From: "Javier BF" <uaxuctum@...>

> OTOH, Czech r^ is a more complex articulation than
> the alveolar trill, because it involves a secondary
> articulation performed with the blade of the tongue,
> which is raised towards the postalveolar area narrowing
> the cavity next to the alveolar main point of articulation,
> producing the accompanying sh-like frication to the
> alveolar lowered trilling. So, from this other point of
> view, using the diacritic for raising makes sense. But,
> apart from introducing an apparent contradiction because
> the main -the alveolar trilling- part of the articulation
> does not undergo a raising but lowering, this point of
> view centered on a secondary part of the articulation
> is not the usual one for interpreting those diacritics,
> because a raised [T_r] is interpreted without doubt as
> a plosive interdental and never as lowering of the main
> point of articulation of [T] from a fricative to an
> interdental approximant with an accompanying _raising_
> of the blade of the tongue that narrows the cavity left
> further back in the mouth producing an accompanying
> sh-like frication.

Thus the misconception that Czech r-caron is pronounced [rZ] as though it
were two sequential sounds. The tongue isn't so much raised as it is held
against the hard palate with more tension, producing a trill-fricative. The
slight retraction cancels out the raising. But since the IPA symbol of
r-with-long-leg was removed, there's no better way of representing the
sound, since there's no separate symbol for a palatoalveolar trill (closest
thing is a retroflex flap and that won't work).

I think it should have its own IPA symbol once again, just like Japanese has
its own (the INVERTED long-leg l!) for /l\/, the alveolar lateral flap
traditionally romanized as r. If the same logic for Czech was applied to
Japanese, does that mean we have to use /r_l/ now?