Newsgroups: comp.ai,comp.robotics,comp.ai.philosophy
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!agate!darkstar.UCSC.EDU!nic.scruz.net!earth.armory.com!rstevew
From: rstevew@armory.com (Richard Steven Walz)
Subject: Re: Minsky's new article
Organization: The Armory
Date: Mon, 5 Dec 1994 13:13:48 GMT
Message-ID: <D0CA3E.9yD@armory.com>
References: <3agf03$qi5@mp.cs.niu.edu> <3blk4d$nqo@jetsam.ee.pdx.edu> <3bmb16$8t9@mp.cs.niu.edu> <3bmj57$59f@jetsam.ee.pdx.edu>
Sender: news@armory.com (Usenet News)
Nntp-Posting-Host: deepthought.armory.com
Lines: 107
Xref: glinda.oz.cs.cmu.edu comp.ai:25716 comp.robotics:15976 comp.ai.philosophy:23151

In article <3bmj57$59f@jetsam.ee.pdx.edu>,
Marcus Daniels <marcus@ee.pdx.edu> wrote:
>rickert@cs.niu.edu (Neil Rickert) writes:
>
>>In <3blk4d$nqo@jetsam.ee.pdx.edu> marcus@ee.pdx.edu (Marcus Daniels) writes:
>
>>>Neil: yes or no, can a robot, in principle, perform scientific
>>>experiments?  Or is there some mental `cosmology' that the robot inherits
>>>but doesn't have implemented.
>
>>For a robot of the future, we might have available completely
>>autonomous robots which can act without being controlled.  We would
>>then be concerned as to whether we could trust the judgement of the
>>robot.  My guess is that we would not trust a robot's independent
>>scientific judgement unless we had already come to attribute to the
>>robot the same type of free will that we attribute to humans.
>
>Strictly speaking, `we' don't do that.
>
>You apparently missed the phrase `in principle', or perhaps
>I wasn't clear. 
>
>By robot, I mean something that is implemented using technology.
>By technology I mean machinery designed according to well understood
>models (if only probabilistic ones).
>
>I don't mean yet-to-be-discovered-quantum-links-to-universal-conciousness
>or any revolutionary physics.  Science, basically as we know it
>today -- only being relevant to the extent finding ways to make
>circuits smaller and faster.
>
>Do you think a robot implemented with, say, 
>massively parallel digital technology can have free will? 
>Don't quibble over cosmology -- I've defined mine.
>
>[And I take it you don't consider mathematics a branch of science?
>This university does, and as I recall Lenat's Automated
>Mathematician re-discovered some basics in number theory.
>Did it have free will?  Or were the results just a function
>of Lenat's free will.  I somehow doubt that would be his
>interpretation, and certainly not his goal!]
>
>The hard part is coming up ideas for experiments, not the execution
>of the experiment itself, or drawing conclusions.  Methods is
>pure algorithims + domain knowledge (+ devotion to the normal distribution
>and linear statistics. :-))  I don't think I'm going out on 
>limb to assert that methods is something a computer could do.
>
>So perhaps feed a pile of texts on methodology into a computer. 
>But, BTW, the courses I've taken on the topic were all about turning
>students into mindless droids -- not the other way around!
>
>One other question: if this free will thing is essentially tied to utility, 
>what if purely determinstic computers/robots start to produce more utility
>than humans do?  Free will, whatever one's definition, becomes a rather
>uninteresting notion, does it not?  Sorta subjective, you might say. :-)
------------------------------
Exactly the point. There are good reasons to believe that "free will" is
only that, and needn't be more than subjective, that is, it is generated by
any system that observes its own function and claims credit for that
function; "possession", if you will! It does NOT even NEED to mean that
such a thing as "free will" exists except as this delusion on the part of
one function of the device which interacts with the decision making process
to some slight degree but which is as determined as a digital watch! To be
"determined" is NOT to be some "mindless droid" but to be a part of the
universe and a normally functioning part of it which reacts to other
actions in the universe upon it or within influence range. Being determined
does not preclude being of the opinion that you exist as an entity and that
what you do is what you feel you "decided" to do, even if it was the only
way things could turn out in this situation. The two supposed opposites, or
antagonists, "free-will" and "determinism" are, in their correct places,
really only different referential explanations for the same thing, and are
NOT true opposites, as they can both exist without contradiction. There is
no magic to saying that the decisions of humans are the result of natural
laws acting in their case and that they claim and bear, in some manner,
credit or blame for them. If a person kills another, and it is what they
did out of complex but understandable cause and effect, then they are still
as much of a continuing threat to others, since we do not know of what they
might still be capable. But their actions have their determination in plain
old physical law, whatever that law may be or whether we understand it or
not! Is any one going to say that the atoms in someone's brain did not do
whatever it was they did the same as if they were in any other chemical
situation? No! But are you going to wish to trust them in the dark on your
street? No! "Will" may not be "free", even though we, being the agents of
awareness, will believe that we decided to do what we did. That's just what
the awareness mechanism does! Awareness is claiming existence by claiming
control which we simply do not necessarily have! Only one set of functions
claim this awareness, this control, and yet they claim the credit or blame
for all the other decision functions and their actions besides the imagining
that they do the controlling! These two approaches seem at odds with each
other because some fear that cause might be found for not holding people
responsible for their actions, which we won't do. And yet we know that a
person who is raised in horrible surroundings will turn out dangerous!
Cause and effect rule these "free-will" decisions just as they do
everything else. It is not that we are automatons, for clearly we are not.
But our decisions are only more complex, even if by far, than an automaton,
and the automaton does not have a recursive function that envisages itself
and seeks its own welfare and those it cares for. we do, and as "caused" as
it is, it IS still us, and that doesn't change. Free-will and determinism
are NOT mutually exclusive, but to be properly understood, it must be
understood that determined does NOT mean null and void, automatic, as we
think of that, or unthinking. And it is, after all, determinism that
produces awareness and the awareness function which calls itself "free to
choose", a legal term between humans, not a declaration of a suspension of
cause and effect!!!
-Steve Walz   rstevew@armory.com

