L C
G
add R, ch, sh, zh, th, dh, ng,
Runes
 1 k 
2k  5k can be reduced
Add my PMF, remove the dashes
Even if Quickscript, Read did not move toward Latin or historic shapes.
So there is something between quickscript and readscript
air = Ar or er
25
Sept 2001
43
phonemes -
What
is the full significance of the order?
What
is the ascii version of this?
roar waw yoy hah
vision-asia-measure
Is this asking too much?
[1] I want a transcription system that is good enough
for a person who knew the code
but not the language
could read aloud a coded paragraph
and be understood by a native speaker.
[2] I want this code to be easy to reproduce - e.g, keyboard
compatible, email compatible
[3] I want to code to be easy to teach and learn [no more
than 10 exception rules]
[4] I want the code to be readable by those literate in
English [without a key].
This means that words have to
be recognizable in context.
Over 50% of the words would
have to be good "eye rhymes"
[5] It should be possible for 80% of the [english speaking]
learners to speed read
this code after 3 weeks of practice.
[6] I want for 80% of the learners to be able to spell
over 90% of the words they hear
with 95% accuracy after 3 weeks
of practice.
Justin Rye thought that [1]
would require that the code include quite a bit of phonetic detail.
It probably would require
that stress be indicated but I think that almost any pronunciation guide
spelling would work for [1].
What do you think?
I think we could teach the
Spanish code and let students read a passage in Spanish.
Their pronunciation would
not be perfect but it would probably be good enough for
someone who knew Spanish
to understand.
My ideas for teaching the
code are probably wrong. I think it is just a matter of
memorizing a correspondence
table.
Has [5] ever been studied.
Beach reported that most of his subjects could speed read
his regularized English
after 2 weeks but they could not spell it.
[6] is probably the hardest
requirement for most transcription systems.
JB
Re: Is spanglish isomorphic with ipa?
> I think you will agree that we can come up
with a
> transcription system that is easier to teach
and use
> than the traditional system.
>
> You have a prototype of such a system that
may be
> better than Spanglish. I would encourage
you to
> present it on the saundspel egroup.
Remember I've only proposed it as a phonemic
notation for my accent,
not a supplanting orthography. And I'm
afraid I'm already operating
at the limits of my free time and network
bandwidth supply, so I
can't promise to subscribe to extra mailing
lists.
> I think that your suggestion of CITATION
spelling is a
> good one. Spell it as you would if
someone asks you
> to pronounce a word clearly. Does
this mean that you
> could drop schwa as a phonogram?
Difference between
unstressed V and @
Is the difference
phonemic
It shouldn't make any difference for that.
The distinction can be
tricky to diagnose, but in the accents that
have one it should be
possible to hear a difference between unstressed
/V/ and /@/ in
citation-form versions of <gallup, hiccup>.
Those both sound the
same to me - hence /g&l@p/, /hIk@p/.
Incidentally, the citation-form rule has its
limitations; it leaves
you having to work out how to handle forms
like <-n't> and <'ve>.
aiv, cant, whee is the problem?
> These hybrids are little more than experiments
at the
> present because there is no one interested
in
> something that is a little better than what
we have.
>
> RITE spelling is a 70% solution.
> [www.unifon.org/rite.html]
>
> Cut spelling is a 60% solution.
> [www.unifon.org/cut-spl.html]
> Neither are systematic. They cannot
be reduced to
> a correspondence table. They are much
like Wijk's
> solutions - they require a 100 rules to
make them work.
Does
cut spelling address the problem? What is the problem that needs
to be addressed?
I know a bit about Cut Spelling - an interesting
kind of
gradualistic reform, but based on an extremely
weird analysis of
what the problem is!
> SPANGLISH is a 90% solution. It is systematic
and can
> be reduced to a correpondence table -- although
a
> rather complicated one. The goal of
Spanglish is to
> be as consistent for English as the Spanish
writing
> system is for Spanish.
Tricky at best - Spanish has it easy with those
vowels!
> Spanglish has a certain reluctance to add
distinctions
> that are never recognized in the traditional
writing
> system. This is because Spanglish
is being developed
> as an alternative i.t.a. -- a way
to teach the
> traditional system.
This I can follow, though it seems a shame.
> Spanglish also is reluctant to change the
look [eye
> rhyme] of international spellings.
science ciencia
Doomed to eternal
irregularity
That's a lot of ordinary English words you're
dooming to eternal
irregularity... it seems a strange concern
for an ITA.
> This means that we get involved in a speaking
reform
> with respect to some Latin and Greek root
words.
> science - sains won't do except as a pronunciation
> guide spelling.
It's a disyllable for me - /s AI @ n s/.
But Spanish-speakers are
content with an orthography that "only" serves
as a perfect guide to
pronunciation... saians does the n need a
>> * realisational differences like [e] vs
[EI] can simply be ignored -
> Are you proposing to spell [great ape] as
[gret ep]?
No, [e] and [EI] are just _phonetic_ representations
of the sounds
(hence the square brackets), not proposed
spellings for the "ape"
phoneme.
>> [..] <<tomeyto>> and <<tomahto>>
don't need a unified spelling
>> any more than do <<egplant>> and
<<owberzhiyn>>.
>
> What spelling are you recommending?
Well, those spellings in the double-angles
are taken from my
"Romanised English" scheme, but as I keep
insisting, I'm not
recommending it, just mining it for examples
(words like
<<owberzhiyn>> are very obviously in
a reformed orthography!)
>> Why would you ever want to merge phonemes?
>
> That is a key issue with the Saundspel group
where
> nearly everyone wants to merge some of the
phoenems of
> speech. The usual rationale is [1]
to keep it simple
?! It may simplify writing, but it complicates
reading. For the
ultimate in simplicity, reduce every phoneme
to /@/!
> and [2] to look more like the traditional
system.
I suppose when there aren't enough simple recognisable
graphemes to
go round this makes some kind of sense; but
the examples you give
below hardly resemble traditional spellings.
Indeed, using <<o>> to
represent an unrounded vowel would strike
British English-speakers
as extraordinary.
> One orthography wants to use [o] for [a:
o and o:]
...That is, to use <<o>> for... three
phonemes, I'm not sure I see
which. If you mean that <balm, bomb,
Baum> become homographs then
a suspicion arises: would the proponents of
these systems just
happen to be speakers of the US accents that
make them homophones?
> and write Army as ormy and Art as ort.
In these
> systems [or] is usually spelled Or.
[@ur]
If that's your "board" vowel, remember that
it isn't a separate
phoneme for me - I say /bO:d/.
What is O, caps suggest a short vowel, but
this one has an extender.
JBR
Ankh kak! (Ancient Egyptian blessing)
24 Sept, 2001
JB,
I didn't become serious about linguistics until
I
started teaching typography classes about
14 years ago.
I tended to add some linguistic detail to
the letter
forms. Unfortunately, many students
did not
appreciate the addition of historical and
linguistic detail.
I was the speaker for the annual banquet meeting
of
the philosophy club where I spoke on the history
of
the alphabet and its cross discipline relevance.
This
talk seemed to spark an interest particularly
with a
few art students.
Regards, Steve
[more below]
[Orthogrpahic challenge near the end]
--- J B Rye <jbr@xibalba.demon.co.uk> wrote:
> Well, there's nothing necessarily wrong with
being a
> self educated linguist. I may have ended
> up taking a degree, but I was a teenage
"self
> educated linguist" before that, and probably
a
> better one than I am now.
Steve wrote:
I want a transcription system that is good
enough
that a person who knew the code but not the
language
could read a coded paragraph and be
understood by a native speaker.
I think that a person can read pronunciation
guide spelling and do this.
I want this code to be easy to reproduce -
keyboard compatible, email compatible.
and I want it to be read by those literate
in English without a key.
