broad-phonetic vs narrow-phonetic transcription
Bob Cunningham
If I write that I pronounce "cat" [k&t], how does anyone know whether
I intend a broad- or a narrow-phonetic transcription?
If [k&t] precisely fits the way I say "cat", then no diacritics or
other modifications are applicable and the transcription is as narrow
as it can get.
If I write '"cat" is pronounced [k&t], but I'm using a broad-phonetic
transcription', then it follows that I would add something to show
more exactly how I pronounce "cat".
That's why I think it's desirable to have separate enclosures for
broad- and narrow-phonetic transcription.
If I write that I use vertical bars for broad-phonetic transcription,
and that I pronounce 'cat' |k&t|, then the reader knows that that's
not exactly the way I pronounce it, and diacritics or other
modifications would be necessary to transcribe my pronunciation more
narrowly.
I would appreciate comments from anyone except Peter T Daniels.
Richard Fontana
> If I write that I pronounce "cat" [k&t], how does anyone know whether
> I intend a broad- or a narrow-phonetic transcription?
There's a sense in which it's obviously somewhat broad, isn't there, in
that you haven't indicated aspiration of the [k].
Peter T. Daniels
>
> If I write that I pronounce "cat" [k&t], how does anyone know whether
> I intend a broad- or a narrow-phonetic transcription?
>
> If [k&t] precisely fits the way I say "cat", then no diacritics or
> other modifications are applicable and the transcription is as narrow
> as it can get.
Let's make sure this doesn't involve your trap from last time. Are you
talking about the English word "cat," pronounced by a native speaker of
English?
In that case, it's obvious that the transcription is quite coarse --
slovenly, even -- because it omits the aspiration from the first
consonant, the nature of the release of the second one, and any relevant
suprasegmentals.
> If I write '"cat" is pronounced [k&t], but I'm using a broad-phonetic
> transcription', then it follows that I would add something to show
> more exactly how I pronounce "cat".
>
> That's why I think it's desirable to have separate enclosures for
> broad- and narrow-phonetic transcription.
>
> If I write that I use vertical bars for broad-phonetic transcription,
> and that I pronounce 'cat' |k&t|, then the reader knows that that's
> not exactly the way I pronounce it, and diacritics or other
> modifications would be necessary to transcribe my pronunciation more
> narrowly.
People don't go around writing phonetic transcriptions out of the blue.
Transcriptions are used for a purpose, and in any particular instance,
that purpose either is known to the audience already, or is stated in
the communication where the transcriptions appear. The readers can then
determine for themselves whether adequate detail for the purpose has
been provided.
Using a phonemic transcription, like the above, however, is
qualitatively quite different: it implies a particular theory of
phonology and a particular analysis within that theory.
> I would appreciate comments from anyone except Peter T Daniels.
>
It's so nice to be appreciated.
--
Peter T. Daniels gram...@worldnet.att.net
Neil Coffey
> If I write that I pronounce "cat" [k&t], how does anyone know whether
> I intend a broad- or a narrow-phonetic transcription?
The terms "broad" and "narrow" are a bit like "little" and
"big". Size forms a continuum, and "little" and "big" denote
a conception of being nearer to one end or other of this
continuum. Similarly, the amount of detail you add to
a transcription lies somewhere on a continuum. A "broad"
transciption is one where there's 'not much' detail; a
"narrow" one is where there's 'quite a lot' of detail. There
aren't really two distinct types of transcription, or
two distinct levels of detail, in the way you suggest.
> If [k&t] precisely fits the way I say "cat", then no diacritics or
> other modifications are applicable and the transcription is as narrow
> as it can get.
I think you're misunderstanding what transcription is. A
transcription will never ever ever ever ever tell you
"precisely" how a word is pronounced. All a transcription
ever does is give an indication of some features
of the utterance which the transcriber thinks are in some way
relevant. There will always be information discarded, no
matter how detailed the transcription is. And given a
particular transcription, there will generally always be
more detail that you could potentially add, if you so wished.
To address your actual question a little more precisely:
When you write, say, a particular vowel symbol, the reader
will generally understand that you mean 'a vowel in the
region of the cardinal vowel denoted by that symbol'.
If you want to tell the reader that the vowel is *very*
close to the corresponding vowel, then you simply tell
them this fact in the accompanying text. And you can
annotate your transcription with footnotes just as you
would any ordinary text. Or you can make up your own
symbol/diacritic where necessary to signal particular
pieces of information when there isn't an appropriate
symbol already defined.
