Re: machine translation

[HACKER G N <c9709244@ALINGA.NEWCASTLE.EDU.AU>]

On Thu, 23 Oct 1997, Mark Vines wrote:
> coi doi lobypli
>
> Someone mentioned that the Cyc system would be superior to
> Lojban for machine translation between natural languages.
> I hope we can discuss this further.

To keep this discussion objective, let me include a snippet from the Cyc
site itself, about CycL:

"CycL, the CYC representation language, is a large and extraordinarily
flexible knowledge representation language. It is essentially an
augmentation of first-order predicate calculus (FOPC), with extensions to
handle equality, default reasoning, skolemization, and some second-order
features. (For example, quantification over predicates is allowed in some
circumstances, and complete assertions can appear as intensional
components of other assertions.) CycL uses a form of circumscription,
includes the unique names assumption, and can make use of the closed world
assumption where appropriate."

Now, who can tell me whether Lojban is capable of the same power and
flexibility as this computer language?

> > the Cyc ontology, which my friend
> Liane Acker helped to assemble;

Cool. I'd love to get involved with Cyc myself. I think they're doing some
very exciting things with AI.
>
> Also, the little I've seen of CycL did NOT seem to approach
> cultural neutrality as closely as Lojban does.

Be more specific. In what ways did it not? Remember, it's just a computer
language; it's NOT mean for communication between humans.
>
> Finally, someone said that CycL can express things that
> Lojban cannot easily express, such as quantification of
> whole predicates (bridi? or bridi-tails?) & something which
> included the string "skolemic".

That's all included in the above snippet.
>
> Sorry, my copy of McCawley is in pieces; it fell apart while
> I was trying to learn the difference between intensions &
> extensions.  What does "skolemic" mean?

I'd like to know this myself.
>
> Also, I thought we could quantify practically anything in
> Lojban.  What gives?

Well, as Jorge already pointed out, Lojban is not well-suited to
quantifying selbri, for one thing. What are some things that Lojban can
quantify besides, say, sumti and to a lesser extent indicators?

Geoff