Newsgroups: sci.lang,alt.uu.lang.misc,alt.usage.english
Path: cantaloupe.srv.cs.cmu.edu!bb3.andrew.cmu.edu!news.sei.cmu.edu!cis.ohio-state.edu!math.ohio-state.edu!howland.reston.ans.net!vixen.cso.uiuc.edu!sdd.hp.com!hplabs!hplntx!rigel!evan
From: evan@hplerk.hpl.hp.com (Evan Kirshenbaum)
Subject: Re: Th - one or two phonemes in English?
Sender: news@hplabsz.hpl.hp.com (News Subsystem (Rigel))
Message-ID: <D2KtzM.MoM@hplabsz.hpl.hp.com>
Date: Wed, 18 Jan 1995 01:11:45 GMT
Reply-To: kirshenbaum@hpl.hp.com
References: <rharmsen.132.000FF592@knoware.nl>
Nntp-Posting-Host: hplerk.hpl.hp.com
Organization: Hewlett-Packard Laboratories
Lines: 40

In article <rharmsen.132.000FF592@knoware.nl>,
Ruud Harmsen <rharmsen@knoware.nl> wrote:
>The (inter)dental sounds dh and th in English are distinguished by all
>speakers, but simple rules with few exceptions can accurately predict
>which sound is used in any given word. Because grammatical category and
>some etymology are needed for the rules, this may not be enough reason
>to conclude that there is only one combined th-dh phoneme in English.
>Yet there are very few minimally differing pairs.
>
>
>The rules for prediction of th/dh, with examples, are as follows:

[Rules omitted]

Correct me if I'm wrong (it's been a while since I took phonology),
but I didn't think that you were allowed to appeal to syntax or
semantics when deciding the phoneme inventory of a language.  If two
segments could occur in the same phonologic environment, then they had
to be counted as distinct phonemes.

Thus the presence of minimal pairs such as /loUT/ : /loUD/ or even
/TIn/ : /DIs/ means that /T/ and /D/ must be different phonemes.  In
fact, appealing to the category in the first example makes no sense,
as the phonemic difference is all that allows the addressee to decide
whether the word is an adjective or a verb.

The reason that you are not allowed to use syntax or semantics is that
traditionally the mapping between phonemes and segments is held to be
an independent step following (in production) syntax and morphology
(which may or may not be independent) and preceding them in recption.
I don't remember what the evidence for this was, but it was convincing
even to a student of non-transformational syntax such as me.

Evan Kirshenbaum		       +------------------------------------
    HP Laboratories		       | Giving money and power to government
    3500 Deer Creek Road, Building 26U | is like giving whiskey and car keys
    Palo Alto, CA  94304	       | to teenage boys.
				       |                   P.J. O'Rourke
    kirshenbaum@hpl.hp.com	       | 
    (415)857-7572
