On 21 Mar 2015, at 04:00, meekerdb wrote:
On 3/20/2015 6:32 PM, Bruno Marchal wrote:
On 20 Mar 2015, at 20:51, John Clark wrote:
On Thu, Mar 19, 2015 Bruno Marchal <[email protected]> wrote:
>> Then the Turing Test works for consciousness and not just for
intelligence.
> It can give an idea that some entity is conscious, if long
enough, and with entity similar to you. It can not work for
intelligence.
Good God!
Let us hope.
In the final analysis passing the Turing Test just means being
observed doing intelligent things,
The Turing test measure the ability in the domain "passing a Turing
tests".
But I think you're imagining the Turing test in a restricted way.
The Turing test can measure learning ability too. You've asked Liz
and me to prove some theorems in modal logic. I don't know about
Liz, but I learned what I know about modal logic from you on this
list. So does my proving those theorems mean I only exhibited
competence, or was it intelligence?
I am not sure.
Do you think it is impossible to distinguish intelligence from
competence?
Competence is domain dependent, and can be evaluated, with exams,
tests, etc.
I could measure your competence in modal logic by given problems.
But with the definition I gave, it is harder to evaluate non trivial
intelligence, and it is very simple to evaluate trivial stupidity. If
a machine says "I am consistent", or "I am inconsistent" I can suspect
it to be stupid. If a pebble says nothing (which is often the case
with pebble) there is a sense to say it is "trivially" intelligent.
You can measure stupidity, by counting the number of times someone
says the same stupid things. Like Albert Camus said: stupidity insists.
But for machine of similar complexity, absence of saying stupidity is
a sort of measure of local/current intelligence, but it cannot be used
to declare the machine intelligent. The machine might repeat the
stupidities some other days.
Intelligence is almost only an attitude. Some people becomes
intelligent after a shock, or some events. Competence usually needs
some long amount of work.
We can say that plant are competent in disseminating their seeds, but
we would not say that it is intelligent behavior.
Intelligence is closer to notion like consciousness, wiseness,
enlightening, and is more a property of the "heart" (in his antic
poetical sense), than a property of a brain owner.
Intelligence is almost what is there by default, and can be destroyed
by mental wounds or psychological problem. Intelligence is affective,
emotional and a probable definition would refer to the first person.
Competence is more 3p and testable.
If you're going to suppose that the world can be emulated by digital
computation I think you'll have to accept that digital
communication, "the domain of passing a Turing test", is sufficient
for any learning, intelligence, and competence.
I agree. This is close to the non zombie principle. But this does not
make "intelligence" and competence equivalent. usually intelligence
accelerates learning and the possible development of competence, but
high competence can make people sleepy and have negative feedback on
intelligence.
Intelligence is well captured by an axiomatic similar to the main
properties of consistency. When true, you will not asserted it about
yourself. But of course such axiomatic works also for happiness and
other "protagorean virtue". In fact, intelligence is the mother of all
protagorean virtue, that you can teach by practising them, and thus by
example behavior, but that you can hardly teach by words, except for
the trivial one by default, for enough complex machinery if you want
avoid the intelligence of the pebble.
Somehow: competence is what make you able to do things and say
correctly "I know" when you know, like "I know how to solve that type
of problem in modal logic. Intelligence is the ability to stay mute
when you can't do the thing.
Competence is well studied by theoretical artificial intelligence
(Putnam, Gold, Blum, Case and Smith, Oherson, Zeugmann, etc.)
Intelligence, up to now, is better studied by the G and G* logics, and
their intensional variant.
There might be different characterizations. I suspect that
intelligence is also an ability to listen to the others (which makes
you saying less stupidities). Stupidity is almost a sort of autism or
neurosis making an entity unable to listen to the others. It does not
exist by default, but develops by affective problems, lack of self-
esteem, etc. may be intelligence is a sort of ability to love people
different from you, and stupidity is an ability to hate people
different from you.
Do you get the idea? I tend to attribute intelligence to all
(universal, Löbian) machines by default, but competence is when some
machine can solve some problems in some class of problems.
There is no universal competence in practice, but intelligence is a
universal notion a priori.
Bruno
Brent
--
You received this message because you are subscribed to the Google
Groups "Everything List" group.
To unsubscribe from this group and stop receiving emails from it,
send an email to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/everything-list.
For more options, visit https://groups.google.com/d/optout.
http://iridia.ulb.ac.be/~marchal/
--
You received this message because you are subscribed to the Google Groups
"Everything List" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/everything-list.
For more options, visit https://groups.google.com/d/optout.
