. When word-initial, it is pronounced
[d], and when medial, it is pronounced [4].


> Q2: Ditto - but for *morphophonology* or otherwise non-general cases.
> E.g., the prefix in- for English (-> r, l, m by context) - it only
> applies to that morpheme, not generally AFAIK. Again, include order of
> application.
>
> (Note: don't go into morphology just yet; that'll be another post. For
> this, all I'm asking for is underlying -> surface representation
> rules.)
Again, for Ebisédian, there's none. It's pretty boring in that sense.
:-)

For Tatari Faran, however, there are quite a number of such changes.
Although individual words may well have repeated syllables, for
example _nana_, a complement meaning "to ridicule", Tatari Faran
generally dislikes repeated syllables (or adjacent syllables that
sound too alike) that result from the juxtaposition of words in a verb
or noun phrase. This drives a lot of sound change processes, among
which are:

1) If a noun ending with _na_ (or _nV_ where V is a vowel) is
immediately followed by one of the receptive particles (_na_, _nei_,
or _no_), /n/ is appended to the noun and the consonant on the
particle shifts to /r/ (which comes out as [d] because it is
word-initial). For example:
	huna + na -> hunan da

If the noun ends with _nan_ (or _nVn_), the particle mutates but the
noun stays the same:
	misanan + nei -> misanan dei

2) Some special cases of nouns and particles involving sV syllables
will assimilate. For example:
	asusu + sei -> asusei
	isi + sa -> isa
	isi + sei -> isei
	isi + so -> iso

3) The feminine case particles merge if immediately followed by the
interrogative particle _ta_:
	kei + ta -> kita
	sei + ta -> sita
	nei + ta -> nita

4) Words ending with _ei_, if followed by a word whose initial
syllable also has _ei_, will shift the vowel to _i_. For example:
	kei + ei -> ki'ei
	sei + ei -> si'ei
	nei + ei -> ni'ei
	fei + kei -> fikei
	fei + sei -> fisei
	fei + nei -> finei

If there are 3 _ei_ words in succession, only the 2nd will mutate:
	fei + kei + ei -> fei ki'ei

One could rationalize this by saying that the change is applied in
reverse order, so _kei_ + _ei_ forms _ki'ei_ first, and then it no
longer has _ei_ in the initial syllable and so the mutation ends
there.


> Q3: Motivation, again? (If anything other than purely aesthetic, or
> you can give details of why you think your choices made for better
> aesthetics.)
The TF sound changes appear to be driven by euphonic considerations
[hunanda] sounds "better" than [hunana], and [fikej] sounds "better"
than [fejkej]. What is euphonic and what isn't, of course, is
determined by the "sound" that I wish TF will have. For example, there
is a predominant use of [a] to get a vaguely Austronesian feel to it,
as well as the use of circumfixes (e.g. the compositive, _i-...-an_).


T

--
Chance favours the prepared mind. -- Louis Pasteur