Newsgroups: sci.lang
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!oitnews.harvard.edu!purdue!lerc.nasa.gov!magnus.acs.ohio-state.edu!math.ohio-state.edu!howland.reston.ans.net!news.sprintlink.net!in1.uu.net!eskimo!rickw
From: rickw@eskimo.com (Richard Wojcik)
Subject: Re: Power spectrum of phonemes?
X-Nntp-Posting-Host: eskimo.com
Message-ID: <DFs1y1.C59@eskimo.com>
Sender: news@eskimo.com (News User Id)
Organization: Eskimo North (206) For-Ever
References: <AC7F815C966813DB1@yarn.demon.co.uk> <rte-2209951437010001@mac-118.lz.att.com> <DFJtKr.AEI@eskimo.com> <rte-2709951124530001@mac-118.lz.att.com>
Date: Sun, 1 Oct 1995 16:03:36 GMT
Lines: 90

In article <rte-2709951124530001@mac-118.lz.att.com>,
Ralph T. Edwards <rte@elmo.lz.att.com> wrote:
>In article <DFJtKr.AEI@eskimo.com>, rickw@eskimo.com (Richard Wojcik) wrote:
>> In article <rte-2209951437010001@mac-118.lz.att.com>,
>>  [snip]
>> 
>> >Incidentally use of phonemes is directly comparable to digital waveforms in
>> >information transmission between machines.  An infinite number of analog
>> >waveform shapes are mapped to a sequence of symbols, each of which is one
>> >of a finite number of siginificantly different symbols.
>> 
>> I doubt this. 
>
>What precisely do you doubt?

That "use of phonemes is directly comparable to digital waveforms in
information transmission between machines."  If it were, then the
development of speech understanding systems would have turned out to be
much easier than it has.  I grant you that many researchers (most of whom
have not studied much phonological theory) have adopted this assumption.

>> Psychological phonemes might not have any physical
>> manifestation in the speech stream.
>
>Huh?

A rather large class of phonological processes have evolved to reduce and
simplify articulation.  For example, English speakers often articulate a
2-syllable word such as "police" /p@lis/ as a single syllable [plis].  One
would say in this case that the schwa has no physical manifestation in the
speech stream.  Consider intervocalic flaps as another example.  Often they
are deleted in casual speech.  For example, the word "kitty" is often
pronounced similarly or identically to "key" ([kIi] or [ki]).  The /t/
phoneme would be said to have no physical or phonetic manifestation in that
pronunciation.  

>>  Whether or not you insist on phonemes
>> having a unique phonetic identity is, again, a matter of theoretical
>> viewpoint.  
>
>Huh again.  Let's try again.  A large number of distinct utterences, varying
>from individual to individual, dialect to dialect, and by context map to a
>finite muber of perceived phonemes (for each individual).  These phonemes
>become so real to their users that many cannot conceive of other phonemic
>distinctions. 

All I meant by my statement was that the definition of phonemes has been
controversial.  That is a matter of historical record.  Not everyone agrees
(or has agreed) that phonemic distinctions can be neutralized in speech.  I
suspect that your difficulty in interpreting my statement may not be that
we disagree on the essentials, but that you are working off of the term
"phoneme" as used in the signal processing literature.  If so, then we are
equivocating on the term and ought to be very careful.  The term "phoneme"
means something very different in the linguistic literature.

>...[snip]... ?no physical
>manifestation?  What can you possibly mean?  If they have no physical
>manifestation text to speech or speech recognition based on a finite
>number of phonological symbols must fail.  Do you propose that my mental
>constructs are communicated to my interlocutor's by telepathy?:->

No, although I'm closer to that position than you think.  ;->  I would not
characterize speech recognition and speech synthesis as failed exercises.
There are some remarkably impressive systems out there.  However, they
don't yet perform up to human standards.  We do not yet know how to use all
of the different sources of information (pragmatic, semantic,
morphological, etc.) that humans use to process speech.  Machines just
don't "hear" phonemes the way humans do.  To give you a better sense of
what I'm talking about, I'll put up a short paper that Jim Hoard and I
wrote a few years ago.  It'll be at http://www.eskimo.com/~rickw/bern.html.

BTW, I am not challenging your claim that there is a finite number of
phonemes for any given person.  I am unhappy with your equating of human
phonological processing with phonetic segmentation performed by machines.
I believe that all machine-based algorithms today simply store all of the
allophonic variation (encoded as "phonemes" [sic]) for words in the
lexicon.  Humans seem to require considerably less memory to store words,
because they only need to assign a single phonemic representation to a
morpheme (ignoring allomorphy) in memory.

>This infinite to finite mapping is the essence of digital communication,
>between humans or machines.

That does not mean that machine-based and human processing can be directly
compared.  Analogies are sometimes enlightening and sometimes misleading.
This one is misleading in that it ignores the role of phonology in human
speech processing.
-- 
Rick Wojcik  rickw@eskimo.com     Seattle (for locals: Bellevue), WA
             http://www.eskimo.com/~rickw/