So, to return to your original question:
> how does anyone know whether
> I intend a broad- or a narrow-phonetic transcription?
the answer is that you tell them. When you transcribe a passage,
or put transcriptions in a text, you give the reader a list of
the transcription conventions you have used. For examples,
see the illustrations in the Handbook of the IPA.
Neil
Bob Cunningham
Okay, that sounds good. Thank you.
Bob Cunningham
said:
>Bob Cunningham wrote:
>> If I write that I pronounce "cat" [k&t], how does anyone know whether
>> I intend a broad- or a narrow-phonetic transcription?
>The terms "broad" and "narrow" are a bit like "little" and
>"big". Size forms a continuum, and "little" and "big" denote
>a conception of being nearer to one end or other of this
>continuum. Similarly, the amount of detail you add to
>a transcription lies somewhere on a continuum. A "broad"
>transciption is one where there's 'not much' detail; a
>"narrow" one is where there's 'quite a lot' of detail. There
>aren't really two distinct types of transcription, or
>two distinct levels of detail, in the way you suggest.
I didn't mean to imply that broad- and narrow-phonetic transcription
were separable categories. If I speak of one person as a tall man and
another person as a short man, I'm not implying that I don't know that
the distribution of heights of men is essentially a continuum, but I
still expect to be understood. "A tall man" refers to a man who is
tall enough to motivate comment about his height, and the same holds
for "a short man". Despite the reality of the continuum, the phrases
will continue to be heard, be useful, and be understood.
But all that I've said about broad-phonetic transcription has really
been a reaction to the idea some people seem to have that there's no
difference between phonemic transcription and broad-phonetic
transcription. Everything I've read about phonemes -- except for
references to linguists who hold that the concept of phoneme is
useless or that there really is no such thing as a phoneme or who hold
that phonemic analysis must be done with no regard to semantics --
emphasizes the use of phonemes as a means of identifying pairs of
sounds in a language that serve to distinguish meaning. When I see
people using phonemic notation for no other purpose than to
distinguish pronunciations, my mind rebels. If the purpose of a
transcription has nothing to do with the phonological attributes of a
language, and everything to do with simply how words are pronounced,
phonemic notation seems grossly inappropriate.
If I were to try to use phonemic transcription to describe the
difference between British and American pronunciations of "stove", I
might find myself writing the useless "The British pronounce 'stove'
/stov/ and the Americans pronounce it /stov/," based upon the
assumption that the British [st@Uv] and my [stouv] are not
phonemically distinct. I could even write, using phonemic
transcription "The British pronounce 'stove' /stouv/ and the Americans
pronounce it /st@Uv/," assuming that [@U], [ou], and [o] are all free
variants of the same phoneme, so it doesn't matter which I use in a
phonemic transcription.
To convey a thought that someone might conceivably find interesting, I
could write "The British pronounce 'stove' [st@Uv] and the Americans
pronounce 'stove' [stouv]."
My underlying motive in discussing this subject is to oppose the use
of slashes when the intent is only to show how words are pronounced,
with no thought of phonemic analysis. I would expect to find that if
an examination were made of a large number of AUE postings in which
slashes were used to transcribe pronunciation, it would be obvious in
almost every case that the writer had no desire or need to imply
phonemic distinction.
The additional complication that some people seemed to think that
slashes were appropriate for phonetic transcription if the
transcription was broad led me to suggest a separate notation for
broad-phonetic transcription. If no one had been using slashes when
there was no reason to assume that their transcription was phonemic, I
might never have made the suggestion.
>> If [k&t] precisely fits the way I say "cat", then no diacritics or
>> other modifications are applicable and the transcription is as narrow
>> as it can get.
>I think you're misunderstanding what transcription is. A
>transcription will never ever ever ever ever tell you
>"precisely" how a word is pronounced. All a transcription
>ever does is give an indication of some features
>of the utterance which the transcriber thinks are in some way
>relevant. There will always be information discarded, no
>matter how detailed the transcription is. And given a
>particular transcription, there will generally always be
>more detail that you could potentially add, if you so wished.
I understand all of that, but my only involvement in all of this comes
from a desire to see ASCII IPA used more often and more properly in
discussions of pronunciation, and in enough detail to make the
discussions significant. It's useful to be able to say that a British
speaker may say [d@iv] for "dive" while an American speaker may say
[dAIv], but it's pointless to use the two transcriptions in phonemic
transcription unless a phonemic contrast can be demonstrated between
[AI] and [@i]. And, again, it's not appropriate for the writer to use
phonemic notation unless he's really thinking in terms of phonemes,
even if the different forms he's transcribing are phonemically
distinct.
