On Mon, Jul 25, 2011 at 9:30 AM, And Rosta <[log in to unmask]> wrote:

> Peter Cyrus, On 20/07/2011 23:34:
>  Surely you don't think that the Latin alphabet is the perfect orthography
>> for English!  We have assigned letters and combinations of letters to the
>> phonemes of English by convention, and Shwa does the same thing, just
>> better.  What are the problems with Shwa that the arbitrariness and
>> unfitness-for-purpose of Latin orthography solve?
> It's very very hard to decide what the ideal script for a given language
> would be like, if we were able to start from scratch. The advantage of roman
> is that it largely frees us from such difficult decisions, while at the same
> time being panlinguistic. Of course, if one were starting from scratch, and
> using roman weren't an option, then you'd want to do much better than roman,
> but it'd be terribly hard to reach consensus on the best script.

It would be ideal if the Latin alphabet could be standardized and extended
to suit even English, not to mention other languages.  There have been
numerous attempts to do this (google "spelling reform"), and I show several
of them on the Shwa home page.  None of them seem to me to do a good job :
 - the Latin alphabet lacks enough letters, especially vowels
 - some letters (c h j x y) have no standard value
 - stress and tone are inconsistently marked
 - using two cases doubles the count of glyphs with no great benefit

The IPA addresses those issues, but at a level too low for everyday use.
 IPA "broad transcription" might be a possibility.  But Mr. Dunn
overestimates how many people are already using the Latin alphabet: only
about a quarter of the world.  So it seemed to me reasonable to explore
starting from scratch, as did Kingsley Read in inventing Shavian or Sejong
in inventing Hangul.

Forgive my evangelism, but I see the other features of Shwa, beyond simply
being a scratchbuilt panlinguistic phonemic script, as adding to its allure:
 - it's featural (like Hangul)
 - it only needs a 20-key keyboard
 - (Mr. Everson) the option of using different gaits may make it more
familiar to people who aren't using an alphabet
 - I hope it will turn out to be beautiful

But of course all the difficulties you discuss in the assignment of
spellings to words are still there, whether the spelling is Latin or Shwa.
 And it's not my role to address those issues, although obviously the script
was designed to enable solutions.

