Re: þe_getisbyrg_adres

On Tue, 2004-08-03 at 20:43, Benct Philip Jonsson wrote:
> Tristan Mc Leay wrote:
>
> [snip correct observations about the non-dispensability of diacritics]
>
> > My problems with it are that it doesn't note stress...
>
> The system *can* optionally note stress by marking light vowels with
> a grave and heavy vowels with a circumflex instead of an acute, thus
> _konsêvd in lìbyrti_.
But that just looks yucky. Dybbl lettrs ar gods gift tu stres!

> > It randomly gives
> > values to schwas so that I took some time to decypher frex. 'konsévd',
> > _onord_; _propósiśyn_ just looks absurd (sorry, can't type s-caron, can
> > type s-acute, treating as equivalent).
>
> Fair enough, but actually using caron rather than acute on _ž, š_ helps
> reading, since they are less likely to be mistaken for an acute on the
> preceding vowel.
Oh, I didn't mean it as a suggestion for improvement, just an
explanation for my alternative. But I did come to like the idea of using
acute as a general-purpose modifier :)

>Also in this sample I limited myself to characters
> available in Latin-1.
You will find that s- and z-caron aren't Latin-1 characters if you check
the character set of Latin-1.

>Otherwise I prefer ŋ to ñ and ʒ to ç; concevably
> one could also use ƶ, ʃ for ž, š. Or to be even more radical use
> distinct letter shapes like this:
>
> Latin-1: a e i o u y á é í ó ú ý
> Unicode  a ɛ ɩ ɔ ʊ y ɑ e i o u ɥ
Oh, that's just scary. No. You lose all the points you gained by doing
that. As it stands, it looks nice and different and icelandically-
influenced. You do that and it just looks like every other screwball
IPA-based awfugraphical reform. Just because you're using some Unicode
characters doesn't mean that Latin-1's out-of-bounds.

> > You might as well write
> > 'propósiśún'.
>
> No, that would be //prQpowsISawn//! tho I guess _propósišn_ would have
> been better! I've changed that now.
I'm still contesting the use of long vowel for a schwa! If I looked hard
enough I could find examples where it's broken (there's plenty of words
where a long vowel has its full value in an unstressed syllable). (There
are current examples of the opposite, but I think the point is that the
English orthography is broken.)

> > _To_ for 'to' but _pur_ for 'poor' strikes me as amazingly
> > and unbelievably backward (historically the both represent the same
> > vowel---I could justify the lack-of-schwa on grounds of history---so why
> > not represent them the same?).
>
>   The idea of the system is
> to have, at least to a very high degree, a consistent spelling of the
> same morpheme save for accents, so thus /@/ *has* to be spelled
> differently -- with the morphonemically underlying vowel if you will.
>
> It's time that I divulge that this example was converted from ordinary
> orthography by a perl script!
See, now my earlier response would've been muchly different if you'd
said that in the first place. A bit less critical of the idea, a bit
more helpful for the implementation...

> I have a longish list of exceptional
> spellings for words and morphemes that are spelled more or less
> irregularly.  Otherwise the script strives to capture the *regularities*
> of current English orthography and convert the spelling based on them,
> and it works rather damn well given the constraints, even tho the list
> of exceptions is up to 56 items by now.  I had still forgotten a couple
> of important exceptions when I converted the text I posted, notably
> _to_ and _of_ (so that /ov/ got confused with /of/!)  Why _their_ became
> _þeyr_ rather than _þár_ is completely beyond me.  One thing which I
> alas can't do anything about without making the list of exceptions
> very long is the variation between /s/ and /z/ for intervocalic _s_.
I'm not sure it's possible. Even <ss>---even <sc>---can represent /z/ in
the right places. Though the Zompist guy did try to work it out and had
a start (something about the value of the vowels...).

Taken from <http://www.zompist.com/spell.html>:
        Voicing of s

        17. s is voiced between two vowels (amuse, design, prison),
        except after a (base, parasite).

        It's easy to find exception to this rule: disagree, opposite,
        analysis-- there's even words where the rule applies only for
        verbs (abuse, house). The rule as stated has more successes than
        failures, and I haven't been able to find merely lexical rules
        that do much better. A better rule might take the language of
        origin into account: the voicing tends to occur in French and
        Latin words (resent, please, reason, miserable), but not if
        they're from Greek (analysis, isoceles) or more exotic languages
        (papoose, Osaka).

        The voicing of s is so almost predictable that there are
        orthographic conventions (borrowed from French) to indicate that
        we really do want an s: double the s (cf. bassoon vs. basin), or
        use c instead (race vs. rase). Annoyingly, there are a few cases
        of unexpectedly voiced ss (dessert, dissolve).

[And 'Aussie', where it was clearly decided that the -ss- meant the
vowel was short, not the consonant unvoiced!]

        As a corollary of this rule, the American use of -ize for
        British -ise was unnecessary, although of course it is more
        foolproof.


> > There's also some random accents e.g. in
> > _háv_, _líviń_ which as far as I know reflects no English pronunciation
> > but rather the irregularity in the English orthography that prohibits
> > <v> from ending words or being doubled.
>
> Sure, that is a limitation of the automatic transcription.  To solve it
> I would have had to list all words with _ve_ in them, and there would
> still be some like _lives_ with two possible pronunciations.
Which just goes to show how silly a language Perl is, that it can't
distinguish between such obvious differences as 'live' and 'live'...
Might I suggest Python? It can tell the difference between everything :)

> > Similarly, _śål_ for 'shall'.
>
> I'll list that one too.
>
> > Also, 'cannot' is one word, not two :)
>
> Not in my sourcetext! :))
Your sourcetext can go jump :))) (The idea of using a text that uses an
out-of-date orthography and idiom to show a reformed orthography has
always struck me as odd. It doesn't give you the full feel of the
change.)


--
Tristan <kesuari@...>