As you imply, "broad" and "narrow" are relative terms. It can be
said, though, that one transcription is "broader" or "narrower" than
another, and if transcription A is broader than transcription B, it's
idiomatic to say that A is the broad one and B is the narrow one.
If I write
A. The British pronunciation of 'stove' is often [st@Uv] and
the American, [stoUv].
B. The British and American pronunciations of 'stove' are both
[stov].
I think I can call 'A' a narrow transcription and 'B' broad,
especially since I've strongly implied that I'm discussing English
pronunciation, so contextually determined pronunciations need not be
specified in a narrow transcription.
>To address your actual question a little more precisely:
>When you write, say, a particular vowel symbol, the reader
>will generally understand that you mean 'a vowel in the
>region of the cardinal vowel denoted by that symbol'.
It's good to know that you think the symbols can be thought of as
denoting cardinal vowels, and that we can think of them as
representing the regions of the vowel quadrilateral corresponding to
cardinal vowels. I've had mixed reception to my own statements of
that useful way of thinking.
>If you want to tell the reader that the vowel is *very*
>close to the corresponding vowel, then you simply tell
>them this fact in the accompanying text. And you can
>annotate your transcription with footnotes just as you
>would any ordinary text. Or you can make up your own
>symbol/diacritic where necessary to signal particular
>pieces of information when there isn't an appropriate
>symbol already defined.
>So, to return to your original question:
>> how does anyone know whether
>> I intend a broad- or a narrow-phonetic transcription?
>the answer is that you tell them. When you transcribe a passage,
>or put transcriptions in a text, you give the reader a list of
>the transcription conventions you have used. For examples,
>see the illustrations in the Handbook of the IPA.
I see that Ladefoged's illustration of American English is a good
example of that as applied to my language.
Thank you for your helpful comments.
Peter T. Daniels
>
> On Tue, 29 May 2001 22:59:56 +0000, Neil Coffey <ne...@ox.compsoc.net>
> said:
>
> >Bob Cunningham wrote:
>
> >> If I write that I pronounce "cat" [k&t], how does anyone know whether
> >> I intend a broad- or a narrow-phonetic transcription?
>
> >The terms "broad" and "narrow" are a bit like "little" and
> >"big". Size forms a continuum, and "little" and "big" denote
> >a conception of being nearer to one end or other of this
> >continuum. Similarly, the amount of detail you add to
> >a transcription lies somewhere on a continuum. A "broad"
> >transciption is one where there's 'not much' detail; a
> >"narrow" one is where there's 'quite a lot' of detail. There
> >aren't really two distinct types of transcription, or
> >two distinct levels of detail, in the way you suggest.
>
> I didn't mean to imply that broad- and narrow-phonetic transcription
> were separable categories.
But that's exactly what you do when you write |---| for one of them and
[---] for the other.
> If I speak of one person as a tall man and
> another person as a short man, I'm not implying that I don't know that
> the distribution of heights of men is essentially a continuum, but I
> still expect to be understood. "A tall man" refers to a man who is
> tall enough to motivate comment about his height, and the same holds
> for "a short man". Despite the reality of the continuum, the phrases
> will continue to be heard, be useful, and be understood.
>
> But all that I've said about broad-phonetic transcription has really
> been a reaction to the idea some people seem to have that there's no
> difference between phonemic transcription and broad-phonetic
> transcription.
Those people are simply wrong.
> Everything I've read about phonemes -- except for
> references to linguists who hold that the concept of phoneme is
> useless or that there really is no such thing as a phoneme or who hold
> that phonemic analysis must be done with no regard to semantics --
NB those are two quite different groups of linguists
> emphasizes the use of phonemes as a means of identifying pairs of
> sounds in a language that serve to distinguish meaning. When I see
That's the misunderstanding, or the incoherence. Phonemes don't
"identify pairs." Phonemes are originally identified by examining pairs
of words for "same" or "different." The pairs don't distinguish the
meanings, the phonemes do.
> people using phonemic notation for no other purpose than to
> distinguish pronunciations, my mind rebels. If the purpose of a
> transcription has nothing to do with the phonological attributes of a
> language, and everything to do with simply how words are pronounced,
> phonemic notation seems grossly inappropriate.
The rebellion is quite proper -- but "how words are pronounced"
certainly does relate to "the phonological attributes of the language."
> If I were to try to use phonemic transcription to describe the
> difference between British and American pronunciations of "stove", I
> might find myself writing the useless "The British pronounce 'stove'
> /stov/ and the Americans pronounce it /stov/," based upon the
> assumption that the British [st@Uv] and my [stouv] are not
> phonemically distinct.
