RE: Metarationality (was: JOIN: Alden Streeter)

From: Ben Goertzel (ben@goertzel.org)
Date: Sun Aug 25 2002 - 12:08:34 MDT


Gordon Worley wrote:
> Rationality is not limited to conscious thought. Thinking rationally in
> your unconscious thought is just more difficult because you have to
> actual think rationally, rather than watch your thought processes and
> correct them if they go astray. This is why further reason why trying
> to moderate your irrational thoughts does not work.

But your unconscious thought processes are precisely those processes that
"you" can't watch!!

"You" being the conscious, reflective aspect of your mind-brain-process

You cannot watch your unconscious processes and correct them if they go
astray. That is not how your brain is built. You can see the output of
your unconscious processes, only. I can accept that in some non-normal
states of consciousness you can go further than we usually do in terms of
observing unconscious processes -- but I'm skeptical about the extent to
which this is really possible...

And actually, trying to moderate my irrational thoughts (e.g. overoptimistic
time estimates, incorrect distance estimates that my visual cortex feeds me)
DOES work pretty well...

> > Musical improvisation is intrinsically nonrational. You're appealing to
> > reason and emotions at once in music, and reasoning about which note to
> > play
> > based on a rational analysis of what emotions and thoughts it can be
> > expected to evoke, is just not workable. Music theorists have been
> > straining in this direction for years with quite limited success.
>
> Maybe it's not possible, I don't know. I can only reproduce music I've
> heard; I'm not very good at producing new music.

Interestingly, I'm just the opposite. I am poor at reproducing music I've
heard, but quite good at creating new melodies, rhythms and song structures,
and improvising within/around them.... One of these years I'll translate a
few of my 100's of 4-track tapes to computer files and give you some smaples
;)

> At any rate, it may be that musical improvisation requires that you
> think faster than rational thought allows. If that's the case, the
> quality of the music will suffer compared to well thought out pieces.

This paragraph gave me a really good laugh!

I really don't think that the greatest works of music in history are the
ones that have been the most carefully and rationally thought out!!

Although I can imagine you and I may have different musical tastes; perhaps
it is the case that you prefer music with a more transparent rational
structure (say, within the universe of Bach, perhaps you prefer the
Brandenburg Concertos to the more improvisational Chromatic Fantasy and
Fugue, which I really love).

Yes, I agree, in theory a superintelligent super-rational being could
calculate, based on data about human psychology, a piece of music precisely
configured to affect human emotions in a certain way. This being could
create a piece of music that I'd love better than Voodoo Child, the
Chromatic Fantasy and Fugue, the Jupiter Symphony, Straight No Chaser, Take
Five and California Uber Alles all rolled up into one.

But this being would not be operating within the limitations of a human
brain.

Within the limitations of a human brain, explicitly rational thought (as
opposed to "rationality" in the broad sense of "whatever works") is just NOT
a reasonable way to improvise music OR to compose music.

Forget music, it is not a reasonable way to go about constructing a
mathematical proof.

Each step of a proof is logical, but the choice of which step to take is
intuitive, based on the unconscious integration of a vast amount of
information learned from reading other proofs, and reasoning in other
contexts. Furthermore, this intuitive judgment is inevitably
emotion-polluted, guided partly by the mathematician's nonrational (and
occasionally antirational) liking for one strategy of proof over another,
etc.

Again I'll say: we humans inevitably rely on our unconscious, which is
imperfectly "rational" in Eli's sense of "rationality as whatever works",
and which embodies a complex mixture of processes that are rational in the
narrow sense (of carrying out incremental inferences going from premises to
conclusions), and processes that are nonrational in the narrow sense.

> Your thought process in the original comment we're discussing may be the
> result of rationalization to prevent you from giving up the irrational
> thinking that makes you continue to try to reproduce assuming you are in
> some sense aware that if you stopped trying to reproduce you'd be able
> to get more work done towards creating the Singularity.

Well I got another good laugh out of this one ;)

Rest assured that I have no immediate plans to reproduce any further!

Three offspring is quite enough for my wife and myself...

> This is my own AI speculation, but:
>
> I think that in an AI thought that ve is getting feedback on will not
> cause a penalty in speed of thought like it does in humans. I can only
> direct my attention at one thought process at a time. If I try to focus
> on more than one thing at a time, it all falls apart and I get very
> confused. A good AI design would allow for parallel analysis of thought
> processes rather than preemptive. (Note: Gordon can converse with you
> and walk at the same time, since walking doesn't require much if any
> attention to do.)

Yes, I agree with you there. Perhaps we agree more about AI's than about
humans ;)

ben g



This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:40 MDT