Hallo!

R A Brown wrote:

> Because of darned GMAIL my reply got sent to Patrick & not to the list.
> Anyway, I have had one or two more thoughts so I'll sort of reply to my
> reply to Patrick  :)
> 
> R A Brown wrote:
> > Patrick Littell wrote:
> >
> >> On 9/29/05, R A Brown <[log in to unmask]> wrote:> It seems to
> >>
> >>> me that apart from on add short interjection here and there,
> >>> it is overwhelmingly Jörg and me corresponding on this thread. Maybe, at
> >>> this stage it could be brought to an end on the list.
> >
> > [snip]
> >
> >> Oh, I had been following with interest; I had simply been too busy to
> >> participate. Here is what I had been thinking the whole time:
> >
> > OK.
> >
> >> (1) Instead of borrowing the Plan B self-segmentation, which I find a
> >> little unwieldy -- neat, but inhuman --
> > 
> > I was not entirely sold on it.
> 
> I assume by inhuman Patrick means it is not a feature that is ever
> likely to appear in human natlangs. But that per_se does not rule it out
> for a loglang, as I understand it (and certainly not for an engelang).

Yes.  It is clarly an unnatural feature, but who asks for naturalism
in a loglang?  Loglangs are _per se_ unnatural.

> But it is, as I wrote, something I shall be looking at again.
> 
> >> use your vowel-deduction
> >> system!  After all, that is the interesting part of this language.  If
> >> you were, say, to eliminate one of the vowels, you wouldn't need any
> >> further system.
> >
> > Ah, a 'zero vowel', or may _virama_ ?
> 
> Yes, I have thought further. Fairly obviously 00 must signify the 'zero
> vowel'. Then, it seems to me, one would have 01 = /u/, 10 = /i/ and 11 =
> /a/.

Nice.

> Now in brz, following Jörg's revision, we have:
> g = [ku, kO]
> k = [ki, kE]
> 
> If I adopt Patrick's suggestion, we would have:
> g = [ku, ka]
> k = [ki, k]
> 
> The two axes, back-front & high-low, have gone. Does [k] fit in with a
> syllabary? Have I now got an abugida?

Sort of, I'd say.  But then, such terms do not apply easily, because
unlike in natlangs, the spoken form is a representation of the written
form here, not the other way 'round.

> Also it seems to me that it might be better to have:
> g = [ka, k]
> k = [ki, ku]
> 
> Umm....
> 
> >> So we take a vowel system i,a,u; if two consecutive
> >> consonants don't give us one of these, it's a word boundary.
> 
> Morpheme boundary, I think, is what we want.

So we are essentially at my old self-segregation scheme where each
morpheme begins and ends with a consonant, with no morpheme-internal
consonant clusters?

> >> (Jörg
> >> may also find interesting the resulting CVCVCVC structure of words,
> >> which will allow additional segmentation fun.)
> >
> > There's also the question how the 'zero consonant' is pronounced when
> > final in such a string.
> 
> Yep - I still haven't had any bright ideas about that.

Perhaps as a velar nasal?  If /m/ pairs with /p/ and /n/ pairs with
/t/, it is only logical to have /N/ pairing with /k/.

> >> This leverages the cool part of the language for another use, and lets
> >> you begin words with any consonant you want.
> >
> > This is true,
> 
> In fact, it is not entirely true. It does constrain what consonant
> symbol begins the word following the final 'boundary consonant' of the
> preceding morpheme.
> 
> > tho there are some advantages maybe in knowing in advance
> > how long a morpheme will be. I'll have to experiment - but it is an
> > interesting idea.
> 
> Also even in brz one can still begin with _any_ consonant letter; it is
> just that the initial letter constrains the length of the phoneme.

Of the morpheme, of course.

> >> (2) Have the vowel quality of the... letter? phoneme?... er... what
> >
> > [snip]
> >
> >> Or something like that; you get the idea.  Something along the lines
> >> of the Gaelic broad/narrow distinction, taken one step further.
> >
> > Yes, I had thought along similar lines at one time. I'll have another
> > think, and try out some ideas.
> >
> >> (3) If a three-vowel system isn't enough, each /a/ can anticipate the
> >> vowel of the next syllable, giving us /e/ (< */ai/) and /o/ (< */au/).
> >>  Or, it could give us /ai/ and /au/, and the /i/ and /u/ could
> >> anticipate an /a/ to give us the /e/ and /o/.
> >
> > Yes - but I am happy with three vowels. If it's good enough for
> > Classical Arabic, I can live with it. Besides quite a few of my earlier
> > experiments with "Roman letter syllabaries" have used just these three
> > vowels.
> >
> >> Anyway, that's my two cents,
> >
> > Thanks - some interesting ideas. I don't know what Jörg will make of
> > them. But I think he and I will be developing 'brz' in our own ways.
> >
> > Meantime, I really need to discipline myself and get in with giving
> > Piashi a proper vocabulary!

Piashi is definitely something worth working out to more detail.

Greetings,

Jörg.