Children already know the language but not
the code. A different problem.
Singer's often learn to sing a song phonetically
How much phonetic detail is required to realize
this goal?
> This automatically requires the script to
go into
> more phonetic
> detail than would be required for a phonemic
> orthography of the
> foreign language in question. If,
for instance, the
> word was the
> French <ici>, "here", phonemically /isi/
with
> absolutely no tricky
> sounds or un-English sequences, it still
might
> easily get mangled as
> ['isi] unless the stress-pattern was marked.
That's
> a phonetic
> detail, not phonemic in French, but nonetheless
> important for
> intelligibility.
>
> And if this transcription is intended to
allow
> _Polish_ speakers to
> read French on sight too... well, we've
had IPA
> since the C19, and
> this is what it was designed for.
>
> > It would be similar to phonetic singing,
but there
> > would be no coach to correct the pronunciations,
> only
> > the "sheet music".
> >
> > Setting it up in this way I don't require
> perfection.
> > There is a clear point at which a system
is "good
> > enough" which is far short of the ideal.
>
> Unfortunately, an Anglophone monoglot's
idea of a
> "good enough"
> approximation to (eg) French <u> is unlikely
to be
> recognisable;
> two sounds that seem almost indistinguishable
to an
> English-speaker
> (such as [y] and [ju:]) can sound utterly
unrelated
> to a French-speaker.
> (Hang on, why am I using French examples?
Hmm,
> because I can't
> think of a good equivalent in Spanish or
> Portuguese...)
Here is more on the orthographic challenge.
THE ORTHOGRAPHIC
CHALLENGE
Three step process: 1. encode speech, 2,
decode and pronounce, 3. interpret the acoustic
productions.
Native speakers using a traditional writing
system can easily do this with 95% accuracy.
The
challenge is
to do it with non-native speakers at stages 1
and 2.
One of the first steps to be established would
be a base line. This could be done with
highly
phonemic
writing systems such as Spanish. Two students
would learn the sound-symbol correspondences.
The
first would listent to a native speaker read a
short paragraph. This could be done
slowly and each
word
could be pronounced twice. This transcription
would be handed to a second student who would
try to
convert the letter sequences into speech. The
speech would be recorded by another literate
native
speaker. The starting paragraph and the two
transcriptions would be compared and the errors
at
each
stage recorded. The hypothesis: 95% of the
text would remain intact.
This is a version of the gossip game where a
story is whispered into one person's ear who
in turn
whispers it to someone else. At the end of
the chain, the original story is compared
to the
ending story to
see to what extent it has been simplified,
embellished and changed.
Re: Is spanglish isomorphic
with ipa?
> When you say no obvious connection with IPA what
> are you saying? I was just saying that it is
> isomorphic with the IPA symbols.
I'm sorry, I no longer remember what it is we're comparing IPA to,
but whatever it is I seriously doubt it has a symbol for
"laryngealised voiceless palatal lateral fricative" (IPA
curly-barred lambda with subscript tilde)!
> Having a one to one correspondence with IPA symbols
> seems to indicate an obvious connection.
(Digs out previous message with chart...) Which columns are you
talking about? None of them match up except the second and fourth,
which are identical. There are vanishing lengthmarks, just for
a
start.
> What is less obvious is the way the symbols are
> used to form syllables and words in two dialects.
> I am still working on that.
>
> Check out
> http://www.unifon.org/14-vowels-10notations.html
Groan... it had better be labelled.
> I have a question regarding the ipa symbols
> @r and 3r as used in some GA pronunciation guides.
>
> It would seem to be that the GA pronunciation of
> RRRRRRR is a pure vowel with little relationship to
> the non-rhotic pronunciation: 33333333333333333
Well, articulatorily they're very similar, but the subtle
retroflexing has a disproportionately major influence on the
acoustics, yes.
> Both are pure or uncombined vowels but in some IPA
> notations they are written as blends.
No, that would be @-backspace-and-overstrike-r! You mean they're
written as sequences or digraphs.
> What is the relationship between
> a stressed @@@@@@@@@@@@@@@@@@@ and 3333333333
> if we can be permited to talk about a stressed
> unstressed midlax vowel.
In my own accent ("http://www.xibalba.demon.co.uk/jbr/phono.html") I
consider "<u> as in <bud>" the same phoneme as "<a> as in
<about>",
and write them both as /@/, but I take it you mean plain schwa.
> Saying that the 3 sound is "her" without the R
> is not very precise.
Very true. Well... I take it you've seen charts of the Cardinal
Vowels? (If not, see under Cardinal Vowels in Crystal's Dictionary.)
> How would this sound differ from h^ or ha:?
[@] aka [@] is dead-centre on the trapezoid, but in fact wanders
into the ranges of the following
two.
[V"] aka [3] is always close to half-way back but varies slightly in
height around the middle;
for me its average location is
slightly lower than the
half-way point, but it's mostly
distinguished by being a
long stressed vowel.
[V] aka [^] is in principle the symbol for Cardinal Vowel Fourteen,
the unrounded counterpart
of [O], but the English phoneme
that carries this label
is normally quite a lot further
forward than that.
The vowel you call "a:" is another one that's very different in
different accents - for me "cart" has a low back vowel /A:/, but
there's no point trying to locate "ur" in terms of a reference point
this wobbly.
JBR
Ankh kak! (Ancient Egyptian blessing) 24Sept
Re: Is spanglish isomorphic
with ipa?
>> (Tut, you're crossbreeding Latin "uni-" with Greek "-phon" and
>> "-graph"... I may have to report this to the neologism kennel club!)
>
> I am not responsible for UNIFON /UnifOn/
> I am responsible for unigraphic which was just a way
> to contrast digraphic. Doesn't the neologism kennel
> club allow mongrels?
Traditionally, no - they really hate coinages like "television" or
"homosexual". Which is why nobody pays any attention to them
any
more.
> Has anyone come up with a cross dialectical
> orthography?
Well, sure: the traditional spelling system, which is equally bad
for everyone. But being serious, the four major categories of
phonological difference between accents are all tractable:
* realisational differences like [e] vs [EI] can simply be ignored
-
it's a phonemic orthography,
not a phonetic transcription.
* systemic disputes can always be settled by the principle that the
orthography makes all the
distinctions any speaker needs -
Cockneys just have to learn
to spell <<thin, fin>> as if
they pronounced them differently.
* distributional problems ditto - I'm perfectly happy to write
<<born>>, even though
my accent drops the /r/.
* selectional variability should just be accurately represented:
<<tomeyto>> and <<tomahto>>
don't need a unified spelling
any more than do <<egplant>>
and <<owberzhiyn>>.
>> In nonrhotic English accents, there's nothing about the phonetics
of
>> [V":] in <furs>, [@] in <corner>, or [A:] in <stars> that
would
>> justify calling it a "combination". And the phonemes /V":
@ A:/ all
>> occur in words where there isn't even an <R> in the spelling
-
>> <colonel, china, half> - so it's just as unjustifiable on that
>> level. Meanwhile you've left "O:" out of this section, even
though
>> in RP its behaviour parallels that of "A:" - it occurs in both
>> <corn> and <cause>.
>
> What about the notational systems that use /iu/?
Which are they? It strikes me as dumb - "ewe" begins with a
cosonant - but what _about_ what about them?
> /ju/ is not a phoneme and not a diphthong.
> It is a combination or compound symbol.
I noticed that fact; you still haven't said what the justification
for it is. Is there a "compound symbol" for /@z/?
>>> through the application of the morphemic standarization rule and
>>> the rule to obscure or merge the various possible A sounds.
>>
>> That's "<A> sounds" (sounds spelled with an <A>)?
>> Why would you want to "obscure or merge" them?
>
> Because you are teaching the traditional system and
> the traditional system uses one symbol for three or
> more sounds.
