Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!purdue!decwrl!shelby!unix!chips2.sri.com!ellis
From: ellis@chips2.sri.com (Michael Ellis)
Newsgroups: comp.ai
Subject: Re: Free will and responsibility.
Keywords: Behaviorism, materialism, dogma, science
Message-ID: <149@unix.SRI.COM>
Date: 12 Jun 89 10:01:17 GMT
References: <10333@ihlpb.ATT.COM> <3850@uhccux.uhcc.hawaii.edu> <52019@linus.UUCP> <533@orawest.UUCP> <2586@rice-chex.ai.mit.edu> <386@edai.ed.ac.uk>
Sender: news@unix.SRI.COM
Reply-To: ellis@chips2.sri.com.UUCP (Michael Ellis)
Organization: SRI International
Lines: 64

> Chris Malcolm

>>It seems to me that the crucial
>>issue is not whether or not a particular action is physically determined or
>>not, but rather whether or not the agent can be said to have "controlled"
>>the action, i.e. whether the cause of the action arose from the beliefs
>>and desires of the actor, or from unavoidable environmental constraints.

>But here you are impaled on the dualist's dilemma: IF the mind is a
>function of the brain (mind-states correspond to brain-states), then
>since the brain, as a machine, is causally determined (brain-state n+1
>can be deduced from brain-state n plus sense-data), it follows that
>mind-states are causally determined by physical events, and are just
>another (perhaps more convenient) way of describing them.

    Lots of "ifs" there. First, and least important, is that the brain
    isn't causally determined because of QM + Chaos theory (either one
    in themselves is not sufficient): Brain state n+1 is provably not
    "determined" by brain state n plus sense data. Also notice you
    neglected to mention "output" or "control data".

>The only
>(obvious) way to restore the causal primacy and efficacy of mental
>events (beliefs, intentions, etc.) is to disconnect mind-states from
>brain-states, creating the mental/physical dualism. That solves the
>causal primacy of will, but at the price of raising the new problem of
>how on earth the non-physical mind affects the brain.

    The "(only) obvious way" is hardly the only way. Screw metaphysical
    dualism. How about anomalous monism and token identity? And wouldn't the
    failure of reductionism imply autonomous levels of description? There
    are so many compelling ways out I am bewildered to see the 19th century
    still haunting our computer science community.
    
    My desires have "causal efficacy" provided they really cause my actions.
    And my mind retains "causal efficacy" provided my present beliefs and
    desires really do control my future beliefs and desires in the way and
    to the extent that I want them to. They can be "made out of" blue cheese
    for all I care.

    You might as well be saying that the bottom level description falsifies
    the top level description. Sort of the arguing over whether "People kill
    people" or "guns kill people" or "atoms kill people". They're all
    valid accounts, none contradicting the other.

    I notice that people have no problem believing their perceptions are
    affected by the world, but then balk at believing their volitions 
    in turn affect the world. Why? 

>But your suggestion is on the right lines: we get into this puzzling
>position by fleeing thoughtlessly from the idea that free-will and
>predictability are contradictory, when even a little reflection shows
>that to be incompatible with our commonplace speech. The hard question
>is what exactly do we mean by "could have been otherwise"?

    Which takes us into modal and intensional logics and straight
    back into that ferocious aristotelian mentalistic jungle again.

    "Could not have been otherwise" as far as it is relevant to free will
    means approximately "any agent would have been forced to act identically
    in all salient respects". This is a different form of free will than the
    others, but is in no way easier to analyze.

-michael