Re: Tunu not dead

On 10/6/05, Taka Tunu <takatunu@...> wrote:
> Yahya wrote:
>
---SNIP---

> My family include French, Puerto-Ricans, Dutch, Khmers and Americans (among
> others), good friends are Japanese and I like Bahasa Indonesia. I wanted sounds
> that all of them could equally well pronounce (not that they know about the
> language though) and I got this few, plus maybe [tS] and a final nasal.
Many Spanish speakers have a hard time distinguishing [tS] from [S] in
their speech.
>
> <<<
> Also, is there any particular reason for not allowing monosyllabic roots?
> >>>
> Yes. Tunu is self-segregating with compound/construct made without a tag. If you
> had ti, ku, tima and maku as valid root words, then timaku could mean either
> ti-maku or tima-ku or tima-kuXX or tima-ku XXXX or else. In the past I tried
> several solutions: a construct tag (like the -u- tag in Wolof), an article and
> you name it but for reasons too boring to explain here I came to the conclusion
> that no-tag construct and CVCV work relatively better.
I, with Jim Henry's help, independently came to the same conclusion
for Konya. Now, if self-segregation were not important in Tunu (or
Konya), then yes, a collection of monosyllabic roots would be better
than not having them.

>
> <<<
> If the 100 or so commonest roots had only one syllable, the language would be
> much more efficient.
> >>>
> I'm not much into "efficiency" :-) I've read this argument a lot in the past
> years on conlang websites together with the "noisy environment issue" but I
> cannot see how saving a second expressing a phrase is more efficient, for what
> purpose and compared to what.
Jim and I did some experiments with Konya's morphology to compare a
version that permitted about 65 monosyllabic words with a version that
permitted fewer than 10 . We found that there was almost no difference
in verbosity in the Babel text translation. We could certainly
_construct_ oddball phrases that would come out longer, but in regular
speech or writing things pretty much averaged out.

Also, in a limited phoneme set, having a lot of very short words may
reduce redundancy, depending on what roles they play and how they are
implemented.

---larry