For me, as a speaker of north-western English English, phonemic quantity
distinctions are very much part of the system. I'm very much aware of them.
Here are a few contrasts I spontaneously came up with. 


/a/ "pat"      ~        /a:/ "part"

/E/ "bed"     ~        /E:/ "bared", "bares" ~     /e:/ "bays"

/I/ "bid"       ~        /I:/ "beard"           ~        /i:/ "bead" 

/Q/ "cot"     ~        /Q:/ "caught"        ~        /o:/ "coat"



From: Of R A Brown
Sent: Sunday, February 04, 2007 6:57 PM
Daniel Prohaska wrote:

> No, there's no contrast with /A:/ because */A/ doesn't exist. 


If there's no contrast then [:] cannot surely be _phonemic_.


> /Q/ does, though. 

I am aware of that - I've been speaking southern British English for more
than 60 years! But I fail to see how that is relevant to whether we have /A/
~ /A:/ or not.


> But vocalic length itself is distinguished in EE, e.g. between /E/

> and /E:/ in /bEd/ <bed> vs. /bE:d/ <bared> 

I think not - there's certainly a _qualitative_ difference in the way I say
it (and I speak a normal non-rhotic SE England variety). In any case, to
give [:] phonemic status because of this pair only seems weak to me.


> (though this could also be analysed as /bE@d/). 

It could, and sometimes is. I have also seen it analyzed as [log in to unmask] But I am
not over fond of such analyses either as it means that we posit one set of
phonemes for non-rhotic varieties and a different set of phonemes for
English speakers in south-west England and among the rural dialects of the
Midlands and several parts of southern England with r-colored/ rhotic vowels
(i.e. /bE`d/ or /be`d/. It also means that we have to have yet a *third* set
of phonemes for those speakers in much of Wales and in the Scottish
Highlands who actually have a apically trilled consonant in _bared_, i.e.


As all these dialect variants are _predictable_, it does not seem to me
sensible to be setting three different _phonemic_ realizations of the |r| in
_bared_ - indeed, unless I have completely misunderstood the phonemic
theory, I would have thought it was wrong to do so.


> Vocalic length is systematic, anyway, so it makes sense to transcribe EE
with /:/. 

I don't understand what you mean by the first clause, and - as you see above
- I do not agree with the second clause.


> /A:/ though it doesn't contrast with */A/ falls

> into the "long vowel" category. 

It certainly does.