You've lost me - weren't you supposed to be explaining what a
"deconstruction" was? Why am I teaching the traditional
system?
(Are you talking about stages in an ITA pedagogy, perhaps?)
>>>> I still don't know how you're spelling <thy thigh banger
>>>> Bangor> (which for me are /DAI TAI b&N@ b&Ng@/).
>>>
>>> thai thhai banger Banngor
>>
>> <<Banngor>>? You're sure you don't mean <<Banggor>>?
>
> I am not sure at all, I am not sure exactly how the
> name is pronounced. I had assumed that there is no
> [ng] in the name.
As I said, <banger> is /b & N @/ whereas <Bangor> is /b &
N g @/. I
pronounce the same /N/ -vs- /Ng/ distinction in "clanger/clangour"
and "linger/linger" ("linger" without a /g/ being a term in
university linguistics department slang).
>> it'll be ambiguous in <withhold>.
>
> Yep. Such words might well be mispronounced.
> As a pronunciation guide it would have to be written
> with-hold. lighthouse = lait-haus.
>
> Aren't there always problems with the digraphic
> approach to an extended alphabet?
Fair enough, I suppose.
>> Pointless anecdote: there was a time a few years ago when I had a
>> flatmate named <John Mather> - pronounced /mEID@/ - and a flatmate
>> named <Jon> whose _girlfriend's_ surname was /mEIT@/.
>
> Should people be required to spell their names in
> a phonemic notation?
Not if they choose to consider their birth certificates definitive.
But I suspect Lucy Mather wouldn't - it probably said <Lucinda>,
and
she's about to change her surname anyway!
Hang on, what am I talking about? (Checks) My birth certificate
nowhere asserts that my surname is <RYE>, only that my parents both
used that surname at the time of my birth! But I have no idea
whether this is true in other countries.
>>> so it is used for /D/.
>>
>> Er, you mean it (that is, <<tthe>>) is used for /T/.
>
> That is an option as in utther which looks so bad that
> I use vther as an option.
>
> iethher
What? A few exchanges back you were advocating /D/ = simple <<th>>,
/T/ = <<thh>>; but now you're using <<thh>> in contexts
where surely
everybody has /D/. Looks to me as if I've found the biggest flaw:
this spelling is impossible to remember, whereas <<dh>> for voiced
and <<th>> for voiceless is easy.
>>>> Oh, and what are the rules for "soft" and "hard" <<C>>?
>>>
>>> No change here. The rules of Latin apply.
>>> citty /sity/ and can be cut to city since y shows
>>> first syllable stress.
>>> Unfortunately science might have to be respelled
>>> sains or saints unless we allow a continental
>>> pronunciation /sEcns/ /si@ns/
>>
>> If you want to make the initial consonant internationally
>> recognisable and your system allows "soft <C>", I'd like to call
>> your attention to the Spanish word <ciencia>.
>
> Which is pronounced sEensEq si:ensi:a:
> What is your point?
Well, the topic of conversation is still there in the context:
"soft" and "hard" <<C>>. Then you raised the subject of
<science>,
with its initial <sc>, so I pointed out that writing it with <<c>>
instead of <<s>> would increase its international recognisability
in
some quarters.
(I wish you'd stop distracting yourself from the spelling-reform
issues with these unimplementable speaking-reform proposals.)
>>>>> yirly is a case in point. y before a vowel is a
>>>>> consonant. Otherwise it is a vowel.
>>>>
>>>> This seems to require that <<hurrying>> becomes <<hurriing>>.
>>>
>>> hurry hurryying
>>
>> Er... but there's no consonant /j/ between the two
>> vowels. At least, not the way I say it.
>
> Can you actually tell the difference between
> hurry-ing and hurry-ying?
The words would sound very similar (as would, say, <narrowing> and
<narrow-wing>), but not identical.
> I don't think it would be noticed in natural speech. A person
> trying to read the word aloud would tend to slur it.
Everything tends to get smeared into an indecipherable blur when
you're drunkenly shouting the words down the phone from a heavy
metal concert during a thunderstorm, but the spelling system can't
rely on such conditions to cover its inadequacies! When someone
specifically asks "Did you say narrowing or narrow-wing?", it's
usually easy to give an unslurred answer. Which leads neatly
on to:
> Give me some more information on CITATION FORMS.
> I think I understand what you mean but please
> elaborate.
A citation form is essentially how you pronounce a word when someone
asks you how it's pronounced - not slurred, not fitted into a
sentential stress contour, just the dictionary entry.
JBR
Ankh kak! (Ancient Egyptian blessing)
24 June
Re: [unifon] Re: Uttak uv thu leengwists
Steve Bett wrote:
> Tom is a self educated linguist of sorts
Well, there's nothing necessarily wrong with that - I may have ended
up taking a degree, but I was a teenage "self educated linguist"
before that, and probably a better one than I am now.
> who has
> relied on his ear and the talking dictionary to build
> a dictionary of phonemic spellings. I think he is an
> air traffic controller in New Jersey. He is connected
> with the FAA in some way.
>
> see www.unifon.org/truespel.html
>
> Among other things, Tom insists that [ng] is not a
> phoneme
You mean /N/ as in <hanging>? Are you quite sure he's not from
Lancashire, where that's true?
> and that the talking dictionary pronounces
> [won't] as wuent [wu:nt].
He also seems to have a rather odd attitude to /i:/!
> You cannot use technical terms in your discussions
> unless you provide examples.
Ha, that's what you think! It turns out to be easy to use technical
terms without explanation, once you get the hang of it.
> Tom has tried to use his ear and his truespel system
And, unfortunately, his brain...
> to transcribe foreign languages. I only know Spanish
> and Portuguese so I can only say that his
> transcription of a portugues speaker was way off.
>
> In the case of "wuent yue help" it would be possible
> to understand what was intended. In the case of some
> of the Portuguese transcriptions, reading them back
> would not be undrestandable or coherent to a
> Portuguese speaker.
>
> This indicates the kind of objective that I have.
> I want a transcription system that is good enough that
> a person who did not know the language but did not the
> code, could read a script and be understood by a
> native speaker.
This automatically requires the script to go into more phonetic
detail than would be required for a phonemic orthography of the
foreign language in question. If, for instance, the word was
the
French <ici>, "here", phonemically /isi/ with absolutely no tricky
sounds or un-English sequences, it still might easily get mangled as
['isi] unless the stress-pattern was marked. That's a phonetic
detail, not phonemic in French, but nonetheless important for
intelligibility.
And if this transcription is intended to allow _Polish_ speakers to
read French on sight too... well, we've had IPA since the C19, and
this is what it was designed for.
> It would be similar to phonetic singing, but there
> would be no coach to correct the pronunciations, only
> the "sheet music".
>
> Setting it up in this way I don't require perfection.
> There is a clear point at which a system is "good
> enough" which is far short of the ideal.
Unfortunately, an Anglophone monoglot's idea of a "good enough"
approximation to (eg) French <u> is unlikely to be recognisable;
two sounds that seem almost indistinguishable to an English-speaker
(such as [y] and [ju:]) can sound utterly unrelated to a
French-speaker.
(Hang on, why am I using French examples? Hmm, because I can't
think of a good equivalent in Spanish or Portuguese...)
>>> This iz aufool. Ie thhaut wee wer trieyeeng tue
>>> straetin out thu leengwwistik mes that thae kreeyyaetid.
>
> The reference was to some IPA spellings such as
> /sing/ which Tom spells as SEENG without making
> NG a phoneme. Its SEEN+G
IPA /sing/ isn't possible as a sequence in English! I think you
mean ess-iota-eng.
> I will have to find some of the other objections.
Well, if that's number one I'm not desperate to hear number two.
