On Tue, Jul 7, 2009 at 5:42 PM, Eldin Raigmore<[log in to unmask]> wrote:
> You need about fourteen or sixteen "bottom diacritics", about as many "bottom
> diacritics" as you have consonants.
> /fl/, /fr/, /sl/, /sm/, /sn/, /sw/, /vr/,
> /sk/, /sp/, /st/,
> /bl/, /br/, /kl/, /kr/, /dr/, /dj/, /gl/, /gr/, /kj/, /mj/, /nj/, /pl/, /pr/, /tr/, /tw/, /tj
> /, all are onsets in English,
> and questionably so are /ps/ and /ts/.

I'm going with the "ignorance is bliss" approach where I only have to
devise symbols to have a close-enough symbol, since the syllabary I
have in mind is NOT strictly phonetic.

as for  /dj/, /kj/, /mj/, /nj/,  /tj/ I'm not sure what words you have
in mind, but my dictionary doesn't have any of those as word initial
clusters, and as I've mentioned in several other posts, I have no
objection to redefining the concept of "syllable" in English to make
the syllabary work with English.

> I don't know how you're planning to handle three-consonant onset-clusters.
> There aren't as many of them; they are all /s/ + /k, p, t/ + /l, r, w, j/.

I will redefine "syllable" so that, for example, "scrap" is a three
syllable word:

scrap -> skø•ra•pø, where "k" is a diacritic above the "sø" symbol.

> /sIksTs/ has a four-consonant coda-cluster.  I don't know what the coda-
> clusters in English are; if your "top diacritics" apply only to continuants, then
> the /ksTs/ will have to be broken into /k@s/+/T@s/, I guess.

What is "/sIksTs/"? I don't find anything like that in my dictionary.

As an aside, only about 15% of words in an average English corpus
require anything other than generic vowels to be distinguishable,
especially when degenerate consonants are replaced by their actual
value. For example, "øm•bø•shøn" can only be "ambition", and "sør•tøn"
can only be "certain". (as opposed to "cør•tøn" which could be
"certain" or "curtain" due to the ambiguity of "c")

There only need to be enough vowel mod diacritics to disambiguate
minimal sets like {alive, olive}, and those diacritics don't
necessarily have to be phonetic in nature, for ø•lø•vø {alive, olive}
it would be enough to mark one or the other syllable symbol as
stressed. Sets like ø•dø•ptø = {adapt, adopt, adept} would require
three distinct vowel mod diacritics, but such sets are very rare,
showing up in less than 1% of the 1,000 most common English words.
(For what it's worth "ptø", "pdø", "btø" and "bdø" could all be
represented by a single symbol since "adapt", "adapd", "adabt", and
"adabd" are all interchangeable without loss of intelligibility.)

It might be enough to simply assume three vowels, (round, broad,
open?) that would be close enough approximations to give unambiguous
symbols for all minimal sets, especially with a diacritic that meant
simply "diphthong" without specifying which diphthong. Thus {abbot,
about, abut would all have the same "spelling" but the first would
show the stress on the first symbol, the second would show stress and
a round diphthong on the second symbol, and the last would show stress
and an open vowel on the second symbol.

In many cases the disambiguation could be left to contextual clues.
The sentence "Is he øløve or dead?" probably references the word
"alive" rather than "olive".

The basic idea is to represent a word using a system loosely based on
unique symbols for open syllables (with a restricted set of allowable
codas), reanalyzing English words as necessary to fit that
requirement, and with no attempt to use the syllabary to specify the
actual sound of the syllable with any precision beyond the bare
minimum needed to distinguish it from other similar words.