Of course trying to do that is impossible -- what you might say is that
the British and American realizations of /stowv/ are [st@Uv] and [stouv]
respectively.
> I could even write, using phonemic
> transcription "The British pronounce 'stove' /stouv/ and the Americans
> pronounce it /st@Uv/," assuming that [@U], [ou], and [o] are all free
> variants of the same phoneme, so it doesn't matter which I use in a
> phonemic transcription.
No, that you can't; there's no such thing as "free variants of a
phoneme."
See Neil's reply to a different message for what "free variation" means.
> To convey a thought that someone might conceivably find interesting, I
> could write "The British pronounce 'stove' [st@Uv] and the Americans
> pronounce 'stove' [stouv]."
Exactly.
> My underlying motive in discussing this subject is to oppose the use
> of slashes when the intent is only to show how words are pronounced,
> with no thought of phonemic analysis. I would expect to find that if
> an examination were made of a large number of AUE postings in which
> slashes were used to transcribe pronunciation, it would be obvious in
> almost every case that the writer had no desire or need to imply
> phonemic distinction.
>
> The additional complication that some people seemed to think that
> slashes were appropriate for phonetic transcription if the
> transcription was broad led me to suggest a separate notation for
> broad-phonetic transcription. If no one had been using slashes when
> there was no reason to assume that their transcription was phonemic, I
> might never have made the suggestion.
Why replace a mistake with an incoherence? (That's not an insult, it
literally means that trying to make such a distinction lacks coherence
-- does not stand up to scrutiny.)
Jacques Guy
> If I write
>
> the American, [stoUv].
>
> B. The British and American pronunciations of 'stove' are both
> [stov].
>
> I think I can call 'A' a narrow transcription and 'B' broad
I'd think that 'B' is just plain wrong. Even /stov/ I'd say is
dead wrong. Mind you, I've recently seen this sort of stuff in print...,
wait... "Phonology" by Abigail Cohn in "The Handbook of
Linguistics". She writes [bot] for "boat", [bet] for "bait",
(she claims to be describing American English, _not_ Scots)
and does not realize that they are diphthongs, since she comments
that, how weird, Japanese speakers perceive them as long (and
her Japanese transcriptions are similarly up shit creek too).
I was starting to believe I'd stepped into a parallel universe,
where America had been colonized by Scots speakers,
but I found the solution in the preceding chapter, by a John
Laver. The fellow analyzes vowel quantity in English as
conditioned by vowel quality, so for instance "bit" = /bIt/ = [bIt]
and "beet" = /bit/ = [bi:t]. Well, he doesn't _analyze_ it, he does
not even _declare_ it to be so, he just plonks it without
further ado. If Lewis Carroll had not done it better and entertainingly
long ago, this could be called postmodernist phonology, where
nothing means anything and everything means anything (see
"English Semiotics" by Humpty and Dumpty).
Richard Fontana
> Bob Cunningham wrote:
>
> > If I write
> >
> > A. The British pronunciation of 'stove' is often [st@Uv] and
> > the American, [stoUv].
> >
> > B. The British and American pronunciations of 'stove' are both
> > [stov].
> >
> > I think I can call 'A' a narrow transcription and 'B' broad
>
> I'd think that 'B' is just plain wrong. Even /stov/ I'd say is
> dead wrong.
I don't see why /stov/ is necessarily dead wrong, if you adopt the
convention that the vowel of "stove" will be represented with /o/.
> Mind you, I've recently seen this sort of stuff in print...,
> wait... "Phonology" by Abigail Cohn in "The Handbook of
> Linguistics". She writes [bot] for "boat", [bet] for "bait",
> (she claims to be describing American English, _not_ Scots)
> and does not realize that they are diphthongs, since she comments
> that, how weird, Japanese speakers perceive them as long (and
> her Japanese transcriptions are similarly up shit creek too).
> I was starting to believe I'd stepped into a parallel universe,
> where America had been colonized by Scots speakers,
Well, it was, to a significant degree. (I think there actually are some
American dialects that have /e/ as in "bait" not diphthongized.)
Peter T. Daniels
"A John Laver" just happens to be (or to have been) the President of the
International Phonetic Association (his *Principles of Phonetics* is one
of the fattest books in the Cambridge Red Series). (You give evidence
that the British distinction between "phonetician" and "linguist" is a
principled one.)
Abby Cohn is the daughter of a friend of mine; her specialty is
phonological theory, so presumably she has reasons for phonologizing
English as she does.
