On 16/04/09 00:19:04, Paul Kershaw wrote:
> ----- Original Message ----
> > From: Philip Newton <[log in to unmask]>
> > I think the point is not "those aren't phonological words" but
> "those
> > words do contain vowels". In the sense that [S:] (when you want
> > someone to be quiet) is a single vowel, I suppose.
> I agree, but I think we can quickly get into the theoretical weeds.
> Given these two statements:
> 1. [S] is a meaningful, independent morpheme in X language.
> 2. All words must contain one phonological vowel.

That's not like what Ray or Phillip have said at all. The word 
"phonological" in "phonological word" does an entirely different job 
from the one you're trying to give to "phonological" in "phonological 

A phonological word is the phonetic output. For instance "the" or "'s" 
are syntactic words in English, but they aren't *phonological* words; 
you can't just say "'s", and it forms a single phonological word with 
the last phonological word before it.

> Do we (a) stretch the definition of "phonological vowel" to included
> /S/, thus having two definitions of "vowel" depending on whether 
> we're
> in phonological (i.e., mental representation) space or phonetic 
> (i.e.,
> articulatory) space*; (b) assume an underlying phonological vowel
> (e.g., /8S/) and a corresponding rule for dropping it when
> articulated; 

So given the explanation above, that doesn't remotely help the 
proposition that every phonological word contains a vowel, because if 
it's been dropped, it's just not there!

>or (c) change the framing of the rule in (2)? Occam 
> would
> tell us to do (c), I'd think. Doing (a) feels like making the data
> conform to the rules, not vice versa, which is bad science and 
> weakens
> whatever relevance the consonant/vowel distinction had in the first
> place.
> -- Paul
> * While [S:] is not strictly speaking a morpheme of English, it does
> have a consistent meaning ("Hush!"), as do [m:] ("I'm thinking"),
> [n:h] ("Darn it!"), [ps:t] ("Hey! You!"), and even [!!!!!] ("Shame on
> you!").

[m:] and [n:] are both easy to analyse as vowels. Look at them in a 
spectrogram; they have lots in common with say [a~] or [i~]. In terms 
of articulation, the sound is produced in the same way for [m] and [a]; 
it's merely modified in a different way. There doesn't seem to be any 
hard-and-fast boundary.

[S] on the other hand, the sound is produced in a completely different 
way, and the resulting sound is of a completely different nature.

Also, the idea that there's some categorical difference between 
consonants and vowels is refuted by the glides one can get like [j] and 
[w], which are essentially vowels sitting in the part of the syllable 
structure meant for consonants. Ray's earlier statement is a good basic 
starting point.

But still it doesn't work. If it's true, then one would say when I say 
the name of the letter "t" on its own, I say [t_hIi], but if I say the 
sound of the letter "t" on it's own, I say [t@_0_X] even though there's 
no appreciable difference between the release and aspiration of the "t" 
in the letter's name and the so-called vowel in the letter's sound.

It begins to make a mockery of vowels, with them as little more than 
carriers, and only recognisable in retrospect. I think this means it 
has little place in science.