THE GUY IN THE TANK
 
                                MCT - 2
                            13 December 1993
 
                 Copyright (C) 1993 Homer Wilson Smith
       Redistribution rights granted for non commercial purposes.
 
>HS>      Have you observed that observation is subjective and prone to
>  > flawed interpretation?
>
>Yes.  I've seen multiple observers have multiple and sometime
>contradictory interpretations, which suggests a potential for
>inaccuracy.  Have you observed that observation is not subjective?
 
     Your implication is that subjective implies flawed.
 
     This is true if subjective observations are being used to learn
about objective referents, but it is not true if the subjective
observation is ABOUT the subjective experience or color form itself.
 
     One can never know with certainty about anything other than what
one is looking at.  If one is looking at his subjective color form to
determine what is true in the outside world, then of course his
observations, being done on a via, are uncertain.
 
     Just because I SEE a Chris Schafmeister doesn't mean there IS a
Chris Schafmeister.  However the fact that I SEE something does mean
that I SEE something, and that is certain.
 
     Subjective observations are always certain when they are about
themselves, and such subjective observations are the ONLY kind of
observations that CAN be certain, because the observation is being done
on a direct contact basis and not on an implication from a via.
 
     People have discounted the importance of observations about color
form because they are so worried about the physical universe that what
the color form IMPLIES about the external world is more important to
them than what it implies about the internal world.
 
     You can't eat immortality.
 
     This is short sighted though.  I will admit we do need to eat, and
so uncertain knowledge gained through our subjective color form ABOUT
the hypothesized outside world is important, but it is also important to
know that our subjective color form ITSELF does not work on space-time
mechanical principles.  If it did, we couldn't see it.
 
     Machines are quite blind.  They can't SEE, they can only respond.
 
     RED COLOR FORM IS ITS OWN THING, machines can not SEE RED.
 
     They can't see anything.
 
>I am certain when I have total confidence that my analysis cannot be
>wrong because it is a flawless application of deductive logic arriving
>at certain conclusions from given definitions or premises.  That sort of
>certainty, however, is conditional on the accuracy of the premises; i.e.
>IF the premises are true, THEN the conclusion is true.  In reality, it
>is a very rare premise that has 0 percent chance of being wrong.
 
     Very rare premise?  That says that there ARE premises that can be
certain.  You can not say I am wrong about the existence of certainty
only because it is rare.  Either it exists or it don't.
 
     Anyhow we agree that the logic of a proof can be certain, but then
we have to ascertain the possible certainty of the premises, which
themselves, to be meaningful to reality, have to be actual observations
about reality.
 
     I agree that finding certainty in observations is hard, but only if
one is using vias to learn.  This means of course that, since we use our
conscious color form to learn about the supposed external universe, and
that we can not know about the external universe EXCEPT through our
conscious color forms, that there is no hope of a certain observation
about the external universe.
 
     Therefore all the certainty that can ever be would have to be about
our own personal conscious color forms and our self.
 
     It also follows that proof of the existence of the supposed
external universe is impossible.
 
     The supposed external 'objective' universe, is at best a theory
developed to model what we observe with certainty in our conscious color
forms.
 
>Many persons have felt and expressed certainty and been
>proven wrong.  You appear to be suggesting a definition of certainty
>that would necessitate agreement that these people weren't really
>certain when they said and thought they were.
 
     Yes.  If certainty is a gradient scale from 0 to 100 percent, and 0
percent means no chance of being right, and 100 percent means no chance
of being wrong, and 50 percent means total uncertainty, then what I mean
by perfect certainty is 100 percent certainty.
 
     It is clear that if a 100 percent certainty can be wrong, then by
default any 100 percent certainty is self discounting, therefore there
IS no 100 percent certainty.
 
     I claim that there are 100 percent certainties that can not be
proven wrong because they are clearly right, and so therefore any OTHER
certainty that was proven wrong, had to have been less than 100 percent
to start with.
 
     Claiming you are 100 percent certain does not mean that you are.
 
     If what you claim to be certain of later proves to be wrong, THAT
proves you weren't really 100 percent certain of it.
 
     Sloppy 'certainties' open the door to being wrong.
 
>Who, then, is capable of certainty?  How, then, may I distinguish
>between the certainty of one who is certain and the quasi-certainty of
>one who may be wrong?
 
     You can not distinguish anything about another with certainty, only
about yourself.  We are all just as blind to the nature of other people,
even their existence, as we are to the physical universe.
 
     All we can see and know for sure is our own color TV set, our own
conscious color forms.
 
     What other people claim, is merely what other people claim, that
must be for you, BY DEFAULT, an UN certainty.
 
>HS>      I am sorry but what you say does not carry much weight with
>  > me as you have already denied the surety of your own words.
>
>Would my words carry more weight if I claimed they were unchallengably
>accurate, correct, and precise?
 
     No just logical.
 
     "All generalities are false.
 
     This sentence is untrue.
 
     There are no absolutes.
 
     There certainly is no certainty."
 
are all logical contradictions in terms.  CERTAINLY.  They are
meaningless and self denying.  If people can live their lives with
circular self denying statements in their repertoire of 'what's true',
then all the more power to them.
 
     You seem to have said that logic is certain, if so, then being
certain you can't be certain of anything is being illogical.  Is it
useful to be illogical?
 
