Return of the minimal pairs

[Robert Whiting]

On Mon, 14 May 2001 pausyl at AOL.COM wrote:

> On Thu, 10 May 2001 18:19:04 +0100, Larry Trask <larryt at cogs.susx.ac.uk>
> wrote:

>> --On Sunday, May 6, 2001 2:29 pm +0300 Robert Whiting
>> <whiting at cc.helsinki.fi> wrote:

>>> To make it clear what my statement was about:  My entire point was that
>>> [T] and [D] do not contrast in initial position in English.  My
>>> statement was not about whether [T] and [D] are phonemes in English or
>>> not.  The two are not necessarily related ([T] and [D] can still be
>>> phonemes even if they don't contrast in some environments).

>> OK; I'll bite.  I will argue that [T] and [D] *do* contrast in initial
>> position in English.  It is merely that we happen to have no good minimal
>> pairs for /T/ and /D/ in the language at present -- a completely different
>> matter.

>> To begin with, 'thigh' and 'thy' are a perfect minimal pair, *if* we accept
>> that 'thy' is a word of modern English -- which you may not want to accept.

<snip>

> I'm actually on Larry Trask's side (for the most part) in this
> discussion, but I assume that the argument from Robert Whiting's side
> would be best stated in terms of "closed-class" vs. "open-class"
> words; _thus_ would seem to be in the closed class.  _thy_ is surely
> English (think of the "Lord's Prayer"), but it's also a closed-class
> item.

Closed-class is part of the definition of function or grammatical words.
See Fowler, _Understanding Language_ (1974), 199.

> And I would judge the Traskian point above about /oi/ never occurring
> in native English words, even if true, to be irrelevant to the
> discussion:

Entirely.

> It seems to me that no amount of introspection would help an
> intelligent (but linguistically untrained) native speaker of English
> to decide that _boy_ (the etymology of which is disputed) is a
> loanword; however, it seems plausible that that same speaker could
> uncover the fact that all words beginning with /D/ are "closed-class"
> or the like.

This is also irrelevant.  The fact that initial [D] occurs only in
function or grammatical (or closed-class) words is not what keeps initial
[T] and [D] from contrasting in English.  What keeps them from contrasting
is the fact that initial [D] in English is always a morpheme.

> Whether all this is relevant to the phonemicity of /D/ vs. /T/ is a
> completely different argument, however.

Thank you.  At least on person understands the issue.

> If we take that same "intelligent (but linguistically untrained)
> native speaker of English" as our judge again, it's clear that a
> minimal pair like "either" vs. "ether" would establish the phonemic
> distinction.

Only in some dialects.  But it is agreed that 'either' and 'ether' are
harder to dispose of.  But this is also irrelevant to whether initial
[T] and [D] contrast in English.

>> Take another case.  It is extremely difficult to find minimal pairs for
>> [esh] and [ezh] in *any* position -- and all the pairs I can think of
>> involve involve either proper names of foreign origin or obscure Scrabble
>> words.

> The best pair I could come up with is _Confucian_ (with [esh]) vs.
> _confusion_ (with [ezh]).

My favorite is 'assure', 'azure', and 'adjure'.  This only works for one
pronunciation of 'azure' however, rather like 'either' and 'ether'.  The
deciding thing for the phonemicity of /Z/ is that it appears to be an
allophone of both /z/ and /j/ (j with hachek) and there is a classical
phonological rule that says that allophones can only belong to one
phoneme.  Since an allophone can't belong to two different phonemes
without violating the bi-uniqueness rules, /Z/ must be a phoneme, at
least by classical phonological rules.

Bob Whiting
whiting at cc.helsinki.fi