>>Also, what are [S] and [Z] doing as post-alveloar fricatives? They
>>used to be palato-alveolar fricatives. Has the meaning or the
>>classification been changed?
> The terms "palato-alveolar" and "alveolo-palatal" are not very informative.
They aren't and while they are retained informally, the current tendency
in more technical descriptions is to use less ambiguous terms.
> I'll paste from a post I coincidentally had just written to sci.lang:
> <begin quote>
> Ladefoged & Maddieson's thorough analysis of the s(h)ibilants reveals that
> the following parameters can play a role:
> 1 blade of the tongue (laminal)
> 2 tip of the tongue (apical)
> 3 underside of the tip of the tongue (sub-apical)
And what about the anterior part of the dorsum?
> place of articulation:
> 1 the upper teeth (dental)
> 2 the alevolar ridge (alveolar)
> 3 behind the alveolar ridge (post-alveolar)
> 4 the hard palate (palatal)
One might also want to squeeze "prepalatal" between "postalveolar" and
"palatal". Polish <s', z', c', dz'> are (antero-)dorso-prepalatal rather
> the shape of the tongue (blade/back):
> 1 grooved \ dental,
> 2 flat / alveolar
> 3 domed \ post-alveolar,
> 4 palatalized / (palatal)
> the position of the tip of the tongue (laminal only)
> 1 against the lower teeth (closed)
> 2 leaving a sublingual cavity (open)
> <end quote>
> Instead of "palato-alveolar" (English <sh>), it's better to use [rounded]
> [apical or laminal] domed post-alveolar. Instead of "alveolo-palatal"
> (Mandarin <x>, Polish <s'>), [apical or laminal] palatalized post-alveolar.
> Instead of "retroflex" (Mandarin <sh>, Polish <sz>) [apical or laminal]
> flat post-alveolar.
>>For voiced dental, alveolar and post-
>>alveolar fricatives, J. D. O'Connor had [D], [z] and [r\], and uses
>>a lowering hook to convert them to frictionless continuants.
> But [r\] (like [j]*) is defined, at least since IPA included an approximant
> row, as a frictionless continuant. You use the raising hook to make it
> into a fricative (as in e.g. English dr-). The sound in <tree> is then a
> reversed r with raising hook and voicelessness ring.
> * the curly-j is therefore equivalent to [j] plus raising thingy.