CXS doesn't suck as much as SAMPA! (was Re: SAMPA sucks! (Or, a few samples of the Ridiculese...))

On Fri, 09 Apr 2004 17:10:16 -0500, Danny Wier <dawiertx@...>
wrote:

Starting to delve into the misty grey marshes that separate X-Sampa from
CXS...

> Two ways to construct affricates: for your Hungarian 'c', [t_s] and
> [ts)].
> Since the underscore/lowline is used primarily for secondary features,
> I'd
> recommend the right-hand parenthesis for affricates and double
> articulations
> like [kp)].
I use (and recommend) ^ before superscripts, _ before symbols that are not
superscripts, and ) to tie two letters (and )) to tie three letters, etc).

Thus /k^j/ for palatalised /k/, /l_e/ for velarised /l/ and /ts)/ for an
alveolar fricative.

I'm trying to find some x such that ^x and _x have clearly different
meanings. I'm drawing a blank. I suspect that if we're willing to leave
the strict IPA behind, we could find a few.

>> 4. X\o:_H_T|\|\6~_X_M
>
> Is it okay to use the tilde with consonants, as in [t~] as an equivalent
> of
> [t_e]
I'd say not, or at least not without special clarification. {~} is quite
unambiguously used as the nasalisation diacritic. No reason you can't vary
 from that, if you tell the reader what you're doing up front, though.

Did we ever resolve the symbols for the following vowels?

between /e/ and /E/
between /o/ and /O/
between /a/ and /A/

I suggest /E\/, /O\/ and /A\/ respectively (for sounds I've seen spelled
{smallcap-e}, {smallcap-omega} and {smallcap-a}), and wholeheartedly
endorse Danny's suggestion of /y\/ for consonantal /y/.

Oh, and just out of curiosity, how to you tell a {smallcap-o} from an {o}?




Paul