Re: Phonological equivalent of "The quick brown fox..."

From: R A Brown
Sent: Tuesday, February 06, 2007 11:58 AM
Daniel Prohaska wrote:

> 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"
 

“If all these had been written between square brackets, I would have no
quarrel with it.

Ray, 

That just depends on your analysis of the phonemes I question, on whether
you analyse them in context with other varieties of English, or on a
historical bases. In a descriptive analysis of my lect the phonemic brackets
above are absolutely correct.

 

“But it seems to me to run counter to the phonemic theory to set up a
phonemic inventory for the English of north-west England and thus separate
ones for…”

Nothing in “phonemic theory” says that you cannot analyse regional variants
separately, or local dialects, or even idiolects – nothing…

 

“…, presumably, north-east England, the west Midlands, the east Midlands,
south west England and south East England. that is, at least _six_ different
phonemic inventories for varieties of English in just _England_ - we'd have
to add others for the English of south-east Wales, of the Welsh valleys, of
south Wales etc, and I don't know how many for Scotland and Ireland.”

Yes, and that just depends on what variety you’re describing. I think you’re
a little hung up on English RP school book phonemic analysis…, but even for
RP /:/ can be justified. In my lect it is most certainly a distinguishing
feature. 

 

“Then we'll have more still for the US (How many different one there?), for
Canada, New Zealand, Australia, the Indian sub-continent, the various
different anglophone countries of Africa and the Caribbean.”

As many as there are Englishes…

 

No - that is simply *not* what I have understood the phonemic theory to be
about. As Mark wrote:

Well, in that case maybe you ought to look into “phonemic theory” again and
ask yourself how you would describe a local variety of English – without
reference to Standard English forms or historical derivation – purely
descriptive. What phonemic analysis would you arrive at?

-------------------------------

Mark J. Reed wrote:

[snip]

 > Sure.  But the choice of phonemic symbols is still arbitrary; so even

 > though

 > the distinction is length rather than rhoticitiy, you could still just as

 > easily say that "part" is /part/ and the phonetic realization of /ar/ is

 > [a:] in that context (vs contexts where it really is [ar] or [a:r] due to

 > epenthesis/liaison).

 

Of course my choice of phonemic symbols is arbitrary. I can call “long-a”
/ä/ if I want to, or even /#/, but what good would it do me? In order to
analyse a phonological system I need to determine what features in this
system allows for contrast. Now I have given various minimal pairs in my
lect to show that they’re distinctive and in contrast by vocalic length
alone, hence vocalic quantity is a systematically contrastive feature and
should be so marked in a sensible phonemic adaptation of my lect. Of course
I could represent the phonemes of my lect with the traditional RP derived
symbols using /p&t/ vs. /pAt/, but that wouldn’t do the phonology of my lect
justice. First of all it would suggest that I have a short front raised
vowel for the former and a back, slightly raised long one in contrast, when
in fact the quality of my two vowels is low and front, distinguished ONLY by
length. 

 

“Yes, indeed - in fact IME many non-linguistically minded southerners think
they do pronounce the _r_ when they say [p_hA:t]! They will explain that
without the _r_ it's [p_h&t], but that [&] + [r\] makes [A:] - naive, I
know, but that has been my experience talking with non-linguists   :)”

Phonemically transcribing voiceless stop aspiration is as redundant in
north-western E as it is in southern E or RP, so as you see I was well able
of abstraction in my phonemic transcription.  

 

 >Such a representation is biased toward rhotic

 > dialects, -but no less valid for that bias.

 

“Indeed it is not, and I would argue that on pan-Anglophone considerations
/part/ is better _phonemic_ representation. To have to posit different
phonemic representations such as */pA:t/, */pA`t/, */pa`t/, */part/ etc is
certainly not using Ockham's razor - IMO it's letting the beard grow wild &
unkempt.

 

 > You chose to render the vowel in "bays" as /e:/, but if there's no short

 > /e/

 > to contrast it with, the /:/ is optional.

 

“In fact phonemically it's redundant.”

Not in a systematic context, as I’ve tried to explain (perhaps not very
well). Since /E/ and /E:/ are contrastive and /E/ with /e:/ by quantity and
quality, whereas /E:/and /e:/ only by quality, we can see that /E:/ and /e:/
share a systematic feature – vocalic length – hence /:/ in /e:/, although
there’s no */e/. We could just as well assign these three vocalic phonemes
the phonemic designations /x/, /y/ and /z/, but that wouldn’t tell us much
about the linguistic system now would it?

No, /:/ makes sense in my variety, and also (in my opinion) on a systematic
level in RP, especially when analysed in contrast with GA (which is
systematically isochronic). 

 

 > One could also choose to include the offglide (/eI/, /ej/, et sim) or
not.

 

“And in Britain we also have varieties like [e@] and [I@] for the same
phoneme   ;)”

Us up North, we ‘ave continental vowels, like, which meks them southerners
laff at us ‘caus of our [e:]s and [o:]s.

 

 > As long as everyone agrees that

 > we're talking about the vocalic phoneme in the word "bays", the
particular

 > symbol choice is unimportant.

 

“Applying Ockham's razor, /e/ would seem the best.”

That would make it appear as systematically short as /E/, which it is not.
Wrong analysis I’m afraid.

 

 > /Q/ "cot"     ~        /Q:/ "caught"

 >

 > That's an interesting one.  (Assuming those are really [Q] and [Q:] in
your

 > 'lect, anyway. :))  It would not have occurred to me that the difference

 > between those two would be realized as length.

 

“Nor I, but I have no reason to suppose that Daniel is not giving true
phonetic rendering  of his 'lect.”

No, you have no reason to suppose that, why would I lie to you? My <cot> and
<caught> vowels are exactly the same in quality, distinguished only by the
length of the vowel in the latter. <caught> on the other hand is
distinguished by quality ALONE from <coat> which I’ve given the phonemic
symbol /o:/, though we have varying phonetic realisations in the North, some
with a more open [O:] (tends to be in the west), others with the more close
[o:] (often more easterly dialects). 

An interesting distinction appears when switching from northern regional
standard (NRS = RP-system in northern pronunciation) to local dialect, i.e.
when “when talking broad”, that NRS has /”tQ:kIN brQ:d/, but in dialect it’s
/”tQ:kIn bro:d/ where <broad> has the <boat>-vowel. 

 

“Down here in the south-east they are [Q] and [O:]”

One of the many ways we’re different… ;-)

 

 >Does that count as a

 > "partial CAUGHT-COT merger", since the qualities are the same?

 

“Presumably it does.”

It’s a qualitative merger at least, and it would serve as an example to
understand how the merger came about in colonial varieties of English when
the lost vowel quantity distinctions. 

 

“But I'm not sure if this particularly aspect of the thread is going to get
us anywhere. It seems to me that Daniel and I do not have the same view of
what a phoneme actually is (it's an abstract construct anyway)."

I rather think we both know what a phoneme is, but that we differ in our
analysis of English phonology – or maybe you’re just unfamiliar with
northern speech patters apart from “oh aye”-clichés.

Dan

 

 

Replies