>>> "er" az in "her" iz u pyer voul,
>>
>> Are you thinking of the retroflex-coloured vowel in
>> GA, Yes, Tom's system is pure GA.
>>
>> And when you say "pure", do you mean a monophthong,
>> like [i:] as in "bee"?
>
> Yes. [er] is a pure phoneme just like /i:/
> Tom does not believe that er can be broken down into
> @ + r which he thinks is the IPA approach.
He's wrestling phantoms - the IPA for the vowel in GA <her> is
reversed-epsilon-hook. The International Phonetic Association
don't
prescribe a system for phonemic transcription - that would be a job
for the Local Phonemic Association!
>>> but "ie" az in "pie" iz naat.
> Tom does not beieve that [ie] can be subdivided into
> a or @ plus i or i: [its not the same as aaa+eeee].
Poor guy. Would it be cruel to show him spectrograms?
>>> This chaart iz raung.
>
> I would have thought that you would defend the chart.
Not without some idea what he was talking about!
> I don't have any trouble with the IPA transcription of
> RP. I have a question with respect to herder being
> represented as /'h3rd@r/
You have? So, er, what is it?
> I realize that you did not invent IPA, SAMPA, or any
> of the various IPA based alternatives. If you have
> problems with any of these start there.
As it happens I do have a couple of minor quibbles with Evan
Kirshenbaum's ASCIIification (eg: it's irritating to need diacritics
on two of my un-American vowel phonemes), but it wouldn't do much
good to whinge to you about them!
>>> PS. Lets staart rieteeng in our sistimz heer.
>
>> Where's here?
>
> The saundspel forum.
I'm not subscribed to the groups, so anything I Cc: to them just
bounces.
JBR
Ankh kak! (Ancient Egyptian blessing) 24 Sept
unifon] Re: Uttak uv thu leengwists
Tom Zurinskas wrote:
> The thing that truespel is trying to do is develop a simple phonetic
> spelling. I find that 40 phonemes is adequate for English and
actually for
> most other languages with the addition of some special phonemes English
> doesnt have.
40 phonemes would do for some accents, but I've got 43 myself, and
you'd have trouble handling one of those African languages with
120-odd! I'd advise not trying - international phonetic
transcriptions and phonemic orthographies for a particular language
are two very different things, and nothing optimised for one role is
likely to be well suited for the other.
> Do you have a simple scheme?
For transcribing random languages, IPA; the 7-bit version I'm used
to is Evan Kirshenbaum's ("http://www.kirshenbaum.net/IPA/").
In
fact there's an ASCII-IPA summary of my accent up on the web at
"http://www.xibalba.demon.co.uk/jbr/phono.html".
For radical reform of English spelling, well, I've got a sketch for
a scheme I called "Romanised English", but I never got it into a
state I considered acceptable. It's used for examples on another
of
my pages, "http://www.xibalba.demon.co.uk/jbr/ortho.html".
I usually advise reformers to think small until things get moving.
Here in the UK the best thing to do might be to start by advocating
catch-up reforms like <color>, and gradually sneak in others nobody
can plausibly argue against, such as <iland>.
> Ankh kak! - Ie ussuem this iz aank kaak in truespel?
To be honest, <ankh kak> is a poor transcription - if you asked an
Ancient Egyptian they'd say something like "`ankha ku'ak", where:
` (Arabic `ain)
is a voiced pharyngeal fricative
kh (Arabic kha)
is a voiceless uvular fricative
' (Arabic hamza)
is a glottal stop
JBR
Ankh ka-k! (May-it-live soul thy) 23 Sept
23 Sep
Re: Rhymes with avian
>>>>> certain errors become conventionalized. e.g.,
>>>>> Shavian is spelled sheivi@n rather than
>>>>> [shau]+vi@n.
>> [...]
>>> You bring up a good point but I think the example
>>> was not one of interdialectical differences.
>>
>> It isn't a disagreement over spelling - it's a question about how
>> the word's pronounced. My own preferred dictionary (Collins
>> Millennium Edition) gives it as /'SEIvi@n/ to rhyme with <avian>,
>> but I've heard Americans pronounce the first syllable as <Shaw>...
>> or was it <Shah>?
>
> That would explain it.
> If you pronounce shavian to rhyme with avian
> then people would just be spelling as they speak.
>
> Does the dictionary rhyme shaw with shave or law?
!? I expect it rhymes it with "throat-wobbler-mangrove".
But I
believe Bernard pronounced it [SO:].
> Is there another analogy where law becomes lave?
> hawk becomes havian.
<Shav-> is a Latinisation of <Shaw>, though not a very good one
(what's <sh> doing in a Latin word?) - compare the way students
at <Harrow> call themselves <Harrovians>. It's mock-scholarly
Anglo-Latin, modelled I suppose on things like <Muscovite> or
<Peruvian>.
> Americans would generally pronounce shaw as sh+awe
> to rhyme with their pronunciation of law.
Well, yes. Who doesn't?
> I think that the British pronunciation of both words
> is tipped toward shaa and laa.
No; there are US accents that have lost distinctions among the open
vowels, so that <Shaw> rhymes with <Shah> (if that's what you
mean),
but in RP <Shaw, law> is /SO: lO:/ (with lip-rounding), and <Shah,
la> are /SA: lA:/ (without lip-rounding).
> By extension, we would say shaavian rather than sheivian.
I don't follow your logic. Not that anyone needs an excuse - I
pronounce <Yugoslavian> as /j u: g @U s l A: v i: @ n/ myself.
> What does kak mean?
> I presume it is a version of the blessing of live long and prosper.
Sort of: "ankh", verb, "may it live"; "ka-", subject
noun, "soul";
"-k", possessive suffix, "thy". "Ankh kak"
= "may thy soul live".
That's the same "ka" that's now a Ford (?) trademark
- /kA:/, as
opposed to an avian "caw", /kO:/.
JBR
/"\ W W
`anxa ku'ak
\ / |___|
(better transcription!)
--+-- _______
Ancient Egyptian Blessing
| \_____/>
23 Sept
Re: Is spanglish isomorphic with ipa?
Steve Bett wrote:
> This one features Unifon II
> a unifonic unigraphic solution
(Tut, you're crossbreeding Latin "uni-" with Greek "-phon" and
"-graph"... I may have to report this to the neologism kennel club!)
> rather than a digraphic one as in the case
> of Spanglish.
>
> Contrasting British & American English Pronunciation
>
>
U2 TRUE SS
> Q Q pot
pQt (Br.) pQt paut pQt
> q A pot
pAt (Am.) pqt paat pott
> E i: easy i:zi
(Br.) Ezy eezee iezy
> E i ease iz
(Am.) Ez eez iez
> ei eI raise reIz
(Br.) eiz --- reiz
> A e raise rez
(Am.) Az aez reiz
> u: u: lose lu:z
(Br.) luz --- luuz
> u u lose luz
(Am.) luz luez luuz
> cu @U nose n@Uz
(Br.) ncuz nu-ooz??nauz
> O o nose noz
(Am.) nOz noez noaz
> O o rows roz
rOz roez rowz*
> o: O: cause kO:z
(Br.) ko:z --- coz
> o O cause kOz
(Am.) koz kauz coz
What's it intended to be used for? It distinguishes between US
and
UK phonetic realisations of the same phoneme, so it's not a
cross-dialectal orthography, but on the other hand it doesn't have
any very obvious connection with IPA.
