Definition of intelligence
John Chelen
intelligence? In particular, I have been thinking about the minimal
functional elements that must be present in an entity for intelligence
to be present. For example, base requirements might include 1.
sensory capability; 2. information storage; 3. information
processing/decision making that can provide a novel output, 4.
manipulation of the environment based on processing.
There might also be required some measure of self-reference. For
example, decision-making might be framed as a determination that
action will increase the proability of ongoing existence of the
entity. I would grant that the self-reference need not be concious
but could be provided an heuristic framework that takes into account
the requirements of the entity.
Might we develop a minimal definition of intelligence simliar to the
bit, perhaps called the "it"? We might then develop a metric for an
"it" and then compute the relative intelligence of any entity as a
multiple of "its". Additionally, we might develop a metric for the
individual sub-functions of intelligence and compute a combinatorial
score based on the sum of scope of the sub-functions.
No doubt this has been worked on already but I haven't been able to
find any decent references to such work.
Wolf Kirchmeir
>Can anyone provide a utilitarian or operational definition of
>intelligence?
Whatever is tested by intelligence tests.
IOW: As the accountant said when asked "How much is 2 + 2?", "What do you
want it to be?"
Best Wishes,
Wolf Kirchmeir
Blind River, Ontario
..................................................................
You can observe a lot by watching
(Yogi Berra, Phil. Em.)
..................................................................
Burkart
Burkart
Cat Halogen
"Burkart" <burkartNO...@web.de> wrote in message
news:apltps$2rjlr$1...@ID-155818.news.dfncis.de...
> For me, the ability of *learning* is most important for
intelligence.
> Burkart
>
*unlearn* - to get rid of fixed ingrained thinking habits.
This is the other side of the coin.
Pat The Cat
Lionel Bonnetier
> Another important feature of intelligence is the ability to
> *unlearn* - to get rid of fixed ingrained thinking habits.
> This is the other side of the coin.
And the most amazing is that we still somehow
remember the things we unlearn. "As a child I
thought that [this-and-that]".
Glen M. Sizemore
I've come to like you, Wolf (not that it matters), but could you possibly
have come up with a more sophomoric response?
"Wolf Kirchmeir" <wwol...@sympatico.ca> wrote in message
news:jbysxveflzcngvpbp...@news1.sympatico.ca...
Burkart
news:apmhho$9kl$1...@mordred.cc.jyu.fi...
Learning can imply "unlearning". Learning is more than collecting knowledge.
More important is the connection between the knowledge. These connections
may become unimportant, so that it has less relevance for the whole system,
so that something seems to be forgotten after a time.
Burkart
Lionel Bonnetier
> Learning can imply "unlearning". Learning is more than collecting knowledge.
> More important is the connection between the knowledge. These connections
> may become unimportant, so that it has less relevance for the whole system,
> so that something seems to be forgotten after a time.
Agreed. Moreover the complexity of the human mind
allows to determine this relevance in more than one
way, contrarily to simple neural network simulations.
A number of factors, such as being put in an old
context, can temporarily revive a discarded pathway.
This is even more stunning when we try to understand
how someone thinks, and we can, in the best case,
really think like that person, even if we totally
disagree with them. I'm scratching my head trying to
imagine how to model this behavior practically.
It's possible that this ability is used to analyse
our own mind, scan our former behaviors, realize our
own mental evolution. Has anyone already tried to
sketch cognitive models of this, maybe using mental
mode switchings, or duplicate mind parts?
Wolf Kirchmeir
>Whatever is tested by intelligence tests.
>
>I've come to like you, Wolf (not that it matters), but could you possibly
>have come up with a more sophomoric response?
Well, if tried really, really hard... :-)
IMO, an _operational_ definition of intelligence is exactly what I said
above. You make your definition, then (try to?) devise a test that will
demonstrate the non/existence of the defined entity. That is, you (try to)
define what the system under test must do to demonstrate that it is
intelligent. Besides, _in practice_, IQ tests have in fact tested what the
customer wanted.
That's how all _operational_ definitions work. As in A Hole Is To Dig. (Not
sophomoric, since it was coined by a child. :-) )
As for my conceptual definition of "intelligence," I'm still working on it.
The more I learn about the abilities of living beings (and our picayune
attempts to imitate them, such as computers), the more convinced I am we are
probably asking the wrong questions. Sofar, I've come to these conclusions:
a) intelligence does not require consciousness. b) a good definition os I
must include a/a set of measurables. (One must be able to say "more/less"
intelligent of the system one is testing.) c) it does not require that the
system under test be biological. I am tentatively including problem solving
and the ability to modify one's behaviour as part of it.
That's as far as I've gotten, and you'll note that in essence it's what
intelligence is not, even where phrased positively.
At any rate, the other responses in this thread illustrate that pretty well
everyone has unstated assumptions about what intelligence is.
Part of my skepticism on this issue arises from my almost 37 years as a
teacher, which convinced me that IQ tests didn't tell me very much about what
a student could do, apart from being able to score well on IQ tests, that is.
