From: Ben Goertzel (ben@goertzel.org)
Date: Sun Dec 08 2002 - 12:17:44 MST
> We don't know how to do these things. Sure, we might currently be
> able to identify the particular nerve-bundles that operate a person's
> right arm and monitor them, but there is a lot more to controlling
> your right arm than just the nerve bundles and the signals that get
> passed along them. A *lot* more. Creating a right-arm-signaling
> model based on this monitoring is, IMO, plausible. But that model
> would bear little relationship to the existing wetware functionality,
> in its myriad connections to everything else. The brain/mind is not
> solely described by its I/O. There is a lot going on in my brain
> that never produces more than a raised eye-brow or two.
>
> It's not worth the time to read.
>
> Michael Roy Ames
You're probably right about the URL -- it was a bit too frivolous to forward
to the list -- sorry!
However, "The brain/mind is not solely described by its I/O" does bring up
an interesting point.
The guy's idea is not that the brain/mind is solely described by its I/O,
it's rather that *each brain subsystem* is solely defined by its I/O. The
I/O of a brain subsystem, in most cases, involves signals that originate &
terminate within the brain and never see the outside world or the rest of
the body....
If we recorded the I/O of each brain subsystem over a period of time, and
built a model of this I/O, then we *might* be able to piece these together
to make a digital replica of the brain. [The idea of reimplanting
digitally-replicated subsystems into the functioning brain opens up a whole
other can of worms, obviously.] The time period would have to be long
enough to ensure that we were capturing the learning/adaptation behavior of
the brain subsystem, not just its steady-state behavior at a given point in
time.
Of course, the viability of doing this brain-subsystem recording in a
nondestructive way is another issue entirely...
-- Ben
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:41 MDT