LINGUIST List 10.1936

Tue Dec 14 1999

Disc: What Exactly Are Allophones?

Editor for this issue: Karen Milligan <karenlinguistlist.org>

Directory

1. Jorge Guitart, Re: 10.1920, Disc: What Exactly Are Allophones?

Message 1: Re: 10.1920, Disc: What Exactly Are Allophones?

For Celso, Dan, Francisco, Benji and anybody else who is interested: a
little parable

There is a universe in which there are only two types of things you can
read: books and magazines, and no book looks like a magazine and no
magazine looks like a book.
Now there are five books and five magazines on the table. Each is a TYR
(for 'a thing you read'). 
No too books are identical and the same goes for the magazines.
I am interested in the theory of the organization of TYRs, which
is the same as the theory of the difference between books and magazines,
or TDBM for short.
Let B and M stand for book and magazine

If I pick up any one of the ten objects on the table and say 'This is a
TYR' I have not said anything relevant to TDBM, but if say 'this is a B'
or this is an M', that is relevant to the theory and the question that
interests me is, how did I know that it was a B or an M. It was because I
matched some physical property of the TYR with the category or type Book
or the category or type M, which are in my head. Let me call these types 
/B/ and /M/
/B/ is not a book and /M/ is not a magazine. They are the idea of book and
the idea of magazine.
Any TYR then is a token of either /B/ or /M/.

The question, What is a book? is equivalent to the question,
What counts as a TYR of the type /B/? and the question What is a magazine
is equivalent to the question, What is a TYR of the type M?
Suppose that I arbitrarily call TYRs 'phones' and types of TYRs
'phonemes'. Now any book is a variant of /B/ and
any magazine is a variant of /M/. Suppose further that the word variant
disappears from the language and is replaced by the word 'allophone'. The
question 'What is an allophone' is equivalent to the old question 'what is
a variant?' 
People have been asking on the list, What is a variant? The word 'variant'
is definable (look it up in your funk and wagnalls). Dan insists that we
cannot ask that question, that we cannot define the meaning of 'variant'. 
As for me, I am not interested in defining what a variant is. I already
know what it is. I am interested in the question, how do variants of the
same type are recognized as tokens of that type and not of some other
type.

Jorge Guitart