Re: 3 possible views of "consciousness" +

From: <hpm.domain.name.hidden>
Date: Fri, 2 Feb 2001 20:12:15 EST

"Jesse Mazer" <lasermazer.domain.name.hidden> wrote:

> It's true that under the right mapping, the ticks of a clock can be seen as
> doing any "computation" you please, including a simulation of an intelligent
>

I replied:
> A clock isn't in fact a possible implementation for something that's
> supposed to remain interactive with its old world, because there's
> no way to give it input. A state machine with input [is a possible
implementation] ...


And I was wrong.

It IS possible to interpret a simple clock as an interactive program
(a Turing-test passing AI, say) by making the interpretation dependent
on the given input. [eg. if your input to the clock at clock reading
T was "Hi there buddy!" the clock showing a later time U might be
interpreted as an output of "Howdy yourself!", while if the input at T
was "Stop bothering me!" the clock reading U would be interpreted as
saying "Up yours, buddy!"] All possible input-history-contingent
outputs could be stored in a humongous lookup table encoding the
interpretation.

In this way of looking at the setup, the intelligence and
consciousness of the AI reside soley in the counting clock. The
Humongous table is simply a translation device that maps the clock's
representation to your representation.

This viewpoint also suggests that it would be mathematically possible
to transform the entire situation of clock, humongous table and
yourself into another where YOU are encoded as a clock, the AI is
transformed into some complicated mechanism much like your present self
that speaks english, and a different but equally humongous lookup
table translates your steady tick tock for the AI into the profound
and witty english dialog that you already understood it to be.

Or both you and the AI could be encoded as clocks, speaking the same
language, with no need for a humongous table to translate. A table
would only be necessary for outsiders who wanted to listen in on the
conversation.



Now, a more conventional view of the original "Clock + input-contingent
Humongous Table + You" setup above is that the clock is just a clock,
like those gigahertz ones pacing processor chips, and the humongous table
is an AI program, encoded in a maximally expanded form (and, of course
You are you).

So, do we need to fight about whether the AI is "really" in the clock
or in the table? I don't want to. It seems obvious to me that
intelligence and consciousness are truly Platonic things (like
cuteness) that can be seen by observers mapped onto many things (or
even onto nothing) in many different ways. Some mappings may have
more utility than others in particular circumstances, but none are
uniquely correct.


Is the cuteness of the little Copenhagen mermaid contained in the
shape of the brass it's made of, or in our brains that predispose us
to like both children and potential mates (and fish?), or in the
bloody evolution that selected those brains, or in the laws of physics
that framed that evolution, or what? Is the question meaningful at all?
Received on Fri Feb 02 2001 - 17:23:13 PST