Re: Critique sought

Bryan Maloney wrote:

> I've put up some notes towards a conlang at
> http://www.people.cornell.edu/pages/bjm10/Gloling/praxname.html
>
> I confess that it's not 100% original--it's based on the Glorantha
> setting.  Thus, I am somewhat constrained.  If I want to remain "true" to
> the "source" material, there is very little I can subtract.  A few things
> in particular bother me:
>
> The vowel structure seems lopsided.  Any suggestions on how to fix it?
Yeah, it kinda is.  Here's what happens in natural languages:  vowels
(like all sounds) tend to develop or be lost with respect to what
phonological characteristic they share with other vowels.  Usually,
a whole class of vowels (like all nasalized vowels, or all front rounded
vowels) will undergo a change (like denasalization, or losing their
roundedness) at the same time.  This leads to a nice, balanced look
to the phonology (but see below).

Usually, there are also certain vowels which pop up far more often
than others.  If a language has a three-vowel system, it'll likely have
/i u a/; 4 vowels, /i u @ a/; 5 vowels /i u e o a/; 6 vowels, /i u e o @ a/,
and so forth.  That's just the norm; you don't have to have a language
like that.  But if you don't, it won't be very naturalistic. The object
here is to make efficient and optimal use of the physical acoustic space
in the mouth, where each vowel will try to be maximally distinct from
its neighbors. Practically speaking, that means it's weird to have vowels
bunched up in the lower-fronter vowel space as you have (with an
unusually high 14 vowels, no dialect of English even does that). You also
have a distinction between lax vowels and tense vowels for the front
high unrounded vowels ([i I]) but not for any of the others -- if the
language were actually used in regular speech, that would almost certainly
change quickly;  either the other vowels would acquire tense/lax variations
of themselves, or (what's more likely) the one distinction made here would
be dropped, most likely in favor of some vowel which approximates [i].

> I think that I've made a goof in adding a labiodental nasal--the "source
> material", since it was written by a Californian game designer with zero
> linguistic background, uses a vary English implicit phonology (from what
> I've been able to gather).  Would a language be likely to differentiate
> between a bilabial and a labiodental nasal?
Well, maybe.  But if it had a labiodental, it would also be more likely
to have a velar [N] too -- the same kind of progression goes for consonants.
If it has 1 nasal, it'll probably have [n] or [m];  2 nasals, [m n]; 3
nasals,
[m n N]; 4 nasals, [m n n_y N] (where [n_y] = the palatal nasal, as in
Spanish).  the labiodental nasal probably would come somewhere soon
thereafter in liklihood.  But it's a bit unusual if you don't also have [N]
and
maybe [n_y].

> The "terminal glottal stop goes away sometimes" rule doesn't quite work
> for me--is it nonsensical?  (All the "combinatorial" rules are my own
> design).
The problem lies in the word "sometimes" -- what does that mean?
Does that mean that there's a rule for it, describing the behavior of
the glottal stop, or that it's completely arbitrary?  The latter is almost
certainly not going to be the case -- language changes most easily
by applying rules, and while allophones may be in freevariation,
(a) they usually aren't, and (b) a full phoneme almost certainly wouldn't
be (otherwise, why would people distinguish it in the first place?).


======================================
Tom Wier <artabanos@...>
ICQ#: 4315704   AIM: Deuterotom
Website: <http://www.angelfire.com/tx/eclectorium/>
"Cogito ergo sum, sed credo ergo ero."

Non cuicumque datum est habere nasum.
It is not given to just anyone to have a nose.
        -- Martial
======================================