Re: Intelligence increase

From: Samantha Atkins (
Date: Tue Oct 10 2000 - 15:46:43 MDT

Robin Hanson wrote:

> >
> >I think you have missed my point. My premise is that it is not the
> >inherent difficulty of communication that is the problem but the very
> >way the human brain is built and therefore learns and stores information
> >that is the difficulty. My its nature "learning" in the human brain (or
> >similarly build artificial brains) takes place by the associations
> >formed and stored. But this bundles up all learning with everything
> >else that is so cross-associated and makes both separation of the
> >learning and efficiently transferring it to another such mind quite
> >difficult.
> But this problem you describe of the excessive context dependence of
> things learned *is* exactly the inherent difficult of communication.
> All learning is naturally context dependent, and it takes substantial
> cognitive effort to strip out person-specific context in order to try
> to communicate what you've learned to other people. Your hypothesis is
> that some other architecture will do better at abstracting from
> person-specific context when learning, which would then directly make
> communication easier.

Fair enough. Although focusing on communication aspects can miss the
implications for learning and the nature of knowledge and knowing that
are also present. It brings up a question of whether there is any brain
architecture that we can conceive of that does not have these
limitations. Doesn't the very act of incorporating knowledge require
blending it with what one already knows and believes? It seems that to
optimize the acquisition of knowledge and its communication to other
minds would require a large degree of common knowledge infrastructure to
which new knowledge could be readily attached/incorporated. But this
would limit novel re-examinations and uses of the knowledge.

Do we get a tension between becoming part of a Collective and less
efficient learning and communicating but with higher inventiveness and

- samantha

This archive was generated by hypermail 2b30 : Mon May 28 2001 - 09:50:16 MDT