Re: Word Construction for a New Conlang

On Thu, 8 Jul 1999, Sally Caves wrote:

> Ed, a person you might ask, who deliberately set out to limit the sounds
> of his language, is Matt Pearson.  What makes Tokana distinctive to my
> mind is that he decided to eliminate our distinction between voiced
> and voiceless stops like t/d p/b k/g.  I was very impressed by that.
> But I can see where your concern is: do you make the rule first and then
> stick to it, or do you experiment with sounds, and then decide which
> sounds occur most often?
My name has been uttered and, like the djinn, I must respond!  :-)

Actually, I don't really have that much to add to this thread.  My
choice to limit the number of sounds in Tokana was a reaction to
previous conlang projects of mine, where the number of sounds had
gotten a little out of hand.  Basically the way I proceeded was not
too different from what other people seem to have done:  I invented
a handful of words which I thought sounded cool, and then extracted
a phonology and phonotactics from that.  Most of the words from my
'cool list' were of the CVCV or CVCCV sort (the first ever Tokana
word was "talhko'", which ended up meaning "because"), so I decided
to limit the syllable structure to CVC.  A wrinkle was added when
two words crept into Tokana with initial consonant clusters:
"klohana" = "to go through" and "sliahte" = "story".  So I decided
to add a limited number of possible initial CC clusters to the
mix - namely, "sp", "st", "sth", "sk", "ps", "ts", "ks" and "kl".
The conhistorical explanation for this is that proto-Tokana
allowed a larger number of consonant clusters, but most of these
were simplified over time.  The clusters containing "s" survived
because "s" is an acoustically salient sound, and resisted deletion.
(Initial "kl" has a different history, being the result of a more
recent set of sound changes:  "klohana" was originally "ytlohana",
with loss of the initial schwa vowel "y" and change from "tl" to
"kl".)

On a different topic, it's funny that Sally should mention Tokana's
lack of a voicing distinction in the stop series, since I was just
thinking today about adding some much-needed allophony to Tokana
stops.  I'm toying with introducing the following rules (ordered
as given) into the language:

(1) Between two vowels, a voiceless stop becomes voiced

(2) When followed by an unstressed vowel, a voiced stop becomes
the corresponding voiced fricative (spirantisation)

Thus a word like "Tokana" would come out [to'gana] phonetically.
A shift in stress from the second to the third vowel would then
cause [g] (phonemic /k/) to become the voiced velar fricative
[G].  For example, the instrumental case form "Tokanane" would
be pronounced [toGa'nanE].

Rules (1) and (2) have yet to be canonised, but I'm definitely
leaning in the direction of adding some voiced stops to Tokana,
and I think it would be better to have those voiced stops be
phonetic variants (allophones) of the voiceless stops, rather
than complicating the basic sound inventory of the language.

So Sally, are you still impressed, or are you woefully
disappointed?  :-)

Matt.