Re: [tied] Re: uvular R

From: Piotr Gasiorowski
Message: 7772
Date: 2001-07-02

Anything is possible, but may not be highly probable. [d] is lower than [r] in any sonority hierarchy, which means that it is a stronger consonant. In general, thanks to the principle of least effort, lenition (weakening) is more common than strengthening except in some characteristic positions (e.g. stem-initially), where the principle of perceptual differentiation comes into play. The theoretic prediction is that in non-initial positions -d- > -r- should be a common type of change while -r-
> -d- will happen rarely if at all in spontaneous phonetic development.
Stem-initially things are different, and trilled [r] could in theory be strengthened into [d]; however, since in "strong" positions the phonemes in question would typically have reinforced (auditorily salient) allophones, confusion -- and therefore direct change -- would be unlikely. A more complex development, with [D] as an intermediate stage, can be envisaged, but in the change r- > D- > d- the first step would involve lenition and the second strengthening. The vagaries of linguistic evolution may surely produce a combination of favourable circumstances to make such a change possible, albeit not very often, judging from real-language data. One favourable circumstance would be the tendency of the inherited /d/ phoneme to have tapped allophones (i.e. [D] as a variant of /d/), preferably (unlike English dialects) also stem-initially -- and this is indeed what some languages have. In such cases we would simply need to assume the confusion with a weakened variant of the trill with an allophone of /d/ and its concomitant strengthening into [d]. I wouldn't exclude it, though to be sure I can't think of any real instance of such a change, whereas there are well-known cases of things going in the reverse direction, i.e. d- (~ D-) > r-.
----- Original Message -----
From: tgpedersen@...
Sent: Monday, July 02, 2001 12:50 PM
Subject: [tied] Re: uvular R

I wonder if that means yes or no? (As you had probably guessed, what
I really asked was whether /r/ > /d/ was possible?)