>  Perhaps I am using the terms "phonetic" and "phonemic" incorrectly.  My
>> intent is to capture the elements of spoken language at the level at which
>> speakers and listeners are aware of them.  I don't want people to have to
>> be
>> linguists to read and write.
> The more phonetic it is, the less suited it is to be a pandialectal
> orthography; and perhaps too also less suited to being the orthography of a
> single dialect in which there is a wide or discontinuous range of allophony
> for a phoneme. A key desideratum of a script is IMO that it should be easy
> to learn to read, and for that criterion at least a roughly phonemic script
> seems the best. Within a single language, phonetic motivatedness of phoneme
> symbols seems a positive disadvantage; any advantage to phonetic
> motivatedness would come only if the script is to be applied to multiple
> languages and it is desirable for one individual to be able to learn the
> orthographies of more than one language.
> If my reasoning is unclear, ask and I'll explain.
>  Having said that, the letters *are* defined by general articulatory
>> mechanism, e.g. *t * is an unvoiced coronal plosive, which covers the
>> range
>> from dental to postalveolar, and from tenuis to aspirated.  Maybe what I
>> mean by "phonemic" is simply "fuzzy phonetic" in your view.
> Possibly, yes. How is _tit_ written if, say, it is phonetically [tsI?]? How
> ia _Anne_ written if it's [i@n] in Chicago, [An] in Aberdeen, and [Ejn] in
> (old) London?
>> To illustrate the distinction I'm trying to make, English written in Shwa
>> would not use the Shwa letters for aspirated plosives to write *p t ch k*
>> at
>> the onset of stressed syllables, even though those sounds *are* aspirated
>> in
>> that context in English, and even though Shwa has dedicated letters for
>> aspirated plosives, because those are "phonetic" allophones of underlying
>> "phonemes".   The letters for the phonemes are not marked as aspirated,
>> but
>> neither are they dedicated to unaspirated sounds, so they are not being
>> "phonetically-defined" in a narrow sense, e.g. enough to write it in IPA.
> How would you decide which phonetic distinctions to ignore? For English,
> for example, aspiration is more important than voicing (on plosives).
>  In contrast, English in Shwa *does* note reduced vowels, such as the *o*
>> in
>> *atom* and the *a* in *atomic*.  Shwa has a large and expandable set of
>> consonants, but only a fixed inventory of 12 (monophthongal) vowels, so
>> the
>> transcription of vowels is more "phonetic", albeit less specific than IPA
>> (with its 28 vowels and numerous diacritics).  My feeling is that even
>> though it is a "phonological" process, downstream of phonemes, that
>> reduces
>> vowels in English (or retracts vowels next to emphatic consonants in
>> Standard Arabic, another similar example), English speakers are aware of
>> it
>> : they *hear* vowel reduction.
> Certainly vowel reduction is phonological rather than (merely) phonetic.
> But dialects and idiolects differ quite a lot in the lexical distribution of
> schwa. And writing <atom> rather than <at@m> has the advantage of
> preserving morphological unity, and thereby assisting the advanced learner,
> at the expense of much greater difficulty for thebeginning learner.
>> Shwa is simply a script - it's up to the users of languages (and their
>> academies) to use it well.For example, voiced plosives *b d **g*
>> open to *bh dh
>> gh* intervocalically in Spanish, but it is not up to me whether
>> Hispanophones choose to write that change or not - my role is simply to
>> provide them with the letters they need.  I'm not enshrining any analyses
>> (although I chose one to illustrate Spanish on the site).
> So in fact Shwa doesn't prescribe how its symbols should be applied to
> English?
>> Your final point is that "any given language would be better served by its
>> own script rather than Shwa".  In practice, quite the opposite is true, to
>> the limits of my experience : I haven't encountered a better writing
>> system
>> for *any* language (although I admire Hangul and Shavian)!  By "better", I
>> mean that Shwa generally adds missing letters (like English *sh th ch ng)
>> *,
>> consolidates duplicate letters (like English *c k q x*), uses fewer
>> strokes
>> when writing by hand, less space on a page, and fewer keys on a keyboard.
>>  In addition, Shwa letters are featural (letters that sound alike resemble
>> each other), universal and usable in several gaits (not prejudiced towards
>> alphabets).  Those are the features I was trying to maximize.
> Let us agree about using fewer strokes, less space, and fewer keys -- or at
> least let us agree that they are a different matter from the main point
> under discussion.
> When I said "any given language would be better served by its own script
> rather than Shwa", what I meant, but clumsily failed to say clearly, was
> that if you were designing a completely new orthography for a language, any
> given language would be better served by its own script designed
> specifically for that language, rather than by Shwa.
> For the purposes of this email message, I've been thinking a little about
> what an ideal orthography for English would look like (in my view), if one
> were designing one from scratch. Tentatively, I think it would have complex
> characters for stressed (sesqui)syllables -- ((schwa) + onset) + peak + coda
> -- and complex characters for consonant clusters flanked by schwas or
> word-boundaries, and also for suffixal -Z and -D (and -TH) too). The
> syllabic characters would straightforwardly decompose into onset and rime
> components, and the rime component would decompose into a peak component
> plus a coda component with the same paradigm as the consonantal characters',
> so that the formula for the characters could perhaps be summarized as
> "(((schwa +) onset +) peak +) coda". But the onset and coda components would
> not all simply consist of concatenations of elements all drawn from the same
> set, though they would have patterned internal structure to show their
> relationship to the other uni
> ts. I stress that this is tentative, without the benefit of refinement
> through prolonged consideration. If this turns into a discussion thread,
> I'll refine the ideas.
> --And.
>> On Wed, Jul 20, 2011 at 2:19 PM, And Rosta<[log in to unmask]>  wrote:
>>  Sam is right about the problems of a phonetic script, but I don't recall
>>> ever having seen a full proposal for a phonetically-based orthography,
>>> and
>>> Peter has explained that his script is phonemic. The phonemic differences
>>> among english dialects are considerable, but not necessarily to the point
>>> of
>>> being a serious barrier to mutual intelligibility (when written in
>>> phonemic
>>> orthography).
>>> By way of the feedback Peter was soliciting, I'd voice some objections to
>>> his scheme. The first is that (IMO) it is nonsensical to use
>>> phonetically-defined symbols for phonemes. A phoneme can have a very
>>> heterogeneous allophonic range and it can overlap with another phoneme's.
>>> And a given phone can be an unambiguous realization of one phoneme in one
>>> environment and of another phoneme in another environment. And different
>>> dialects use very different realizations, so basing phonemes' symbols on
>>> realization either privileges certain dialects (that have standard
>>> phoneme
>>> symbols based on their realization in that accent) or turns into a quasi
>>> phonetic orthography if different dialects get different symbols.
>>> A second objection is that phonemic analyses are rough and ready - a
>>> given
>>> analysis may be good enough for a certain job, e.g. for representing
>>> pronunciations in a dictionary, but that doesn't mean it's the Best
>>> analysis
>>> or the Right analysis. So how would Shwa go about deciding which analysis
>>> to
>>> enshrine in the orthography?
>>> Lastly, a morphophonologically based orthography (i.e. one that abstracts
>>> away from phonologically-conditioned alternations and perhaps from all
>>> allomorphy) would arguably be superior. Harder for a child to learn to
>>> read,
>>> but a truer representation of the lexicon.
>>> Having said all this, if you were designing one script to serve every
>>> language, it's not easy to see how to improve on Shwa's approach. But
>>> given
>>> that my points amount to saying that any given language would be better
>>> served by its own script rather than Shwa, there's a sense in which the
>>> roman alphabet's unfitness-for-purpose is actually a virtue: it forces a
>>> degree of arbitrariness and language-specificity and so is less exposed
>>> to
>>> the problems Shwa has.