LINGUIST List 2.730

Thu 31 Oct 1991

Disc: R-Linking

Editor for this issue: <>

Directory

1. Jim Scobbie, Re: 2.707 R-Linking
2. Jim Scobbie, Re: 2.707 R-Linking
3. Richard Goerwitz, Re: 2.720 R-Linking

Message 1: Re: 2.707 R-Linking

I don't think I made myself very clear at one point:
>>From: Geoffrey Russom <EL403015brownvm.brown.edu>
>>My understanding is that you get r-linking primarily to avoid a hiatus
>>between vowels. With tense "u," there's an off-glide "w" to avoid the
>>hiatus, so it's not surprising to find that Ellen doesn't get r-linking
>>in that environment. ... But a word-final "-r" after
>>tense "u" amounts to a final consonant cluster [wr], ...
>My understanding of /r/ suggests the above is correct.
This cryptic comment is meant to convey the idea that /w/ and /r/ are
constrasting glide elements of the vowel system, and that you can't get
both.
The system doesn't allow /Vwr/ or /Vrw/ as vowels, only /Vw/ /Vr/.
I should have noted my change of [] brackets to // brackets: someone asked
about the [uw] vowel, where if Geoffrey had used /uw/ the confusion
may not have arisen. In my opinion, '[]' are generally used nowadays to
flag the alphabetic transcription of some speech, '//' are used to
flag the alphabetic shorthand for a theoretical representation. If they
aren't, they should be, damnit. ;-) --
James M. Scobbie: Dept of Linguistics, Stanford University, CA 94305-2150

Message 2: Re: 2.707 R-Linking

David Stampe writes:
>Isn't "intrusive r" just an underlying /r/?
Yes. :-)
>That's the analysis that
>Patricia Donegan describes in a forthc. book on historical phonology
>ed. by Ch. Jones.
It's an analysis that crops up from time to time. I presented a paper
myself at the LAGB meeting a few years back which argued for underlying
/r/ in words such as 'idea'. I also posted the basis of the argument to
sci.lang. I was adapting a Wells-ian/Firthian vowel system analysis in
a unification-based framework. I was also attempting to use a 'licensing'
approach to account for the distribution of [r] and [schwa], taking
parts of Goldsmith and Ito's work. COnsequently you'll not be at all
surprised to find that I agree with your conclusions:
>The hypothesis that intrusive r is actually an underlying r seems to
>fit the facts better than the alternative hypotheses I know of, viz
>(1) Analysis of intrusive [r] as a "linking" sound (a consonant
> inserted to avoid vocalic hiatus).
>(2) Analysis (e.g. by Bill Labov) of intrusive r as due to false
> analogy or hypercorrection:
>(3) Analysis (by Theo Vennemann) of intrusive r as "rule reversal":
The reference for the talk I gave at LAGB, and to the unpublished draft
paper:
Scobbie, JM (1989) The r-0 alternation of non-rhotic English. ms-draft.
Scobbie, JM (1990) A lexical account of [r]-sandhi in English. Paper
presented at Spring meeting of Linguistics Association of Great Britain on
Wednesday April 4th 1990.
Jim Scobbie
PS... For some interesting discussion of /r/ in some rhotic dialects
of English, see Tom Veatch's 1991 dissertation 'English vowels: their
surface phonology and phonetic implementation in vernacular dialects'.
--
James M. Scobbie: Dept of Linguistics, Stanford University, CA 94305-2150

Message 3: Re: 2.720 R-Linking

	I somewhat incongenial statement was recently made here in
response to David Stampe's bit on r-insertion:
 David Stampe makes a fundamental error in reasoning in his recent
 posting, arguing that the linking /r/ is "underlying". (I put
 'underlying' in quotes because, as a good declarative phonologist,
 I reject that notion as a snare and a delusion, but for the sake
 of the argument, I will pretend I know what is meant by this
 term.)
Every one of the innumerable schools of linguistic thought that have
sprung up out of the primaeval structuralist mud seems to think the
other schools a foolish ruse. It would be nice if we showed a bit of
mutual respect around here. Also, don't forget that we're not all
phonologists. Some of us (heaven forbid!) are philologists, and don't
even know what declarative phonology is.
	Anyway, the objection to Stampe's assertions seems to come
down to this:
 The error lies in assuming that, since r-insertion would be
 phonetically unnatural, it cannot be involved in a process which
 is characterized by a great degree of automaticity. But that's
 circular, since the correlation between phonetic naturalness and
 automacity was supposed to be a discovery, not a postulate, of the
 school of natural phonology.
Let's rephrase your statement a bit, and see if we can evoke an answer
to an interesting question, rather than just dump on David (who always
seems to stand at the eye of one or another storm): What evidence do
you (David) have that r-insertion is phonetically "unnatural"?
Presumably this evidence would have to overrule the specifics of the
r-insertion case in English (where, if the evidence were taken simply
on the basis of this one phenomenon, your argument would in fact be
circular).
	The objections to David's analysis continues on with a hint at
counterevidence:
 The linking r, as well as numerous other examples (such as the /n/
 that shows up between a word-final /t/ and a following
 word-initial /y/ [and perhaps elsewhere] in Korean), are simply
 counterexamples to this.
You know, not all of us know Korean here. Those that do probably only
know it superficially. And even if this is a valid analysis (which I
do not doubt), I do not see how dropping the velum after /t/ and
before /y/ in Korean brings David's theories crashing to the ground.
Nor does the following blanket statement:
 ...I think that, since in many other cases the claims of natural
 phonology seem to be brilliantly vindicated, the appropriate
 response to such counterexamples is not turn a factual theory into
 an a priori formal system, but to see what makes these
 counterexamples different and to modify the theory to account for
 them.
You mean as in "Why Phonology is not Natural" and what not? If you
feel the claims of natural phonology have been brilliantly vindicated,
then why bother with a response? You certainly haven't offered any
evidence yourself that David's analysis is unreasonable. If we are
supposed to know, a priori, that it's unreasonable, then I don't
really see why you felt the need to post. I feel a bit bad to put you
under the flame thrower, but your response really offers very little
in the way of constructive criticism. It is merely a pooh-poohing of
David's particular theoretical framework.
	Incidentally, David has on many occasions - both on Usenet
(sci.lang), and in papers he has given - tried to respond to these
criticisms. Have you read any of these responses? It appears you are
readily willing to admit you have not. It is true that Stampean
papers are not always as accessible as one might like. This does not
prevent us, however, from addressing his ideas seriously here, and
from making reasonable requests for clarification on points that might
seem to cloud, or even undermine, his position. David's certainly
heard the sorts of objections you raise before. Ask him nicely, and
I'll bet he'll tell you why he's still a natural phonologist :-).
d