Re: Real Conlangs Here, Made-to-Order!

En réponse à Tristan McLeay :


>Apparently only because we don't know how to tell a computer how to be
>creative and don't have the power to let it be creative.
Because we don't understand ourselves how creativity works. But whatever
the reason, my argument still holds.


>I find that unlikely.
Give me evidence for your claim.


>Again, the problem here is that we don't know how to be able to tell a
>computer how to do it.
Whatever, you only prove my point.


>But the human input is necessary for a human to create something, too.
Not always. We can discover how to create by ourselves, without human
input. Or else, how could I have learned how to conlang while I had never
followed a conlanging course?

>Do you imagine that a human who'd never come in contact with music would
>be able to create something that we'd consider music?
Yes. I truly think somebody who would have never come in contact with music
could have by himself the ideas of putting sounds together, and that if it
pleased him it would be music. After all, there are so many musical
traditions in the world that in our ears don't sound musical at all, so who
are we to judge whether something is music or not? Art is not only in the
eye (or ear) of the beholder.

>  They'd need to be
>taught! (or at the very least, influenced).
Only if they want to make music according to our Western rules. But why
should music be restrained to those rules?

>  Just because we don't have
>to be told absolutely everything because we come with a built-in
>grounding doesn't mean that a computer is less capable.
Yes, because a computer cannot come up with something it hasn't been told
about. A human being can, and it's something we don't know how it happens.
We may find out how it happens in the future, but we're currently not even
at the edge of the beginning of an understanding of our own creative
processes, so it won't happen any time soon.


>You needn't.
Yes you need. It's the essence of creation!

>  You program obedience (unnecessary I guess) and the idea of
>what a language is, or what music is, or what art is.
But that's not enough. The computer can only at best put together already
known parts in a random manner. Creation is something else.

>  Admittedly most
>people would reject it as art, but if something is sufficiently advanced
>that it can create something that's indistinguishable from art (barring
>knowledge of its source), what difference does it make?
The same as between the result of three open pots of paint falling on the
floor by accident and the result of a person painting the floor with the
same three pots of paint. You may like the first result better than the
second, but does it make it art? And if so, who's the artist?

>  (Personally, I
>don't know if I'd call the output art, at least, not until it's been
>appreciated by others (i.e. the thing itself isn't art, but the feelings
>it evokes bestows that title upon it) but the program which created it
>surely is.)
Of course, such a program would be a piece of art. But such a program would
have had a human programmer, who would be the artist. Whatever happens, the
human input is there.


>What exactly is a philosophical lang?
A language working with a small amount of roots and which builds words with
them in a taxonomic manner, with the added thing that it must be easily
parseable.

>  (And again, 'cannot'? What
>evidence have you got?)
I was replying to Andrew who claimed there was a program out there which
can create conlangs. I don't have to prove anything I say. He has to prove
me wrong by pointing out which program he's talking about.


>Neither of those have to do with creation; merely with comprehension.
Of course. But if they cannot even do that, how could they possibly create?

>And human translation is often dangerous; why should a computer be any
>better than that which programs it?
You're proving my point.

>  Furthermore, I contend that
>speech-to-text has orders of magnitude less to do with the ability to
>create (or even comprehend) than sight has to do with creating a
>painting or hearing has to do with composing music.
Prove it.

>  Illiterates are
>perfectly able to comprehend spoken English, yet they can't perform
>text-to-speech (which computers can), much less speech-to-text.
But they can be taught it, much better than any computer can currently
achieve it. The difference in potential is big.

>  Someone
>who's only ever heard one dialect spoken, or a few dialects spoken,
>often has difficulty understanding speakers of quite divergent dialects.
But they only need a little time of immersion to learn. A computer cannot
achieve anything like that currently. Even if the difference between human
beings and computers is only of degree, the difference is so big that it is
no different from a difference in quality.

>I know people who can play beautiful music yet can't read, much less
>write, a note (in the sense of one of those circular things, generally
>with a vertical line coming out of it, placed on a bunch of horizontal
>lines)!
And here again you're proving my point. Human beings don't need to be
taught anything to be able to create.


>Nor is denying undisproven things,
I didn't. I know there's no such program existing today, and you don't
argue by referring to the future, as the future is by essence uncertain. My
arguments only have to do with the present and our current knowledge.

Christophe Grandsire.
http://rainbow.conlang.free.fr
You need a straight mind to invent a twisted conlang.

Replies