/stov/ is indeed an odd phonemicization of "stove" -- IF you use a
Smith-Trager-type system (where it would be /stowv/). But in fact you
can't say it's entirely wrong without seeing the whole phonemic system
in which it finds its place.
Bob Cunningham
<rf...@sparky.cs.nyu.edu> said:
[...]
>(I think there actually are some
>American dialects that have /e/ as in "bait" not diphthongized.)
In Peter Ladefoged's Illustration of American English in _The Handbook
of the International Phonetic Association_, he uses [e] in his 'narrow
transcription' for the 'a' vowel in words like 'came', 'making', and
'take', while using diphthongs like [aI] ('shined') and [aU]
('around') elsewhere.
Jacques Guy
> I don't see why /stov/ is necessarily dead wrong, if you adopt the
> convention that the vowel of "stove" will be represented with /o/.
The vowel of "stove" is a _diphthong_ in British English and in
American English. I have no idea what it is in Scots as
spoken in Edinburgh, but seeing that there most diphthongs
are pure, long vowels, I guess that it might perhaps be /sto:v/.
At any rate, you can only write /stov/ and claim it is American
English because American "shop" is English "sharp". How the
bloody hell would you represent the three-way contrast
cot/court/coat if you transcribed "coat" as /kot/ like you
want "stove" as /stov/?
Let's adopt the convention that the whatever of "stove",
between the "t" and the "v" will be represented as
Donald Duck's hat on top of Minnie Mouse's knickers,
hey? It will be just as true to fact, and a lot more
fun. But it don't matter none, as long as you remember
what Donald Duck's hat on top of Minnie Mouse's
knickers symbolizes (hint: that is NOT Donald Duck's
hat on top of Minnie Mouse's knickers).
Bob Cunningham
<gram...@att.net> said:
>> I could even write, using phonemic
>> transcription "The British pronounce 'stove' /stouv/ and the Americans
>> pronounce it /st@Uv/," assuming that [@U], [ou], and [o] are all free
>> variants of the same phoneme, so it doesn't matter which I use in a
>> phonemic transcription.
>No, that you can't; there's no such thing as "free variants of a
>phoneme."
I believe that that assertion is in conflict with what others say.
It seems to me that saying "free variants of a phoneme" is another way
of saying "various realizations of a phoneme that are not in
contrastive distribution". Do you disapprove of the latter way of
saying it?
>See Neil's reply to a different message for what "free variation" means.
I had read it and concluded that he was merely disagreeing with the
_Britannica_ article's definition of "free variant". I respect Neil's
opinion, but I can't assume that the _Britannica_ is wrong. I must
conclude that this is yet another term for which linguists can't agree
upon a definition.
Niel said:
"Free variation" is usually used to refer to
interchangeability of segments _within_ a particular accent.
Either that one realisation of a phoneme can be subsituted
for another without any particularly well-defined contextual
motivation. Or -- "phonemic free variation" -- that in
particular words, one phoneme can be substituted for
another-- cases like 'scone' = /skOn, sk@Un/; 'schedule' =
/skEdjul, SEdjul/ in British English.
The _Britannica_ article says:
Phones that can occur and do not contrast in the same context
are said to be in free variation in that context, and, as has
been shown, there is a permissible range of variation for the
phonetic realization of all phonemes. More important than free
variation in the same context, however, is systematically
determined variation according to the context in which a given
phoneme occurs.
That seems to be implying clearly enough that free variants of a
phoneme are realizations of that phoneme that are not in contrastive
distribution.
David Crystal's _Cambridge Encyclopedia of Language_ (page 161) seems
to agree with the _Britannica_ interpretation of "free variation". He
says:
If the sounds do occur in the same place in a word, then they
can belong to the same phoneme only if they do not change the
meaning of the word. [Gives examples] Such sounds are said
to be in "free variation".
(He also covers the point that for sounds to belong to the same
phoneme, they "ought to display a reasonable amount of physical
similarity".)
Richard Fontana
> Richard Fontana wrote:
>
> > I don't see why /stov/ is necessarily dead wrong, if you adopt the
> > convention that the vowel of "stove" will be represented with /o/.
>
> The vowel of "stove" is a _diphthong_ in British English and in
> American English. I have no idea what it is in Scots as
> spoken in Edinburgh, but seeing that there most diphthongs
> are pure, long vowels, I guess that it might perhaps be /sto:v/.
