Re: Turing Test

From: Adam Boardman (
Date: Thu Jun 01 2000 - 22:19:24 BST

    [The following text is in the "iso-8859-1" character set]
    [Your display is set for the "US-ASCII" character set]
    [Some characters may be displayed incorrectly]

> > If you take a T3 passing robot then remove all external
> > non-computational components (sensors & limbs etc), then re-attach all
> > its inputs/outputs to other computer systems that simulate an
> > environment, identical to the real one.
> If you take a T3 passing robot and remove all external (and internal! --
> why only external?) noncomputational components, then you no longer have
> a T3 passing robot but an implemented symbol system. If you run the code
> in a virtual world, we're back to a virtual system in a virtual world --
> a place we've been many times in these discussions!

Hmmm, I think that you've miss-understood me, If you take a human and remove all
its non-brain components, give it a mechanical heart, mechanical limbs, cameras
for eyes etc. OK so this is sci-fi for now but were beginning to get the medical
abilities to do this kind of thing, not long before its possible. Cochlea
implants, neuron interfaces etc?

It doesn't die? Brain in a jar, wired up to it's basic sustenance, blood etc.

> It's all squiggles.

Yes but is that relevant, I've assumed a T3 is using squiggles for its inner
core mind container, the computational component. The rest is robotic actuators
and sensors.

> > Have we killed it? Its now entirely computational, could be simulated
> > inside another person.
> Would we kill one of us if we removed all of his noncomputational
> (implementation-independent symbolic) components? You bet we would.

Ummm, not shore if that's entirely true. Ok so a human if connected to a
simulation of a heart would die due to lack of oxygen etc, but computer
simulations of electricity are still electricity so no problems their?

> And whatever you might do with the (implementation-independent) code
> that was left over (whether you put it in a simulated world or a
> payroll programme), there would be no one left in there...

Fair enough.

> > If we take it that Searle is running this new computational system
> > that has had the benefit of grounding in the real world and is now
> > placed entirely inside his head within its virtual world. Then surely
> > we must have a T4 before we have a mind?
> Implementation-independent squiggles are not grounded; squiggles that
> are functional components of a T3 robot are grounded -- the whole system
> is grounded. Remove the noncomputational parts of the system and the
> grounding's gone too.

Even if you took the robotic appendages, which would be exactly computer
simulateable (this is T3 not T4) un-coupled them and replaced simulated ones
representing a simulated world. Why would they loose their grounding?

> T4 refers to a T3 robot that is also internally equivalent and
> indistinguishable from the innards of a real human being. The
> disembodied squiggles you are talking about here couldn't be further
> from that.

Yes my point was that we would need more than T3 the robotic appendages would
have to be like human ones i.e. T4.

I think that maybe the problem is that I'm imagining a T3 that is a central
computer thing running realistic neuron simulations attached to robotic motors,
sensors etc. Rather than using realistic neuron hardware stuff for the rest of
the body and muscles etc.


This archive was generated by hypermail 2b30 : Tue Feb 13 2001 - 16:36:29 GMT