> Combinations @ and @r
>
> R 3: furs f3:z
(Br.) fRz
> Rr 3` furs f3`z
(Am.) fRrz fcrz ferz
> c @ corner "kO:n@ (Br.)
kornc korner
> cr @` corner "kOrn@`(Am.)
korncr
> q: A: stars stA:z (Br.)
stq:z staaz
> q A` stars stA`z
(Am.) stqrz staarz
> ic I@ fears fI@z
(Br.)
ficz fi-uz
> ir I@` fears fI@`z (Am.)
firz firz
> ec e@ stairs ste@z (Br.)
stecz ste-uz
> er e@` stairs ste@`z (Am.) sterz
stairz
> Uc jU@ cures kjU@z (Br.)
kUcz kyoo-uz
> Ur jU@` cures kjU@`z (Am.) kUr
kyoorz
In nonrhotic English accents, there's nothing about the phonetics of
[V":] in <furs>, [@] in <corner>, or [A:] in <stars> that
would
justify calling it a "combination". And the phonemes /V": @ A:/
all
occur in words where there isn't even an <R> in the spelling -
<colonel, china, half> - so it's just as unjustifiable on that
level. Meanwhile you've left "O:" out of this section, even though
in RP its behaviour parallels that of "A:" - it occurs in both
<corn> and <cause>.
Oh yes, and /jU@/ isn't a phoneme - it's just a sequence of /j/ and
/U@/, though it happens to be the context where /U@/ has shown most
resistence to the merger with /O:/.
>>> You spell "cats and dogs", then you apply the morphemic
>>> principle to arrive at cats and dogs.
>>
>> Don't you mean you start by spelling it phonemically as, erm,
>> <<catts and doggz>> and then standardise the ending as <<-s>>?
>> Or <<-z>>, less familiar but closer to what you get in <horses>.
>
> Right, you seem to have picked up the Spanglish
> conventions better than I have.
> catts or caets aend doggz [daagz] or dogz [daogz]
> devolves to "cats and dogs" through the application
> of the morphemic standarization rule and the rule
> to obscure or merge the various possible A sounds.
> @ ae aa
Thats "<A> sounds" (sounds spelled with an <A>)? Why would
you want
to "obscure or merge" them?
>>> You indicated that there were cases where there
>>> was not one reading per grapheme in Spanglish. I may
>>> not be aware of these.
>>
>> They may all have been figments of my imagination - I still don't
>> know how you're spelling <thy thigh banger Bangor> (which for
me are
>> /DAI TAI b&N@ b&Ng@/).
>
> thai thhai banger Banngor
<<Banngor>>? You're sure you don't mean <<Banggor>>?
> The traditional th for /D/ is the dominant or more
> frequent spelling. [thh] is so rare
Well, it occurs about as often as /tS/, and considerably more
commonly than /OI/, /Z/, or (in RP) /E@ I@ U@/. Never mind the
ugliness of <<thh>>, what I'm worried about is that it'll be
ambiguous in <withhold>.
> that it would rarely be invoked except in the case of a possible
> confusion such as [the thhug] and [thai thhai].
How about verb-vs-noun cases such as <mouthing, teething>?
I also
have a /T/ - /D/ distinction in <loth, loathe>, but most RP speakers
seem to use /l@UD/ for both.
Pointless anecdote: there was a time a few years ago when I had a
flatmate named <John Mather> - pronounced /mEID@/ - and a flatmate
named <Jon> whose _girlfriend's_ surname was /mEIT@/.
> so it is used for /D/.
Er, you mean it (that is, <<tthe>>) is used for /T/.
> finger singer fingger singer
Fair enough. But this makes <single> <<singgl>> - I can
see that
you might not want to bother applying the rules here, but it makes
<single> a bad word to use as an example of how the rules work.
>> Oh, and what are the rules for "soft" and "hard" <<C>>?
> No change here. The rules of Latin apply.
> citty /sity/ and can be cut to city since y shows
> first syllable stress.
> Unfortunately science might have to be respelled
> sains or saints unless we allow a continental
> pronunciation /sEcns/ /si@ns/
If you want to make the initial consonant internationally
recognisable and your system allows "soft <C>", I'd like to call
your attention to the Spanish word <ciencia>.
>>> yirly is a case in point. y before a vowel is a
>>> consonant. Otherwise it is a vowel.
>>
>> This seems to require that <<hurrying>> becomes
>> <<hurriing>>.
>
> hurry hurryying
Er... but there's no consonant /j/ between the two vowels. At
least, not the way I say it.
>> All sorts of "impossible" sequences of sounds occur in connected
>> speech - great strings of consonants; interjections like /fE/, /hV/,
>> /p&/; merged phonemes (I recently heard someone pronounce <angler>
>> with a single long nasal-velar-lateral sound in the middle!); all
>> kinds of stuff. That's one of the reasons I've always said
the
>> standard spelling should be based on "citation forms".
>
> I am not sure that I follow you here. Can you expand
> it to a paragraph on citation forms.
For a definition, see Crystal's dictionary - I don't see anything
there I'd quibble with. People can write "jawanna" instead of
"do
you want to" if they feel like being slangy, but there's no need for
orthography-designers to introduce special rules for their
convenience - after all, part of the point of slang is to show
you're not interested in the rules.
JBR
Ankh kak! (Ancient Egyptian blessing)
Sep 20
Steve Bett wrote:
> Thanks for your critical analaysis.
> I wish you had given me a little more to go on
> in terms of how to repair the problems.
Constructive criticism? That would involve some effort on my part,
though, wouldn't it?
> I have attached a chart showing the correspondences
> between the Spanglish and SAMPA-IPA. This may help.
Ta.
> You have indicated that you have a phonemic
> notation for English. Would you care to share it?
Did I? Well, there's my personal ASCII-IPA transcription page
("http://www.xibalba.demon.co.uk/jbr/phono.html")... or are you
thinking of the unreleased "Romanised English" spelling-reform
scheme lurking in the background on
"http://www.xibalba.demon.co.uk/jbr/ortho.html"? I hope not -
for a
start it's partly morphemic rather than purely phonemic.
>> By the way, what happened to
>> "http://pages.whowhere.com/community/sbett/spel-link.html"?
>
> Whowhere went out of business or at least out of the
> hosting business.
My link-checking robot's had a busy couple of months...
>>>> I think that any phonemic notation can be fully described
in a
couple
>>>> of diagrams such as the ones below.
>>
>> In principle, yes, but only if no special principles are being
>> followed and it's already been established what phonemes exist to
>> be notated... and the diagrams would have to be a bit better
>> organised.
>
> ANY SUGGESTIONS FOR BETTER ORGANIZATION?
Well, the structure of the diagram needs to be determined by the
structure of the information it's designed to convey - at the most
basic level, what are the units, and is it a plain unordered list,
a
set of sublists, a "periodic table" of rows and columns, or what?
>> Did you consider the option of calling it "Spanglosaxon"?
>
> NO. THANKS FOR THE IDEA.
I've always wondered how the Anglo-Saxons all ended up calling
themselves Angles while their neighbours decided to call them
Sasnacs. And those poor Jutes must have had such a mismanaged
PR
department.
>>>> is a simple phonemic transcription system for English based on
>>>> traditional Latin symbol-sound correspondences and using all of
>>>> the devices found in Old English. The devices such as double
>>>> consonants after short stressed vowels are used today, but not
>>>> consistently.
>>
>> *Old* English? I thought this trick didn't turn up
>> until the Middle English period.
>
> It turned up in the period between OE and ME. Orm used in in
1180.
Which is a good generation into the ME period; I'll agree that Orm's
orthography is in the Anglo-Saxon tradition, less Normanised than
later attempts, but it's designed for the phonology of post-Conquest
English.
>> Er, you are aware that the phonological system has
>> *changed* since then, and might therefore not
>> necessarily be well-served by a mediaeval-reenactment
>> orthography?
>
> I am oversimplifying. I am calling an augmented Latin
> alphabet a saxon alphabet but that may be where the
> similarity ends. It is only a partial restoration.
> There is no point in restoring sounds that are no
> longer part of present day English.
> I was trying to come up with an angle that suggested
> that I was not using a foreign orthography for
> English.
