I wouldn't say that apical trills exist
because there is eternal equilibrium between [r] and other rhotics. They exist
because the common-or-garden apical trill is an important stage in the natural
life cycles of consonants.
Apical taps or flaps often cooccur with
trilled [r] as positional or allegro-speech allophones of /r/. A tap can be
regarded as a minimal, one-phase variant of the corresponding trill.
However, if [r] changes into a completely untrilled or uvular rhotic, this
certainly represents an irreversible or hard-to-reverse qualitative shift. But
then all lenitions are like that -- and if it's true that uvular [R] is for
some reason easier to pronounce than [r], or cheaper in terms of effort and
neuromuscular control, then [r] > [R] _is_ a kind of lenition. [R] may be
further untrilled, producing dorsal fricatives or glides. There are also other
well-known "decay series" involving trills, e.g. alveolar trill > alveolar
tap > apical approximant > zero; or alveolar trill > postalveolar
flap > retroflex approximant > zero. Such processes, as well as
uvularisation, would have eliminated all the apical trills in the world's
languages a long time ago if lenition, or the principle of least effort, were
the only factor that counts in linguistic evolution.
But as old [r]'s succumb to gradual erosion
(rather slowly, since [r] is a relatively stable consonant diachronically), new
[r]'s come into existence. First, rhotics can themselves be products of lenition
affecting obstruents or nasals -- e.g. [t] > [d] > [D] (in this posting,
[D] = alveolar tap). Next, if lenition is common in syllable-final and
intervocalic positions, strengthening is equally common morpheme-initially
or at the beginning of a stressed syllable -- that is, in contexts which
typically require phonological reinforcement. Strengthening may transform taps
or voiced alveolar fricatives into their phonetically enhanced
counterparts, i.e. trills -- this is how rhotacism usually works. Strong trills
may also develop via assimilation from sequences like [-tD-] or [-Dz-]. Finally,
rhotics may develop from laterals in dissimilatory environments.
In Scotland, by the way, /r/ is mostly a
postalveolar approximant or tap as in most of northern England;
trilled [r] represents an occasional patriotic enhancement thereof. Dorsal
rhotics are all but extinct in Britain, even in their Northumbrian focal area.
But even without Scotland, the vast majority of European languages still have
the apical trill with the usual minute proportion of various idiosyncratic
pronunciations -- here spreading, there receding on a local scale. Uvularisation
seems to be an areal phenomenon diffusing through imitation across France
and much of Northern Europe, but it isn't general even in German-speaking
countries.
Piotr
----- Original Message -----
Sent: Thursday, June 28, 2001 6:30 PM
Subject: [tied] Re: uvular R
Against your steady
state uvular/trilled /r/ theory, I would venture a big bang uvular /r/ theory:
According to my phonology teacher, Hans Rischel, there was a tendency away from
trilled to flapped apical or uvular /r/ in Europe (Note that Scotland and North
Sweden still have trilled /r/. The standard "explanation" for this tendency is
that
since people spend more time indoors, there is no need for a noisy
trilled /r/. But there is such a thing as a subclinical poisoning, ie. without
symptoms serious enough that any doctor would care.
Torsten