Abstractness (Was Re: [j] v. [i])

--- In cybalist@yahoogroups.com, Robert B Wilson <han_solo55@...>
wrote:

>
> On Thu, 05 Jun 2003 16:42:54 +0200 Piotr Gasiorowski
> <piotr.gasiorowski@...> writes:
> >
> > ----- Original Message -----
> > From: Jens Elmegård Rasmussen
> > To: cybalist@yahoogroups.com
> > Sent: Thursday, June 05, 2003 4:20 PM
> > Subject: [tied] Abstractness (Was Re: [j] v. [i])
> >
> > Modern phonology
> > demands,
> > however, that phonological representations should be "natural",

i.e.

> > reasonably surface-true. To use a real-life example, Englidh /h/

and

> > /N/
> > ("ng") are in complementary distribution, but no-one has ever

dreamt

> > of
> > simplifying the system by assigning them to the same phoneme.
>
> actually, people have... why would we know that they are in

complementary

> distribution if no one ever dreamt of assigning them to the same

phoneme?

I expect the complementary distribution became apparent when
someone mechanically went through all the pairs of consonant
phonemes. Does anyone know the history of the observation? Why
aren't /w/ v. /N/ and /y/ v. /N/ considered problem pairs? Is it
because some Americans would actually accept 'low' v. 'lung' and
'toy' v. 'tong' (or, using a word of limited distribution, 'dray' v. 'dren=
g')
as minimal pairs?

Richard.