On Tue, May 14, 2013 at 10:21 AM, Jacob Errington <nictytan@gmail.com> wrote:
On 12 May 2013 12:56, Bob LeChevalier, President and Founder - LLG
<lojbab@lojban.org> wrote:
More
broadly, really complex constructs are more easily expressed using
gi'ebo/gi'eke for right and left grouping in afterthought, and ga/gi for
forethought. Since these constructs connect multiple "bridi-tails"
(selbri plus trailing sumti) within one .i sentence, for maximal
flexibility, you have to but all the sumti including x1 after the
selbri. These are also described in that same chapter of CLL.
Furthermore, which I didn't
realize earlier, the gi'eke construct is in my opinion unintuitive
because it breaks the ordinary rules of ke..ke'e brackets:
{.i broda gi'eke brode ke'e brode gi'e brodu} doesn't
parse, even though {ke brode ke'e brode} does outside of a gihek. This
makes it impossible to achieve in a gihek the tanru parse {(broda brode)
(brodi (brodo brodu))} which would be done with a combination of
ke..ke'e and bo as such: {.i ke broda brode ke'e bo brodo bo brodu}
My suggestion would be to remoke gi'eke
brackets altogether because they cause unexpected behaviour and limit
the possibilities of our speech more than anything. They also have
presumably little to no usage.
.i mi'e la tsani mu'o
{bo} is hackish, but its meaning is at least determined by the grammar. Grammatically the worst thing about it is failure to elide certain terminators. For example, if {I TAG BO} were instead {I TAG BOhOI} or something, {.i fi'o broda fe'u bo} could be {.i fi'o broda bo'oi} with the {fe'u} elided. This isn't that bad, however, all things said; I wouldn't say it's worth having its own cmavo. Incidentally, why is it that the PEG can't make {.i fi'o broda bo mi brode} parse correctly, whereas it can fix things like JOI connecting LE-sumti without {ku}?
Evidently {ke} is actually even worse, since you can't coerce {GIhA KE SELBRI} to not make a gihe-kek. Why were these designed to be so awkwardly general purpose?