Newsgroups: sci.lang
Path: cantaloupe.srv.cs.cmu.edu!bb3.andrew.cmu.edu!nntp.sei.cmu.edu!news.psc.edu!hudson.lm.com!news.math.psu.edu!chi-news.cic.net!newsfeed.internetmci.com!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: <DG4ysI.7JE@eskimo.com>
Sender: news@eskimo.com (News User Id)
Organization: Eskimo North (206) For-Ever
References: <AC7F815C966813DB1@yarn.demon.co.uk> <rte-0210951046590001@mac-118.lz.att.com> <DG03p4.EJA@eskimo.com> <rte-0610951143500001@mac-118.lz.att.com>
Date: Sun, 8 Oct 1995 15:24:17 GMT
Lines: 60

In article <rte-0610951143500001@mac-118.lz.att.com>,
Ralph T. Edwards <rte@elmo.lz.att.com> wrote:

>In article <DG03p4.EJA@eskimo.com>, rickw@eskimo.com (Richard Wojcik) wrote:
>> Phonemes are phonetic units (or classes of units) that function in a
>> language to distinguish morphemes or words.  Early phonemic theory (roughly
>> 1880-1930) assumed that phonemes could be completely neutralized in speech
>> from a phonetic point of view.  For example, one could take the "flapped
>> /t/" in `writing' and the "flapped /d/" in `riding' as a phonemic overlap
>> or neutralization.  This was what Bloomfield did.  After the turn of the
>> century, linguists began to challenge the notion of phonemic
>> neutralization, and phonemes actually came to be defined as phonetically
>> unambiguous at one point.  Many linguists today still use the term
>> "phoneme" to describe a phonetically unambiguous sound unit or category.
>
>If I've understood you correctly, the debate is about whether the word
>phoneme should be used to describe the mental symbols used to store the
>deeper morphemes /rait/ for write, (or /to:d/ for the German word for
>death) or a later stage of linguistic production after rules for context
>have been fully applied.
>/rait/ vs. /rai<flap>IN/ or /to:t/ vs. /to:d@s/.

The flap is normally considered allophonic in American English, so it
wouldn't be taken as a phoneme by most linguists.  When speakers do notice
flaps, they tend to categorize them as the phoneme /d/.  So you would
not represent "writing" with flapped [D] at the phonemic level.
Furthermore, what you call "rules for context" would apply both before and
after shallow phonemic representations.  Just the same, you have understood
the gist of what I was saying.  

Let's take your German example as illustrative of the basic controversy.
The progenitor of phonology, Baudouin de Courtenay, held that the [t] in
[to:t] (i.e. German "Tod" or 'death') was really the phoneme /d/ in a
voiceless variant.  He did not identify it with the German phoneme /t/.
Later on, in the 20th century, one of Baudouin's students, Shcherba, came
to identify the devoiced /d/ with the phoneme /t/.  This caused a real
split among Russian linguists.  Those from the "Moscow School" tended to
side with Baudouin's original formulation, but those from Shcherba's
"Leningrad School" went the way of shallow phonemes.  It is not well-known
in the West, but Shcherba had considerable influence on phonemic theory in
England, which strongly influenced other schools, too.  That is one of the
reasons that virtually all schools of phonemics outside of Eastern Europe
came to adopt the shallow version of phonemic theory--the one that does not
allow phonemes to have allophones that look like other phonemes.  (The best
known Western linguist to retain Baudouin's original usage was Sapir, who
was roundly criticized for it.)

>It does seem we should have separate words for the two concepts, both seem
>real. 
>I always assumed it meant the latter, but I see the need for some word for
>the former.

Right.  The best term for the deepest level of "phonemic" representation
would probably be just "phonological representation".  You have to be very
careful in using the term "phoneme" because so many people mean so many
different things by it.  Then again, they mean different things by
"phonological representation", too.  ;-)
-- 
Rick Wojcik  rickw@eskimo.com     Seattle (for locals: Bellevue), WA
             http://www.eskimo.com/~rickw/