Fair enough. Odd to combine it with the word "Spanglish", then!
>> It would be a lot easier to "unpack" information from these
>> diagrams if you'd put it in the diagrams in the first place -
>> quite a lot of it seems to have been left behind in your head
>> somewhere.
Sorry about the tone of this - I must have written it at at 3 am.
>> The boxes (each of which represents one... what? Not
>> grapheme, not phoneme...) are arranged in a random
>> order (largely traditional, but random); and their contents are
>> unintelligible.
>
>> To take just the top-left corner for instance: "A e,i age the".
I
>> deduce with some effort that you're using revised-orthography
>> grapheme <<A>> to symbolise the phoneme common to <ago>
and <the>
>> (that is, ASCII-IPA /@/). But what's the "e,i" about?
Your
>> diagram simply doesn't contain the information.
>
> Yes a is used for @. If you are familiar with SAMPA
I don't use it as much as Kirshenbaum ASCII-IPA, but I do know of it.
If I ever switch away from the usenet Kirshenbaum standard it'll be
to go over to full Unicode, though - come to that, why are you using
SAMPA when real IPA's just as easy to put in a gif?
Of course SAMPA's phonetic rather than phonemic, so the most
important layer of organisational information is still implicit...
> there is a chart that shows the SAXON - SAMPA
> correspondences. Are you saying that you have
> problems with SAMPA?
No (for a start, it's /@/ in Kirshenbaum ASCII-IPA too); I was
saying the diagrams you'd sent didn't contain any SAMPA information!
[questions moved to RFC1855-compliant position _before_ answers:]
>> I could go through the rest of the boxes commenting on the missing
>> items (if <<URR>> gets a box, why doesn't <<EAR>>?),
poorly chosen
>> examples (<<heir>>?) and weird allocations of graphemes (<<X>>
to
>> mean /ks/?!) but it would take me all day.
[and back to your answers:]
> URR = SAMPA 3'
That is, retroflex-hooked [3], which Unicode makes "025D,
reversed-epsilon-hook". Of course not all accents make that sound
retroflex-coloured anyway - for me the vowel is plain [3:], which
Kirshenbaum ASCII-IPA rather clumsily renders as [V":].
> EAR is the unstressed version of EIR = SAMPA e@'
"Unstressed" in what sense? The words /E@/, /DE@/ are no more
"stressed" than /bE@/, and even if there was a difference in stress
between these words, why represent it by using different vowel
symbols?
> It does not get its own box because it is not
> and uncombined sound.
<<EAR>> doesn't get its own box? Oh, you mean in the one
you sent as
"C:WINDOWSTEMPnsmailBT.gif" - "Saxon Spanglish Alphabet". (I
hadn't
realised you were typing your reply in above my question - I'll
reorder things so it makes more sense... okay, done.)
"Not an uncombined sound"? I don't follow. In my accent,
it's a
single phoneme - normally diphthongal, I'll grant you, but no more
so than /@U/ as in <bode>... and at times it's a long pure [E:]!
Of
course for my Scottish flatmates (with their fully rhotic accents)
it's not a phoneme at all, it's a sequence of vowel plus /r/; but for
them exactly the same applies to the box "UR ER surrfer".
> EAR does have its own box or cell in the 14 vowel chart
> which lists many combinations.
"14 vowel chart"... that'll be "C:WINDOWSTEMPnsmail8V.gif" - "14
Saxon Spanish Vowels with 14 combinations". 14+14? So why
are
there 37 boxes, and why is the filename "8V"? Ah, the yellow
ones
are the first 14 (and a half). The second 14 must be lost somewhere
among the greens and whites... what exactly are they "combinations"
of?
> There are three charts. One is in traditonal alphabetical order.
With a few oddities such as "D, UR, E" and "U. W UU V W"... and of
course it's been stuffed in an irrelevant 6x7 grid.
> One is in alphabetical aeiou order.
No, "A E I O W U" order. The column labels make at least a bit
of
sense, though.
> and one does has a few wierd [i.e., quasi arbitrary] locations for
> phonograms.
There's a third chart? You mean the SAMPA one you just sent?
> It does not have an X however.
>
> The X appears in the alphabet but is not listed in the
> chart of mostly uncombined phonemes.
How are you planning to spell (eg) "anxious, hacksaw, xanthoxyl"?
>> > [Image]
>>>
>>> The no stress column is unique to
>>> Spanglish although the traditional
>>> system often follows this
>>> convention. The point is that words
>>> such as sofa, the, very merry, go,
>>> intu, and other have a free vowel in
>>> the final position and that free
>>> vowel has one, usually unstressed,
>>> sound.
>>
>> I could spend a day pulling this one apart, too - in what sense
are
>> <go> and <through> "unstressed"?
>
> They are not since one syllable words cannot be said to be
> stressed or unstressed. "silo" might be a better example.
So you wouldn't actually spell "go through" as <<go thru>>?
What
would it be - <<gow thruu>>?
>> Why is the phoneme /@/ spelt differently in <<the>> and <<ago>>?
>
> Good point. There are a number of quasi IPA schemes
> that are more consistent. if a=@ the the is spelled
> tha.
>
> Spanglish has to add a rule which lists the two
> exceptions, after the crossed D
There's a crossed D?
> and before the R, @ is represented with the [e].
Before I lose track completely here, can we sort out our bracketing
conventions? I'm used to:
[phonetic transcription]
/phonemic transcription/
<standard orthography>
<<tentative orthography>>
What do you mean by the [e] here? Surely not Cardinal Vowel Two?
>> Why is <<EI>> shown twice? Why is the sequence /ju:/
as in <you>
>> called a "diphthong" when the equivalent sequence /wi:/ as in
>> <we> is ignored? And what's the idea with <<irrate>>?
> ei corresponds to SAMPA e
> eir "
er
Yes, but <<EI>> (with no <<-R>>) appears as _two_ of the
"14 pure
vowels" - once in the box "EI EY vein they" and once in the
half-size box "EI rein". Meanwhile the other half-size box is
"EA
bread", which for me is the same short-vowel phoneme as the one in
"E. edj eppic"...
> In a later version of the chart I called
> ju a combination rather than a diphthong.
Why are you disseminating an old version, then?
> It would not have to be in the chart since it is
> a combination. The usual practice is to include
> high frequency combinations that are somewhat
> problematic.
What's problematic about it? It obviously isn't a vowel phoneme
(or
we'd say "an yew tree") - it's a sequence of /j/ and /u:/. As
for
frequency, is it more common than /@Z/? I say "YU" is no more
worthy of a box to itself than "HA" or "WI".
> ir is a combination used for the sound in the words
> ear, irrate, irrigate, ....
Assuming that second one's "irate", that's three different initial
stressed vowel phonemes for me -
<ear>
/I@/
<irate>
/AI r EI t/
<irrigate>
/I r I g EI t/
... and all of them are accounted for elsewhere in the chart.
>> But my most fundamental objection is that you're using digraphs
>> and trigraphs by the bucketload with no sign of any effort to
>> make them systematic and unambiguous. How do you spell
>> <towering>, <alien>, <hurrying>?
>
> Perhaps tauring,
"Towering" for me is three syllables, and doesn't end in /g/.
> eilian,
So <<I>> can be an unstressed vowel, but it isn't listed in the
"unstressed vowels" column of the "14 Vowels" chart.
> hurrying
Why isn't the <<yi>> in <<hurrying>> pronounced the same
way as in
<<yiddish>>?
Also, I don't use my <<URR>> vowel in <hurrying> (it's plain
short
<<U>> as in <upper>). In fact the word <slurry> can
be pronounced
in two different ways in my accent, depending on whether it's the
noun "slurry" as in "slurry pit" or the adjective "slurry" as in
"slurry voice":
<slug> /s
l V g/
<slurry> /s l V r i:/
(pit)
<slur> /s
l V":/
<slurry> /s l V": r i:/
(voice)
<Furrier> is another ambiguous case: the fur-trader is /f V r i:
@/,
the comparative of furry is /f V": r i: @/.
