Re: CHAT: XS vs. Kirshenbaum vs. Who-knows-what

Tristan McLeay wrote, quoting myself:

> > I've mentioned in reply to Roger that [x] is no less logical a choice
> > than [K]. X-sampa also uses [F] and [J] for nasals. I find that far
> > less logical than most of your reservations about my scheme.
>
> Well, yes, it is a less logical choice. x is a character of the IPA, K
> isn't. So doesn't it make more sense to map IPA x to Ascii IPA x, even if
I mentioned earlier that I was not going to allow the shapes of IPA
characters to dictate my choices of how to represent them.

Trying to imitate the shapes of IPA characters is a perfectly sensible
design priority if that's the way you like it, but that doesn't make
mutually exclusive *alternative* design priorities any *less* sensible
or logical. Just different. If you try to incorporate mutually
exclusive design priorities into the same scheme, you tend to get the
committee effect.

For me, the shapes of the underlying IPA characters are of very little
importance. A much higher priority was making the table of (e.g.)
consonants relatively easy to read, digest and remember by taking
advantage of mnemonics relating to the ascii characters themselves
*without* having to refer to original IPA characters.

> we do something utterly silly for IPA symbols for which there's no simple
> translit? Not to mention the fact that x is a nice, simple, single
> character, K is a more complicated double character (typing-wise, it's
> shift+x), whereas CXS [x] (FITS [K]) is used a lot more frequently than
OK, so your ideal scheme would have the following design priorities

  - use characters that look like IPA characters whenever possible
  - use characters easiest to access on keyboard for most common phones

That's OK. Mine doesn't. I'm not too worried about the use of the
shift key because I assume people can touch type.

> CXS [K] (FITS [x]). Sure, you might argue that X is no more logically
> applied to an l-ezh ligature than a chi, but I think you can't find any
> argument in favor of the change other than ... other than the fact that
> you felt like it?
My consonant table is much more regular. You can change palatal, velar
or uvular plosives into fricatives simply by converting to upper case.
That sort of relative regularity, with plenty of compromises but not
too many, finds favour with my tastes.

> > > [{]  begin double articulation or affricate (when necessary)
> >
> > Is that prefix, a postfix, a roundfix or an infix?
>
> The line you quote explicitely answers your question. Did you not
> notice my use of the word "begin"? "Begin" means "at the beginning",
> i.e. prefix, no? As in [{tS]
Oh, selective reading. Sorry bout that.

> As I recognised ('... certain mnemonic properties ...'). The problem is
> also aesthetic, of course, but that hardly counts for much... But how
> would you the equivalent of [d&:n(t)s]? Would you just have to say
> '[d&:ns] or [d&:nts]'? Even '[d&:ns], [-nts]' seems awfully long-winded.
I'm not sure what you're trying to say. You seem to be asking me
something about x-sampa [d{:n&tQs] (I don't have a cxs chart in front
of me, sorry).

All other points of view you express further down have been dealt
with in general terms above.

Adrian.

Replies