Re: Your first "linguistic" memory

Ekkehard Dengler wrote:
>
> Ruud Harmsen schrieb:
> > Sun, 15 Jan 2006 13:34:24 GMT: "Peter T. Daniels"
> > <grammatim@xxxxxxxxxxxxxxxx>: in sci.lang:
> >
> > >Ruud Harmsen wrote:
> > >>
> > >> Sat, 14 Jan 2006 21:35:36 GMT: "Peter T. Daniels"
> > >> <grammatim@xxxxxxxxxxxxxxxx>: in sci.lang:
> > >>
> > >> >No, pronuciation variants are differing realizations of the same
> > >> >phoneme. The two pronunciations of "economics" involve substitutions of
> > >> >a phoneme.
> > >>
> > >> These two statements seem contradictory to me.
> > >
> > >The two pronunciations of "economics" are not "pronunciation variants,"
> > >but an example of "free variation."
> >
> > I don't understand the difference.
>
> Neither do I. In fact, I doubt that such a clear-cut distinction exists.

I don't see what the problem is. Free variation is when anyone
recognizes that there are two different pronunciations of a word:
either/either, economics/economics, etc., and either one can be used by
any speaker with no difference in meaning. Since they are recognized as
two different pronunciations of the same word, they involve different
phonemes. In "either," the phonemes are /iy/ vs. /ay/; in "economics,"
the phonemes are /iy/ vs. /e/ (i.e. [E]).

In [sit] vs. [sith], no naive English-speaker would be aware of the
difference; there is no free variation. There are different realizations
of the final /t/.

Sometimes unconscious variation is conditioned, as in English [#th] vs.
[#st]. Sometimes it isn't.

> I've got a feeling we're all talking at cross purposes, Ruud, and I've
> probably added to the confusion. While I don't think the term "pronunciation
> variants" says much about the phonological status of the sounds involved,
> the term "free variation" does have a basic meaning most people seem to
> agree on. Here's one definition
> (www.departments.bucknell.edu/linguistics/lectures/05lect06.html):
>
> <Sounds do not represent two separate phonemes if they are in "free
> variation"; that is, if you may use one in any position you may use the
> other without any semantic effect. For example, aspiration may be omitted
> from stops at the end of words in English [...], however, whether it is
> dropped or not is indifferent; the meaning of the word does not change.>
>
> So far, so good. A logical problem arises, however, if this definition is
> taken to include lexically restricted variation. Let me quote from David
> Crystal's "Dictionary of Linguistics and Phonetics":
>
> <"free": The main use of the term is in the phrase "free variation" in
> phonology, referring to the substitutability of one sound in a given
> environment, with no consequent change in the word's meaning, as when a
> speaker articulates a word like "sit" with an unreleased or released
> plosive, or different pronunciations are given to "either" (/i:D@/ v.
> /aID@/). These different phonetic realisations of a phoneme are called "free
> variants".>

This could only have been written by someone for whom the concept of
"phoneme" is no longer operative. I'm afraid I must consider it the
first mistake I've ever been aware of in a work by Crystal.

> Now the fact that the meaning of "either" doesn't change if you replace one
> vowel with another which happens to be a typical realisation of another
> phoneme seems difficult to reconcile with the whole concept of phonemes.
> After all, the reason for treating /i:/ and /aI/ as separate phonemes in the
> first place is precisely that [i:] and [aI] aren't normally interchangeable.
> (Incidentally, note Crystal's use of slashes.)
>
> There are various ways of explaining this contradiction away. You can
> postulate an underlying phoneme which for some reason can be realised in
> either of two discrete ways although the phonetic context remains unchanged.
> This idea is implicit in the above quotation from Crystal, unless I
> misunderstand. However, suspending the rules in this way raises the question
> of why "see" never sounds like "sigh", for instance. After all, words like
> "either" are the exception, not the rule.

Or you can postulate that Crystal is conflating two different things,
phoneme substitution and conditioned (or not) alternations.

> I personally think /i:D@/ and /aID@/ can be more usefully regarded either as
> variant word forms that differ by one phoneme or as belonging to alternative
> lexicons. I believe Wells treats these two pronunciations as phonemically
> different, putting them in the "fleece" and "price" lexical sets,
> respectively.

Of course they're phonemically different. But they can occur in any
single speaker's utterances without indicating any other difference at
all.

> Anyway, I shouldn't have implied that there's an uncontroversial dichotomy
> between allophonic and lexically restricted variation, since for some, the
> two concepts are compatible.
--
Peter T. Daniels grammatim@xxxxxxx
.

Relevant Pages

  • Re: Your first "linguistic" memory
    ... This isn't fundamentally different from saying that a phoneme ... >>> Meaning is important. ... > Two or three possible pronunciations of the same word. ... > does that tell us about the essence of the notion of "phoneme"? ...
    (sci.lang)
  • Re: Your first "linguistic" memory
    ... >>> These two statements seem contradictory to me. ... the meaning of the word does not change.> ... taken to include lexically restricted variation. ... phoneme seems difficult to reconcile with the whole concept of phonemes. ...
    (sci.lang)
  • Re: Your first "linguistic" memory
    ... >> Ekkehard Dengler wrote: ... This isn't fundamentally different from saying that a phoneme ... >>> The fact that words often have more than one meaning doesn't really ... >>> isn't a necessary condition for a change in meaning. ...
    (sci.lang)
  • Re: Transcribing rhotics for ESL
    ... Changing a phoneme changes meaning. ... > Aspiration has no effect on comprehension. ... Look at your IPA chart (evidently your French students are more familiar ...
    (sci.lang)
  • Re: Tamil phonology
    ... pronunciations are acceptable). ... In the second word, the is not ... difference between and affect phonemic analysis? ... what one says about the phoneme system. ...
    (sci.lang)