>
> At any rate, you can only write /stov/ and claim it is American
> English because American "shop" is English "sharp". How the
> bloody hell would you represent the three-way contrast
> cot/court/coat if you transcribed "coat" as /kot/ like you
> want "stove" as /stov/?
As long as you have three distinct symbols you can do this.
/kA.t/ /kOt/ /kot/ .
> Let's adopt the convention that the whatever of "stove",
> between the "t" and the "v" will be represented as
> Donald Duck's hat on top of Minnie Mouse's knickers,
> hey? It will be just as true to fact, and a lot more
> fun. But it don't matter none, as long as you remember
> what Donald Duck's hat on top of Minnie Mouse's
> knickers symbolizes (hint: that is NOT Donald Duck's
> hat on top of Minnie Mouse's knickers).
I don't really understand the strength of the objection. We're talking
about phonemic transcription, not broad phonetic transcription. I do not
claim to be knowledgeable in these matters, however.
Peter T. Daniels
That doesn't seem "narrow" at all.
Peter T. Daniels
>
> On Wed, 30 May 2001 21:15:38 GMT, "Peter T. Daniels"
> <gram...@att.net> said:
>
> >> I could even write, using phonemic
> >> transcription "The British pronounce 'stove' /stouv/ and the Americans
> >> pronounce it /st@Uv/," assuming that [@U], [ou], and [o] are all free
> >> variants of the same phoneme, so it doesn't matter which I use in a
> >> phonemic transcription.
>
> >No, that you can't; there's no such thing as "free variants of a
> >phoneme."
>
> I believe that that assertion is in conflict with what others say.
>
> It seems to me that saying "free variants of a phoneme" is another way
> of saying "various realizations of a phoneme that are not in
> contrastive distribution". Do you disapprove of the latter way of
> saying it?
That's (subphonemic) free variation; it's not free variation of a
phoneme. Examples might be "checked" vs. "released" versions of final
stops in English.
> >See Neil's reply to a different message for what "free variation" means.
>
> I had read it and concluded that he was merely disagreeing with the
> _Britannica_ article's definition of "free variant". I respect Neil's
> opinion, but I can't assume that the _Britannica_ is wrong. I must
> conclude that this is yet another term for which linguists can't agree
> upon a definition.
>
> Niel said:
>
> "Free variation" is usually used to refer to
> interchangeability of segments _within_ a particular accent.
> Either that one realisation of a phoneme can be subsituted
> for another without any particularly well-defined contextual
> motivation. Or -- "phonemic free variation" -- that in
> particular words, one phoneme can be substituted for
> another-- cases like 'scone' = /skOn, sk@Un/; 'schedule' =
> /skEdjul, SEdjul/ in British English.
For the former, cf. eckonomic/eekonomic.
> The _Britannica_ article says:
>
> Phones that can occur and do not contrast in the same context
> are said to be in free variation in that context, and, as has
> been shown, there is a permissible range of variation for the
> phonetic realization of all phonemes. More important than free
> variation in the same context, however, is systematically
> determined variation according to the context in which a given
> phoneme occurs.
>
> That seems to be implying clearly enough that free variants of a
> phoneme are realizations of that phoneme that are not in contrastive
> distribution.
I don't see that it says anything of the sort! For instance, my example
of checked vs. released is unconditioned finally, but the phonetic
realization of that difference -- in the form of aspiration -- is, as
you know, strictly complementarily distributed elsewhere.
> David Crystal's _Cambridge Encyclopedia of Language_ (page 161) seems
> to agree with the _Britannica_ interpretation of "free variation". He
> says:
>
> If the sounds do occur in the same place in a word, then they
> can belong to the same phoneme only if they do not change the
> meaning of the word. [Gives examples] Such sounds are said
> to be in "free variation".
>
> (He also covers the point that for sounds to belong to the same
> phoneme, they "ought to display a reasonable amount of physical
> similarity".)
Without the examples, the description isn't too helpful ...
Brian M. Scott
The transcription is narrower than it looks, though I don't find it
particularly narrow. Bob neglected to mention the list of conventions
associated with the transcription, one of which is that '[e] and [o]
are usually slightly diphthongized'. One of P.L.'s points is that
there are many correct ways to do the transcription depending on what
conventions one adopts.
Brian
Neil Coffey
> >No, that you can't; there's no such thing as "free variants of a
> >phoneme."
>
> I believe that that assertion is in conflict with what others say.
I think I agree with Peter on this -- the specific wording
"free variants of a phoneme" isn't one I can honestly say I've
heard much before now. It's not entirely clear what it means,
either, I don't think.
> It seems to me that saying "free variants of a phoneme" is another way
> of saying "various realizations of a phoneme that are not in
> contrastive distribution".