> I have a problem with alien because I would prefer
> not to respell international words. I would prefer
> the spelling aalien
What, as if it was pronounced ahleen? That's an astounding number
of irregular spellings you're committing yourself to...
>>> [Image]
>>> What would the English writing
>>> system look like if all the devices
>>> used to distinguish long and short
>>> vowels and different diphthongs were
>>> applied rationally. That is,
>>> applied in such a way to avoid code
>>> overlaps?
>>
>> "Code Overlaps"? This is intended as an *explanation* of the
>> previous phrase?!
>
> Yes, rational, consistent, means no code overlaps.
But what _is_ a code overlap? Is this "code" as in espionage,
etiquette or software development?
>>> The result would be a
>>> phonemic transcription system
>>> similar to the one described here -
>>> Saxon Spanglish.
>>
>> Except that the system described here is making no particular
>> attempt to be "phonemic" (or even strictly graphemic).
>
> It makes an attempt.
> The attempt is supposed to be on par with Spanish.
Madrid Spanish at least is considerably closer to "one spelling per
phoneme" (and more or less perfect for "one reading per grapheme").
The only exceptions are the systematic alternations of <b> with
<v>,
<j> with soft <g>, <z> with soft <c>, initial <h> with
zero, and a
handful of special-case stressmarks.
> Chart attached. Tell me if it helps.
"21v-sampa2-ss.gif". I've got a bit more time today so I'll go
through it in more detail than the previous charts - row-by-row,
ignoring the columns for "unifon" and "englik".
* the "pot" row - _you_ may use IPA [a], but I have a rounded vowel
there (Unicode 0252, turned-script-a).
This kind of
phonetic difference shouldn't
matter to the orthography, but
since you aren't differentiating
phonetic from phonemic
analyses I feel obliged
to point them out.
* the "cost" row - nope, for me this is the same phoneme as in
"pot" (I suspect this time
it's a systemic difference, and
thus more important).
* "short vowels" - why put section headers _after_ their sections?
* the "raise" row - _you_ may say [e], but I say [EI].
* the "nose" row - _you_ may say [o], but I say [@U] (just as much
a
diphthong as the one in
"rouse").
* the "corner" row - for me this schwa is entirely nonrhotic -
"coda" is a perfect rhyme
for "coder". I happen to agree
that it makes sense for
the orthography to distinguish them
by including <<R>>s
where I don't pronounce them, but you
should be aware that this
makes the orthography more
"morphemic" and less "phonemic"
for RP-speakers.
* the "rise" row - for me it's phonetically [AI]
* the "are" row - for me it's phonetically [A:], and the phoneme
/A:/ as in "calf" and "cars"
belongs alongside /O:/ as in
"cause" and "corn".
* the "ore" row - nonphonemic for me: it's merged with [O:].
Indeed, I pronounce <or,
ore, oar, awe> as homophones!
* why are [3`], [@`] repeated in the "herder" row?
* what no "cure" row? I have an extra /U@/ phoneme, though again
it's in danger of merging
with /O:/.
And throughout, your use of different graphemes for "stressed" and
"unstressed", whatever you really mean by those, is an inexplicable
extra complication which makes it impossible to work out how you're
proposing to spell actual words.
=====================================================================
21 Sept
Steve Bett wrote:
> There is, for instance, a Shavian discussion group
> that tries to use a coding system that makes it almost
> impossible to write a paragraph in Shavian without a
> dozen errors. One thing that has happened is that
> certain errors become conventionalized. e.g.,
>
> Shavian is spelled sheivi@n rather than [shau]+vi@n.
Well, serves them right for trying to skip the difficult part!
It's
simple enough to list the phonemes of your accent and map them onto
graphemes - the hard tasks are sorting out reasonable compromises
with interdialectal differences, morphophonemic questions and
learners' traditional-orthography habits.
> I think that there needs to be a simple ascii
> unigraphic system that anyone who is developing
> extended alphabets can use. A kind of universal
> phonemic keyboard.
The problem is that only phonetics is universal - phonemics is
necessarily local. A transcription system that handles your
phonology perfectly won't match mine; what's needed is an
"archiphonological" transcription that can cope with all the major
accents of English. Probably possible, but it'll involve enough
judgement calls that it's unlikely to serve as a neutral basis for
development.
> I don't think that unicode is the answer to phonemic
> speed typing. It is a good solution for a web page.
It will be once Unicode support has sunk in properly - at present
ASCIIifications are the only way not to lock out users of text-only
browsers.
> How do you deal with @u and ei? Are you calling them
> both diphthongs?
Well, for me they are diphthongs, and in ASCII-IPA transcriptions I
just write them as /@U/, /EI/, even though in principle that's
ambiguous - cf /tS/ for the <ch> phoneme, ambiguous with /t S/ as
in
<hotshot>).
In spelling reform schemes I advocate giving common phonemes the
simplest available graphemes regardless of whether or not they're
phonetically compounds. As it happens in my own pet scheme I
gave
/@U/ and /EI/ the compound graphemes <<ow, ey>> before a consonant
(simple <<o, e>> elsewhere), but that doesn't mean they're sequences
of vowel plus semivowel any more than <<dh>> is a sequence of
consonant plus aspirate.
> I was calling them pure vowels which
> are often pronounced as diphthongs in some dialects of
> English. Do you have a category for compounds such as
> ju: and wi: ?
They're just plain sequences of phonemes, no more special than the
common sequences /tr/ or /@z/ (oops, I wrote that as /@Z/ yesterday;
much less common!). The /ju:/ sequence is more interesting for
phonologists than /wi:/ since it's one of the main diagnostics of a
big US/UK accent thing (CNN nooz vs BBC neeooz), but that's
perfectly explicable in terms of the distribution of /j/. (To
go
into detail: US accents have extended postconsonantal yod-drop to a
wider range of contexts than most, though the front-runner is still
Norfolk dialect, where "beauty" and "booty" are homophones.)
And
come to that, /@z/ is pretty interesting too, since most Americans
pronounce the plural-marker <es> with a schwa while many Brits say
/Iz/.
JBR
Ankh kak! (Ancient Egyptian blessing)
Hi Allan,
I dont necessarily want one symbl for one sound.
I want 'trap' and 'bath' each spelt one way, and one way only,
even tho u and i pronounce 'bath' difrntly from Steve.
How would u sujest Steve spell 'bath', and how should u spell
it?
In my publishing work we are looking at distinguishing lexical sets
such as this by using acceptable alternatives in the way a letter can be
written.
For instance we could distinguish the 'a' in 'trap' from the 'a' in
'bath' in this way. For Steve and you the letter 'a' would sound
the same.
The letter 'a' on the other hand would be sounded differently by you
and Steve (but be spelt the same).
Chris
Steve,
Obviously, I am a sucker for fine graphics, so I saw nothing terribly
wrong with your saxon-alfa webpage. And as for Justin Rye's comments,
all I can say is that he may be perfectly correct, on all counts.
But
I
wonder if he has ever made any such presentations of his own (which
he
allowed to be criticized), or whether is has been content to merely
copy
what the IPA folks have come up with, and be their cheerleader.
So,
it's a hell of a lot easier to criticize, but a whole lot more
difficult
to try to be innovative. Besides, what you have done has BEEN
FUN, no?
So who cares, as long as you are satisfied yourself that what you have
done, and are doing, is worthwhile.
Steve Bett wrote:
>
> So far most of the English linguists that I have
> talked
> to seem to be of the persuasion that if you cannot do
> it right, then you shouldn't do it.
