Re: Lojban ML: Syllogism and sophism

[Lionel Bonnetier <leo@EASYNET.FR>]

Xorxes wrote:

> Lionel:
> >That's right. I was trying to reach the most concise and complete physical
> >description of the linguistic contradiction. Something like a common-sense
> >AI description mandatorily telling apart a whole thing from its components,
> >and an absolute quantity change from a relative one.
> >
> >   cheese.whole.quant.abs.up -> cheese.holes.quant.abs.up
>
> But that is not a logical conclusion. Only a very likely one based
> on the more likely way in which the cheese.whole.quant.abs
> could increase. If you increase it by filling in the holes then it is
> not the case that the quantity of holes will go up.

It wasn't so clear, but I meant a proportional increase of all the
components of <*.whole> by its increasing. Filling in the holes would be
another operator such as <cheese.hole.replace: cheese.matter>


> >   cheese.holes.quant.rel.up -> cheese.matter.quant.rel.down
>
> Same here. If you increase the quantity of holes by adding cheese
> so that each hole is split into two holes, then the cheese.matter.quant.rel
> goes up.

Here I meant the total volume of holes, not the number of holes, but it was
imprecise. A real AI description language should be even more finicky.


> >Natural languages don't provide any short expression for such precisions,
> >not necessarily because they are obvious, but mostly because they are
> >fuzzily understood, which makes language manipulations so easy.
>
> I don't think Lojban goes beyond natural languages in this respect.

I wasn't expecting Lojban to be an expert-system language, just trying to
find a very short expression for disambiguation.


> Articles are mandatory in Lojban. So you may have something like:

Thank you. I still have to study a lot. And not by browsing the grammar like
a frenzied kangaroo :-/ I regret there is not more discussions in Lojban on
the list -- I'm learning English mostly on the net, since I'm rather lazy at
swallowing the so-called 'fine' litterature. I need real time spinning.


> In any case, the whole point of the aphorism is that it's a seeming
> paradox. If we resolve the paradox it loses its whole raison d'etre,
> so a good translation has to maintain the apparent paradox.

Indeed, but my primary goal was to mount a flashing evidence of the paradox
knot. There are so many cases where Lojban is much more precise, for
instance with the conjunctions, that I wouldn't find it a weakness not to be
able to translate a pun or a sophism adequately.

I can imagine the very first articulated-speech languages were rich of many
jokes and mistakes that would be totally impossible nowadays. Imagine that
several words such as 'climb', 'fly off', 'rise', 'lift up', 'sky' were one
word whose meanings were told apart through the context. Let's take another
word meaning 'fire', 'light', 'bright', 'sun'.

Then the tribe around the night fire is talking about how big was the
mammoth they killed yesterday, and how hard it was to get it. Then a spark
jumps off the embers. A child amused by it shouts something like 'light
up!', and another child sleeping there wakes up and wonders why he can't see
the morning sun...

Now try to translate the mistake into some modern language... Except through
wild poetry, I guess it's impossible.

If Lojban is to be the next step for human or AI languages, it will have to
add many explanatory footnotes when translating today's little jokes. And
I'm sure it will create its own puns and logical puzzles, out of reach for
the current languages.

Cheers!

   Lionel Bonnetier <leo@easynet.fr>   Ph: +33 478 601 862