Quoting Tristan McLeay <[log in to unmask]>:

> On Mon, 26 Jan 2004, Andreas Johansson wrote:
> > Quoting Tristan McLeay <[log in to unmask]>:
> >
> > > Unless of course you use the approach where /a2)/ is the diphthong and
> > > /a_2/ is /a/ on the second tone
> >
> > If I use that approach and my audience is aware of it, yes!
> Given all the problems with /a_2/-for-diphthong, I think it almost make
> more sense to remove it as an option in CXS.

I rather agree.

> > > (IIUC, [a_2] wouldn't make sense because
> > > there's no phonetic second tone).
> >
> > I'm positive I've heard 'bout a second tone in Chinese, and tone pretty
> much
> > has to be phonetic, hasn't it? Anyone sinaically enlightened feeling like
> > commenting?
> Well, what I meant was that the phonetic realisation of tones will differ
> from language to language, so there's no 'phoneitc second tone', only
> phonemic ones. A phonemic second tone might be _H or _L or _H_R or
> something. Or so it seems to me, I might be wrong.

Well, we've clearly been speaking past one another here a bit. If you use the
X-SAMPA's allowance for using traditional language-dependent tone numbers, of
course a given tone number specifies a specific phonetic content for that
language, but you're of course right there is no "absolute", god-given second

ObTangent: On the subject on refining the CXS, what would people say about
completifying the correspondence between the alveolar and dental series by
reassigning [r`] to a retroflex trill and introducing [r\`] for the retroflex
approximant? Besides making it easier to transcribe my 'lect, it does, IMHO,
improve the mnemonic properties of the scheme. I've been using that, in the
mistaken belief it was kosher X-SAMPA, anyway, without anyone complaining!