Roger Mills wrote:

> I find this quite well organized, logical, and probably neater looking than
> XS, but one would want to see lots of it in use for a final judgment.

A couple of proposed changes to the draft are suggested herein.

To begin with, I'm thinking of swapping around the symbols for extra
short [.] and non-syllabic [\.] - logically the monograph should be
whichever is most often used, which on reflection is probably the

> --the abandonment of some cases of IPA=Roman, such as c j, and IMO worst of
> all x X =  your lateral fric. Though I can't offer an alternative.

Fits [j] has the same value as IPA [j]

Getting rid of [c] frees it up to be a vowel :-)

As for [x], you'd have a hard time persuading me that [K] is logical :-)

If it helps, you could think of [x] as being the shape you'd get if
you took the Welsh "ll" and put one of the "l"s over the other, in
much the same manner as the Roman numeral X represents two hands
crossed over at ninety degrees.

> Incidentally I don't see anything for voiceless l (unless that's your "x"?),
> or other such possibilities-- would you have to use the voiceless diacritic
> "l\f", similarly r\f , m\f , a\f etc.??

My IPA chart doesn't have anything for voiceless l.

> --You're presumably following IPA, but why have a separate symbol for
> velarized L (#) rather than using your "velarized" diacritic [l\K]

Because x-sampa does (fits [#] = xsampa [5])

> --Most major carp:   please please please introduce ' for primary, " for
> secondary stress.  One tick, primary; two ticks secondary-- get it?? isn't
> that logical???? :-))))

Absolutely not!!! :-) Two tick, greater stress, one tick, lesser stress.
Logical, because primary stress is louder than secondary stress.
This may be something where we'll have to agree to disagree. I'm quite
definite and unshakeable in my opinion on this matter.

> On the vowels, a minor question:  if R is the vowel of "heard", what about
> r-less speech???  If I recall my IPA, [3] is lower-mid, about at the same

As I explained in my commentary, I'm using [R] as a generic non-schwa
mid central vowel, since when transcribing words like "heard", people
seem to use anything from xsampa [3:] to [8:] pretty much

Perhaps I should set [R] to be rounded schwa (my "heard" vowel is
definitely rounded), and set:

(proposed change)

  xsampa    fits
  ------    ----
    @\       @-
    3        @+
    8        R-
    3\       R+

> (hence the use of epsilon for E, reversed eps. for 3, [and "3" thus makes
> sense, though I know you want to avoid using numbers]).

Well, avoiding numbers frees them up for possible use in tones.

> I take it % represents V??? Where then is 6 (IPA inverted a), a low vowel??

xsampa [6] corresponds to fits [%]

To review, xsampa [V] is the unrounded counterpart of [O], but for
historical reasons often (mis)used in English transcription in place
of [6]. The fits equivalent of true xsampa [V] is [Q], just as the
equivalent of xsampa [7] is [q].

> What is "(      A)" supposed to mean?

It means that the rounded counterpart of [a] is [(]
and the rounded counterpart of [A] is [)]

> What would be its rounded
> counterpart, XS Q?? Low [&] also has a rounded counterpart, admittedly rare,
> but not indicated in your chart.  Perhaps you need a "rounded" diacritic???

There is no rounded counterpart to xsampa [{] cxs {&] on my IPA chart.

To summarise the above

  xsampa   fits
  ------   ----
    6        %
    &        (    (note: x-sampa [&] not kirsch/cxs [&])
    A        A
    Q        )

> A final minor carp:  why not retain = for "syllabic"??

I've allocated | to xsampa [M] and = to xsampa [U_c], but since the
latter doesn't actually have an IPA (or xsampa) value, you're quite
right that I could free up [=]. It's just more symmetric to have a
symbol for unrounded [U]).