Come now! They really can't be serious, can they? Who was
the last
Nobel Prize winner who was afraid to do things any way but the "right"
way? And just who decides what that "right" way is? So,
who died and
made such an English linguist God?
> They will unload on any attempt to develop something
> a little short of IPA and see no point in using it as
> a writing system.
Just let them drown (or be bored to death) in their own lack of
imagination. If they are over thirteen years old, there's no
need to
even try to teach them how to think creatively. It's just too
late, no
matter how many Ph.D.s they have.
> They believe that you cannot come up with a system
> that
> will be suitable for both RP and GA.
That may be perfectly true, but who says we can't keep investigating?
> Perhaps you can do a better job of describing this
> position. If these people play around with a writing
> system, they rarely ever publish it.
I don't publish on such things either. I wouldn't want to compete
with
Mad magazine, would I?
> I am not sure how Well's pulled off SAMPA.
> Perhaps he did receive lots of critiques.
> We just do not see the critiques of the different
> attempts to describe English speech.
SAMPA fulfilled a definite need, and still does. If you want to
transcribe something in ascii, it's the most complete, and perhaps
the
best, simply because it's SOOOOO simple. It appeals to my sense
of
admiration for things that are ridiculously understated.
> I wonder if there is a way to draw linguists into the
> challenge. Is there a way to get them to provide
> constructive criticism?
It's hard NOT to have a linguist criticize something, but then, most
of
it does tend toward the "destructive" criticize type. The SaundSpel
forum is great for getting criticism, and it's not so cluttered up
with
opinionated Ph.D.s, so you should count your blessings.
David
20 sept Paul
I imagine everything I have done has been done
before because I
just followed my nose. I know there have been a lot of people
working to improve our alphabet and could only go so far as
individuals. It's great that Unifon has a web page to help with
the
development of some of these ideas. Maybe the internet and the
high
tech environment we live in will make the difference.
I appreciate the efforts Steve Bett is making
to the cause. He
has been a great help to me. I never thought I would find the
dictionary interesting.
Paul Stought
ADDENDA: some of the points about the first two charts that I didn't
get round to last time. Just one question from the 14 Vowels
chart:
in what sense is V a "semivowel"?
Over to the SS Alphabet:
* I gather this is essentially a list of graphemes, but why in that
case are some of them (for
instance "IE I. Y") forced to
share a box?
* I understand the dots and the colour-coding, I think, but what
does the use of italics,
small caps, large caps, asterisks,
and apostrophes signify?
* Using <<heir>> as an example of <<h>> is needlessly confusing
-
remember, <heir> has
no /h/, and readers can't be expected
to know that <<heir>>
means <hair> or <hare> until they've
understood the rest of your
system! What was wrong with
<<hill>>?
* What's <<tippy>>?
* When you say "judj jvj", what does the "jvj" mean, and why isn't
it <<jujj>>?
* I hadn't noticed "K qu kick quit". You're keeping <<Q>>
as well
as <<X>>? How
do you spell <awkward>?
* If <<littl>> is the two-syllable word <little>, how do you
spell
<waddling duckling>?
* What does "N 'n nvn n'n" mean?
* How (to repeat myself) are you distinguishing /N/ (as in <singer,
sinking> from /Ng/ as in
<finger, single>? And what's the
dot for in ".NG"?
* "O AO dog _awe_" - I pronounce <dog> with the same short rounded
vowel as <otter>, not
the /O:/ in <awe> (etc).
* "T tot tott" - er, so why don't you also say "B bib bibb", "D did
didd" and so on?
* I don't know what's going on with "Th thh _thy_ _thhai_", but it
leaves me entirely at a
loss as to how I'm meant to be
distinguishing the voiced
dental fricative phoneme /D/ as in
<thy, loathe, this'll>
from the voiceless dental fricative
phoneme /T/ as in <thigh,
loth, thistle>. The obvious
approach is <<dhidher>>
vs <<thertieth>>, but you only seem
to have allocated one box,
and I can't tell which one it's
for!
* Using <<v>> as a vowel is insane.
* Using <<w>> a vowel might just work, as long as people can
cope with
Welsh-looking spellings
for words like <woodenly>...
* Why "W." and "Y."? What's the dot for?
20 sep
You can use o for aa if you are content with
> ort and ormy as spellings. This is the way that
> art and army are spelled in Unifon I and the only
> trouble is with the eye rhymes. I have a hard time
> pronouncing or as aar.
Yes, I agree that the words with an "a" (pronounced as broad "ah") in
them, traditionally, look very strange if spelled with the "o" in
"pot". But then, either you keep the "o" for broad "ah" or you
don't.
And when you don't, of course, you get "pot" being spelled "paat" and
that looks almost as strange. There is really NO satisfactory
solution,
as long as you use General American pronunciation. Received
pronunciation is almost as bad (but not quite), with its double and
triple uses of the short "o". So there you have it. If
you want joint
spellings for all varieties of English, the best thing to do is keep
the
present spelling system. It's not consistent, but the Americans
know
how to pronounce "pot", as do the British.
Of course, there are compromises, but people don't like them very much
either. Based strictly on historial development, we should keep
the
British short "o", and forget the innovative American pronunciation.
On
the other hand, we should keep the American "ae" in "path" and forget
the innovative British pronunciation. In all such cases of variant
pronunciations, we can use historial linguistic information to find
which modern pronunciation is most conservative, and then key the new
spellings to that. But THAT whole process will go over like a
lead
balloon, too.
19 sep 2001
Poem in 2 transcriptions -- www.
Steve,
I think Unifont II looks a lot better than
Unifont I.
I'm not crazy about your new symbols though.
The standard alphabet has well established names for all the letters.
New symbols are going to be awkward. I think we can do pretty well
without them.
I used to think we didn't need the crutch of
stress identification but I see that it's not that hard to do in a lot
of cases, and so it would seem a sin of omission to leave it out.
I like your use of -y- for an unstressed -e, but I still think you
should use the -u- for a stressed schwa. (up, ul, ur).
I think you should drop the single symbols
for the -th- sounds etc. I think ch, ng, Th, (small caps), th and
zh, should be used. Except for Th and zh, they are known values,
and Th and zh would quickly become obvious. You are already using
letter pairs, au, oi, what's a few more? It looks better and more
natural also. If you use ch, that will free up c to be used for the
unstressed schwa.
You are using small caps sometimes, but not
always, for words like ape, eagle, and oats. Seems like you should
make a choice that shows consistency. I think the small caps look
good and can be made distinctive enough to use.
You seem to be going to use -xr- but still
show -R- in your chart. Also, in your chart, you are using the word
fqx. You don't plan on
using -x- in two positions do you?
Your reference to the pronunciation of the
word -pair- is interesting. I couldn't find a reference to -pAr-
but I don't think it matters. I was recently reading the pronunciation
section of a couple dictionaries, and was impressed by the fact that they
don't tell you exactly how to pronounce their symbols. They make
allowances for different dialects and leave it to the reader to put his
own spin on it. For instance the use of -o- for awe: suggested
keywords are, saw, caught, and for. The word sought, could be pronounced
saw+t or saw+ut. For- is not expected to always be pronounced faw+r,
but to be pronounced according to how the reader is used to pronouncing
-or- words that are identified by the dictionary with the aforementioned
symbol. Sort in Unifont II would not likely be saw+rt, but sort (to).
Some -or- words have a different symbol suggested and would be pronounced
with another set of options. Merry and air should be spelled, mery,
and er. No other options required. It would just confuse.
If you wish to pronounce Mary as mAry, you would use the symbol identified
already as the -A- sound.
I have an idea about speed typing a phonemic
transcription. You have seen my work on Shortrit and Foksrit.
I'm thinking I could write in Shortrit and have that converted to Foksrit
or Unifont II or Unifont I. What do you think?
Paul
|