IME, tests generally tell you only how well a student will do on that type of
test. One of the most arresting examples of this is the traditional spelling
test: most students can learn do very well on these - and then misspell as
usual in their written work. Spelling tests don't measure what really counts:
proof reading skills.
But that's getting OT.
Burkart
"Wolf Kirchmeir" <wwol...@sympatico.ca> schrieb im Newsbeitrag
news:jbysxveflzcngvpbp...@news1.sympatico.ca...
> On Tue, 29 Oct 2002 23:20:14 -0500, Glen M. Sizemore wrote:
>
> >Whatever is tested by intelligence tests.
OK.
> As for my conceptual definition of "intelligence," I'm still working on
it.
> The more I learn about the abilities of living beings (and our picayune
> attempts to imitate them, such as computers), the more convinced I am we
are
> probably asking the wrong questions.
But what is the correct question?
> Sofar, I've come to these conclusions:
> a) intelligence does not require consciousness.
The main problem is perhaps that "intelligence" is such a fuzzy notion.
For example: Can subconsciousness be (seen as) intelligent?
> b) a good definition os I
> must include a/a set of measurables. (One must be able to say "more/less"
> intelligent of the system one is testing.)
Hm, isn't this still conceptual or already operational?
> c) it does not require that the
> system under test be biological. I am tentatively including problem
solving
> and the ability to modify one's behaviour as part of it.
Exactly. The ability to modify = learning is very important. For me, it is
the main aspect of intelligence.
Burkart
Glen M. Sizemore
> >I've come to like you, Wolf (not that it matters), but could you possibly
>have come up with a more sophomoric response?
WK: Well, if tried really, really hard... :-)
GS: Just ask Lester....you're bound to acquire a response that is more
sophomoric. But, of course, you'd then have several sophomoric responses,
each "defined analytically in terms of each other."
WK: IMO, an _operational_ definition of intelligence is exactly what I said
above.
GS: God Almighty I love you Wolf. This is EXACTLY what I wanted you to say.
WK: You make your definition, then (try to?) devise a test that will
demonstrate the non/existence of the defined entity.
GS: But see....this is absolute crap. The "existence" of the "entity" is a
priori....that's exactly the problem. To ask "what is the thing that we call
'intelligence?'" is already to give it "thingness." Can't you see that? No,
I suppose not.....but, damn it, I still love you!
WK: That is, you (try to) define what the system under test must do to
demonstrate that it is intelligent. Besides, _in practice_, IQ tests have in
fact tested what the customer wanted.
GS: Once again, Wolf, I must express my overwhelming love for you. And I
think you love me too! Why else would you play right into my hands? Uncle
Sam was the original "customer," and what "he" wanted was a way to label
people as "mentally defective" so he could either "send them back to where
they came from," brutalize and sterilize them - but they did get what they
wanted.
WK: That's how all _operational_ definitions work.
GS: Yes, operational definitions which were, at first, suggested as ways to
purge specious concepts from physics, have been utilized to make terms
"scientifically acceptable" that would otherwise be rejected. What we have
to show for them is mainstream psychology.
WK: As in A Hole Is To Dig. (Not sophomoric, since it was coined by a child.
:-) ) As for my conceptual definition of "intelligence," I'm still working
on it. The more I learn about the abilities of living beings (and our
picayune attempts to imitate them, such as computers), the more convinced I
am we are probably asking the wrong questions.
GS: Well, as I have said numerous times, you don't "get" to invent
definitions. If we are trying to look at what is "meant by a colloquial
term" which is typically the way things go in psychology, then "inventing a
definition" serves only to infuse the subject matter with one's assumptions.
WK: Sofar, I've come to these conclusions: a) intelligence does not require
consciousness. b) a good definition os I must include a/a set of
measurables. (One must be able to say "more/less" intelligent of the system
one is testing.) c) it does not require that the system under test be
biological. I am tentatively including problem solving and the ability to
modify one's behaviour as part of it.
GS: Or maybe one should say that the term is not scientifically useful.
Behavior can already be measured - it is repeatable, has temporal locus, and
has temporal extent. These properties can be assigned dimensional
quantities.
WK: That's as far as I've gotten, and you'll note that in essence it's what
intelligence is not, even where phrased positively.
GS: Don't know what you are saying here.
WK: At any rate, the other responses in this thread illustrate that pretty
well everyone has unstated assumptions about what intelligence is.
GS: But some of us don't think that "it" is anything. One thing is for sure,
it is surely stupid to say we behave intelligently because we "possess
intelligence."
WK: Part of my skepticism on this issue arises from my almost 37 years as a
teacher, which convinced me that IQ tests didn't tell me very much about
what a student could do, apart from being able to score well on IQ tests,
that is.
GS: See? You can be reasonable.
WK: IME, tests generally tell you only how well a student will do on that
type of test. One of the most arresting examples of this is the traditional
spelling test: most students can learn do very well on these - and then
misspell as usual in their written work. Spelling tests don't measure what
really counts: proof reading skills.
GS: See? You can be reasonable.
WK: But that's getting OT.
GS: Hardly.