Do you have any specific objection to simply saying
"allophones in free variation"? There's not necessarily any reason
to adopt the stance that things can't be called 'allophones' if
they're not in complementary distribution. For comparison, Goldsmith
says the following:
"The simplest situation we might find is that 'x' and 'y'
are allophones in complementary distribution or in free
variation. In the former case, we find there is no phonetic
environment in which both 'x' and 'y' appear, while in the
latter 'x' and 'y' may occur freely with no lexical or
grammatical difference involved."
(Handbook of Phonological Theory, p 10)
It seems fairly clear from this that Goldsmith views phones
in free variation to be just as much allophones as those in
complementary distribution.
> The _Britannica_ article says:
>
> Phones that can occur and do not contrast in the same context
> are said to be in free variation in that context
I think that more or less amounts to the definition I gave,
doesn't it? -- it's just that the Brittanica is being a bit
vague by using the word 'phone', whereas it's arguably
worth separating out cases where the items in free
variation are specifically phonemes.
Incidentally, as far as I can recall, none of the passages
you've quoted from the Britannica say that 'segments
in free variation cannot be called allophones', I don't
think.
> David Crystal's _Cambridge Encyclopedia of Language_ (page 161) seems
> to agree with the _Britannica_ interpretation of "free variation". He
> says:
>
> If the sounds do occur in the same place in a word, then they
> can belong to the same phoneme only if they do not change the
> meaning of the word. [Gives examples] Such sounds are said
> to be in "free variation".
This is more or less true, but in order to be correct I think
it relies on you interpreting "in the same place in a word" as
"in a particular phonological context". I wonder if the *word*
has specifically to be the domain, too.
> (He also covers the point that for sounds to belong to the same
> phoneme, they "ought to display a reasonable amount of physical
> similarity".)
Though with the usual problem of deciding what 'physical' actually
means (does it mean 'articulatory' or 'auditory' or 'acoustic'
or maybe a mixture of all three, in which case what mixture
exactly?).
Neil
Bob Cunningham
If you were familiar with the _Handbook of the International Phonetic
Association_, you would know that each language illustration has a
section labeled "Conventions". In the narrow transcriptions, this
makes it unnecessary to show an elaboration of extra marks that are
common to all occurrences of a symbol.
My reason for mentioning the two diphthongs was to show that the use
of [e] couldn't be explained by his eschewing diphthongs entirely in
his transcription.
Steve Hayes
>Richard Fontana wrote:
>
>> I don't see why /stov/ is necessarily dead wrong, if you adopt the
>> convention that the vowel of "stove" will be represented with /o/.
>
>The vowel of "stove" is a _diphthong_ in British English and in
>American English. I have no idea what it is in Scots as
>spoken in Edinburgh, but seeing that there most diphthongs
>are pure, long vowels, I guess that it might perhaps be /sto:v/.
I thought the Brits called the "cookers" anyway :-) At least in the South they
did.
>At any rate, you can only write /stov/ and claim it is American
>English because American "shop" is English "sharp". How the
>bloody hell would you represent the three-way contrast
>cot/court/coat if you transcribed "coat" as /kot/ like you
>want "stove" as /stov/?
Americans seem to have the caught/cart merger.
Steve Hayes
http://www.suite101.com/myhome.cfm/methodius
Peter T. Daniels
>
> On Thu, 31 May 2001 21:02:51 GMT, "Peter T. Daniels"
> <gram...@att.net> said:
>
> >Bob Cunningham wrote:
>
> >> On Thu, 31 May 2001 00:53:17 -0400, Richard Fontana
> >> <rf...@sparky.cs.nyu.edu> said:
>
> >> [...]
>
> >> >(I think there actually are some
> >> >American dialects that have /e/ as in "bait" not diphthongized.)
>
> >> In Peter Ladefoged's Illustration of American English in _The Handbook
> >> of the International Phonetic Association_, he uses [e] in his 'narrow
> >> transcription' for the 'a' vowel in words like 'came', 'making', and
> >> 'take', while using diphthongs like [aI] ('shined') and [aU]
> >> ('around') elsewhere.
>
> >That doesn't seem "narrow" at all.
>
> If you were familiar with the _Handbook of the International Phonetic
> Association_, you would know that each language illustration has a
> section labeled "Conventions". In the narrow transcriptions, this
> makes it unnecessary to show an elaboration of extra marks that are
> common to all occurrences of a symbol.
