At 10:52 PM 12/18/01 -0500, Rob Speer wrote:
On Tue, Dec 18, 2001 at 02:29:18AM -0500, Bob LeChevalier (lojbab) wrote: > >I've probably expressed this before, but I think that the separation of > >tense selma'o is going to be the first thing to go when the baseline > >ends - which would for the most part bring the language more in line > >with usage anyway, and with the goal to remove restrictions on thought. > > If it goes, then we return to the TLI Loglan state where any agglomeration > of tense words is a tense, whether or not it could possibly mean> anything. We tried to err on the side of over-specification - at one point> PA was several selma'o so that we could rule out invalid "number" strings> and thereby make other number strings more meaningful. We could not devise> an unambiguous grammar for numbers in multiple selma'o so we abandoned the > effort. In contrast, we have a grammar for tenses that works and is only > clumsy when you push it in new directions. It is more flexible and > powerful than any natlang tense system, and it was the best we could do. Okay. I'm willing to accept that this system is more powerful, but is there any document which explains how this power should be _used_? The Book rarely uses more than two tenses at a time, so it doesn't give much of an indication of how they interact.
Not sure I understand. If you know how any two tenses interact, then you should be able to predict larger conglomerations (using the default left grouping if it matters). Now, TLI Loglan documentation used only single tense examples, which made it really hard to figure out how the system was supposed to work.
Also, I know from other situations like (as you point out) numbers, and UI, that meaningless conglomerations of words can be grammatical. It's not the grammar's job to restrict semantics.
It is the grammar's job to as much as possible 1) rule out ungrammatical strings and 2) define the internal structure of all strings. If the semantics is always clear using strict left grouping then there is no problem, but this is rarely the case with tense strings.
The real reason for the current tense system is that originally pc and I designed a system for Lojban that was quite rigidly structured, but was ugly. We redesigned it around 10 years ago but kept the basic concepts of the old system, which included distinct time and space tenses.
> The different selma'o within the tense system DO reflect distinctions in > meaning. Sometimes unrelated words ended up in one selma'o (as cu'e and > nau) because we could not think of a difference in how they would be> used. But a CAhA is not a CA or a TAhE, and whatever a CAhA+NAI might mean> is not determined by parallels with those other selma'o. Hence it is a > separate selma'o How does this justify VA and ZI being separate?
They are separate for historical reasons dating back to the TLI Loglan era. The story of the evolution of the tense system would be quite long, if I could even manage to tell it straight. The short version is that "pu/ca/ba" and "vi/va/vu" used to be parallel constructions for time and space tenses respectively, and each could take "zi/za/zu" as a distance, with somewhat different semantics in each case. In the abovementioned redesign, we recognized that pu/ca/ba and vi/va/vu were not in fact parallel and zi was thus being used to mean two unrelated things that overlapped other selma'o (the interval sizes), and using parallel constructions with zi made no sense. We then dropped the space usage of zi, but retained vi/va/vu for space tenses for backwards compatibility.
Given the major changes we had made in the tense system, we would never consider eliminating selma'o that were merely apparently similar in grammar, because we could not be sure that there would be no further redesign of the tense grammar. Backwards compatibility was always more important than elegance in the grammar description.
lojbab -- lojbab lojbab@lojban.org Bob LeChevalier, President, The Logical Language Group, Inc. 2904 Beau Lane, Fairfax VA 22031-1303 USA 703-385-0273 Artificial language Loglan/Lojban: http://www.lojban.org