On Sun, 5 Jan 2014 09:02:30 -0800, Gary Shannon <[log in to unmask]> wrote:

>On Sun, Jan 5, 2014 at 3:03 AM, J. 'Mach' Wust
><[log in to unmask]> wrote:
>> On Fri, 3 Jan 2014 12:29:08 -0800, Gary Shannon wrote:
>>>Empirically, and without reference to any existing theory of
>>>phonology, "y" can, with complete consistency, be mapped to the same
>>>class which includes {a, e, i, o, u, oi, ou, au, ow, aw, wi, we, ...}
>> Wrong. English has a few words that are sensitive to the distinction
>> between consonants and vowels. The most obvious is the alternation
>> between "a" and "an". Before vowels, it is "an" (an eel, an idea, an
>> ill-gotten something), before consonants, it is "a" (a keel, a muse,
>> a strange one). What is it before [j] and [w]? Right, it is "a" (a
>> youth, a use, a wood), so these sounds are in the same class as
>> consonants -- with regard to English.
>And that makes "h" a vowel then. This aN Historic occasion!

Well, at least in the stylebooks of those who treat grammar as a game of gotchas, whose rules were handed down graven on tablets from Strunk and White on their holy mount.  But not generally in English as she is spoke these days, AIUI, except by hypercorrection.  And I think when it wàs current, dropping the /h/ in that position was current too.  

>That's an English-specific convention, just like "a youthful lad"
>instead of "an youthful lad". What I'm trying to do is not to describe
>English usage of vowels and consonants, but to come up with a general
>system of pronounceability for coining non-English conlang words. 

Well, if you're coming up with _a_ system of pronounçability -- as we would generally say around here, coming up with a phonology -- you can do as you like.  A conlang phonology in which all sequences of (optional Y or W) + (one of A E I O U) + (optional Y or W) behave identically is entirely believable.  Calling each of these sequences "a vowel" for the purposes of describing that language is also not unreasonable.  

>find no objection with a conlang using a sound sequence like "an
>youth" CaN You see what I'm driving at here?

Do, however, keep straight "describing what's pronounçable in my next conlang" from "describing what's pronounçable" tout court!  More or less any combination of segments, no matter how horrible, càn be pronounced by the suitably limber-tongued; what you care about is whether your conlang's hypothetical speakers are up to it.  

I would also warn that substituting "describing what's pronounçable in my next conlang" by "describing what _I_ can pronounce" is liable to lead to unexciting phonologies which are quite close to English's, but given your scorn for mouth noises I suspect this isn't particularly a concern for you.