>I don't insist on defining what machines can have as consciousness, but
>I do see a strong similarity between bioconsciousness and computer
>consciousness.
 
     Yes, bioconsciousness is being used as a TV monitor to display data
about the external universe, much as a soldier in a tank would use a
real TV set to monitor what was going on outside the tank, if he had
external TV cameras attached to the top of the tank on the outside aimed
at the external universe.
 
>I do see a need to discriminate between them, but I am
>not at all convinced that the presence or lack of "certainty" is the
>characteristic by which one may reasonably assign a CU candidate to
>its appropriate category.
 
     I understand this, you have here demonstrated a complete grasp on
the challenge put before me.
 
     Just as a foretaste of things to come, notice that the guy in a
tank looking at a TV monitor in order to 'learn' about what is going on
in the outside world, can not really be certain of the data coming in,
because anything might be interfering with the cause lines coming in
from the outside world to his TV screen inside the tank.
 
     The image on the TV monitor is a SYMBOL for the external referent
that the symbol refers to and implies things about.  A REFERENT is the
object in the external universe that is being referred to or referenced
by the symbol.
 
     What he is trying to learn about is the external referent, HOWEVER
what he SEES is the symbol.
 
     This is called learning about a cause by looking at it's effect.
 
     The symbol is the effect.  The referent is the cause.
 
     The symbol is HOPEFULLY an accurate REPRESENTATION of the external
referent.  As long as the symbol is an actual effect of the external
cause, then perhaps an argument can be made that the state of the symbol
accurately implies, follows and parallels the state of the external
cause.  But there is no way to prove the integrity of the cause lines
coming into the symbol from the external referent.
 
     At any point in the cause lines, someone or something, even God,
could be feeding in false data to the data lines.  In a worst case
scenario the guy's data lines could be fed a recording of an external
universe that bore no resemblance to any actual universe present or
past.
 
     Even if he turned the external TV cameras on themselves and their
own data lines to the internal TV monitor, if the lines were corrupt,
anything they might report about their own integrity could be wrong, and
therefore could not be trusted.
 
     Therefore such a mechanical system of learning, which is dependent
on vias or internal symbols to symbolize external referents, HAS to be
uncertain by design.
 
     This is what every one is screaming at me about when they tell me
there can never be any certainty.  I KNOW THIS!
 
     It is true because looking at effects (the internal TV monitor)
does not prove cause (the implied external objects represented by the
images on the TV monitor).
 
     The establishment of the absolute CERTAIN UNCERTAINTY inherent in
learning via symbols or learning by looking at effects IS the machine
certainty theorem, namely a machine can not be certain of anything.
 
     The only way a machine can learn about anything including itself is
via 'symbols and referents', or effects implying causes, or changes in
states here implying prior changes in states there.
 
     There are two parts then to the MCT.
 
     The first is to prove WITH CERTAINTY that a machine can not be
certain of anything.  Considering how many of you are screaming at me
that certainty certainly doesn't exist, is dangerous, is immoral etc, I
would think this part would be easy.
 
     The second part is to prove that a Conscious Unit CAN be certain of
SOME things concerning its own internal TV set and color forms.
 
     This would imply that the INTERNAL nature by which a CU comes to
know of its own existence AND the nature of its own internal TV set, is
NOT a space-time mechanical process, but something else entirely.
 
     If the process by which a conscious unit came to know of its own
existence and the nature of its own internal TV set of color forms, were
a mechanical process, then the CU could only surmise that maybe it
existed, and it could not know for sure.
 
     It is CERTAINTY of self existence that implies that the conscious
unit is NOT learning about it self by looking at effects which merely
imply its own existence, but rather that it has direct contact WITH its
own existence.
 
     A conscious unit learning about itself is NOT the same thing as a
TV monitor hooked up to a TV camera which is able to look at the TV
monitor.  THAT may produce pretty infinite regressions, but it will
never produce a certain conclusion ON the TV monitor that the TV monitor
is there or even that the TV camera is there.  The line could be
corrupted or fed a false image anywhere along the causal path.  Wouldn't
you be able to see the spurious cable coming in providing the false
feed?  Not if the data fed in by the false feed didn't show it!
 
     Even if you have TWO TV cameras looking at each other hooked up to
the same TV monitor you still don't get certainty.  What ends up on the
TV monitor may or may not be representative of what is really going on
in reality, in the TV cameras, in the data lines, and in the TV monitor.
 
     There is no way you can get the TV monitor to display WITH
CERTAINTY what is being picked up by the TV cameras.
 
     There is NO way you can hook a TV monitor up to a TV camera in such
a way that the TV monitor ALWAYS CERTAINLY displays what the TV camera
is actually pointed at.
 
     Certainty is not possible in a mechanically linked learning system.
 
     It has nothing in particular to do with nuts and bolts, or wires
d and electrons.  It has to do with learning about causes VIA effects,
regardless of what they are made of.
 
     There is CERTAINLY no certainty in that game.
 
     Are we clear on this?
 
     Homer
 
-----------------------------------------------------------------------
Homer Wilson Smith           This file may be found at
homer@rahul.net              ftp.rahul.net/pub/homer/homer/mct2.memo
Posted to usenet newsgroup:  alt.clearing.technology