I have looked through it in stores several times and each time
determined that it wasn't worth spending 20-odd dollars on. And I saw a
review somewhere (probably in *Language*) that agreed it was less than
satisfactory.
> My reason for mentioning the two diphthongs was to show that the use
> of [e] couldn't be explained by his eschewing diphthongs entirely in
> his transcription.
A certain modicum of essential information, then, was omitted.
Peter T. Daniels
>
> On Thu, 31 May 2001 14:24:30 +0000, Jacques Guy <jg...@alphalink.com.au> wrote:
>
> >Richard Fontana wrote:
> >
> >> I don't see why /stov/ is necessarily dead wrong, if you adopt the
> >> convention that the vowel of "stove" will be represented with /o/.
> >
> >The vowel of "stove" is a _diphthong_ in British English and in
> >American English. I have no idea what it is in Scots as
> >spoken in Edinburgh, but seeing that there most diphthongs
> >are pure, long vowels, I guess that it might perhaps be /sto:v/.
>
> I thought the Brits called the "cookers" anyway :-) At least in the South they
> did.
"The whale cookered in the side of the whaling ship." ???
Frances Kemmish
Wouldn't that have to be "The whale cooker in the side of the ship"?
But the English don't use "stove" in that way, either.
"Stove-enamelling" I've heard, but otherwise it's not too common a
word in English English.
Fran
Bob Cunningham
wrote
[courteous, instructive, and persuasive arguments to support the use
of "allophones" to refer to the various realizations of a phoneme
whether or not they are in complementary distribution.]
Okay. I will henceforth use "allophone" that way, which is the way
I've long wished I could use it, and I will quit complaining about
others using it that way.
Thanks, Neil.
Greg Lee
: Bob Cunningham wrote:
...
: > David Crystal's _Cambridge Encyclopedia of Language_ (page 161) seems
: > to agree with the _Britannica_ interpretation of "free variation". He
: > says:
: >
: > If the sounds do occur in the same place in a word, then they
: > can belong to the same phoneme only if they do not change the
: > meaning of the word. [Gives examples] Such sounds are said
: > to be in "free variation".
: This is more or less true, but in order to be correct I think
: it relies on you interpreting "in the same place in a word" as
: "in a particular phonological context". I wonder if the *word*
: has specifically to be the domain, too.
Crystal probably really does mean "word". Consider the upwards
pitch bend at the end of English yes-no questions. It clearly
distinguishes meaning, but not everyone would care to assume
an abstract "question" phoneme. The pitch bend does not occur
in a word, really, but at the end of a phrase, so stating the
criterion for free variation relative to words lets us discount
phrase phenomena.
...
--
Greg Lee <l...@Hawaii.edu>
Richard Herring
We do, but it's not exactly an everyday occurrence. (Though I
doubt if most people could tell you the present tense.)
> "Stove-enamelling" I've heard, but otherwise it's not too common a
> word in English English.
"Cooker" does seem to have taken over (it's a broader
term - one wouldn't have a microwave stove) but I think my parents'
generation would have used them interchangeably.
--
Richard Herring | <richard...@baesystems.com>
Rob Bannister
When I was young in England, we always said 'oven'; eg 'Light the gas on the top of the
oven.'
-- Rob Bannister
Perth, Western Australia.
T. Mike Keesey
<rf...@sparky.cs.nyu.edu> said:
>(I think there actually are some
>American dialects that have /e/ as in "bait" not diphthongized.)
Some northern midwestern dialects/accents (cf. FARGO) have /e:/ instead of
/e<j>/ or /eI/ (and also /o:/ instead of /o<w>/ or /oU/), which I think
comes from the Scandinavian accents of settlers. This is probably a
feature of a lot of foreign accents, actually.
_____________________________________________________________________________
T. MICHAEL KEESEY
Home Page <http://dinosauricon.com/keesey>
The Dinosauricon <http://dinosauricon.com>
personal <kee...@bigfoot.com> --> <t...@dinosauricon.com>
Dinosauricon-related <dino...@dinosauricon.com>
AOL Instant Messenger <Ric Blayze>
ICQ <77314901>
Yahoo! Messenger <Mighty Odinn>
Rich Wales
> Some northern midwestern dialects/accents (cf. FARGO) have
> /e:/ instead of /e<j>/ or /eI/ (and also /o:/ instead of
> /o<w>/ or /oU/), which I think comes from the Scandinavian
> accents of settlers.
This is also a common feature of Canadian English accents. I would
assume it is probably explained by the Scottish origin of many immi-
grants to Ontario.
Rich Wales ri...@webcom.com http://www.webcom.com/richw/