sean@aipna.ed.ac.uk (Sean Matthews) (05/18/88)
Arguments for `free will' from itrospection boil down to `I think I have free will because I think I have free will' which is not even worthy of the label `tautology'. The (contradictory?) arguements go: I experience what I believe to be `free will' all the time since I have yet to experience myself doing something In spite of my desiring to do the opposite, even though I am not subject to any coercion.[1] I have demonstrated `free will' since I wanted to drink Coca Cola, but instead drank Pepsi Cola (That I also wanted to prove I had free will is overlooked). People often seem to confuse an impulse to perversity for `free will'. There are two problems: 1. does anyone have the slightest idea of what `free will' looks like. 2. perfect introspection is a logical impossibility[2], so anyone who invokes it allows their `mind' the sort of logical transcendance that Thomas Aquinas explicitly denied his god. This seems to be arrogating rather a lot. Se\'an Matthews Dept. of Artificial Intelligence JANET:sean%sin@uk.ac.ed.aiva University of Edinburgh ARPA: sean%uk.ac.ed.aiva@nss.cs.ucl.ac.uk 80 South Bridge UUCP: ...!mcvax!ukc!aiva!sean Edinburgh, EH1 1HN, Scotland [1] I don't know enough about psychiatric medicine to say whether it would be possible for a schitzophrenic to have an experence superficially akin to this---two aware `minds' in one brain, only one of which is exercising control---but the possibility raises interesting questions which can be, and are, dealt with in other places. [2] if a `mind' x had perfect introspection, it would need to contain within itself a complete model of itself, as well as the structures which are used to model x and to reason about it, but these would need to be inside x and therefore would be in the model. This is just a version of the barber paradox. PS. This has all probably been said before, but I didn't read it; if it has, I apologise for wasting your time.
gilbert@cs.glasgow.ac.uk (Gilbert Cockton) (05/19/88)
In article <38@aipna.ed.ac.uk> sean@uk.ac.ed.aipna.UUCP (Sean Matthews) writes: >2. perfect introspection is a logical impossibility[2] That doesn't make it impossible, just beyond comprehension through logic. Now, if you dive into Philosophy of Logic, you'll find that many other far more mundane phenomena aren't capturable within FOPC, hence all this work on non-standard logics. Slow progress here though. Does anyone seriously hold with certainty that logical impossibility is equivalent to commonsense notions of falsehood and impossibility? Don't waste time with similarities, such as Kantian analytic statements such as all "Bachelors are unmarried", as these rest completely on language and can thus often be translated into FOPC to show that bachelor(X) AND married(X) is logically impossible, untrue, really impossible, ... Any physicists around here use logic? -- Gilbert Cockton, Department of Computing Science, The University, Glasgow gilbert@uk.ac.glasgow.cs <europe>!ukc!glasgow!gilbert The proper object of the study of Mankind is Man, not machines
sean@aipna.ed.ac.uk (Sean Matthews) (05/21/88)
in article <1193@crete.cs.glasgow.ac.uk> gilbert@cs.glasgow.ac.uk (Gilbert Cockton) writes >In article <38@aipna.ed.ac.uk> {I} write >>2. perfect introspection is a logical impossibility[2] >That doesn't make it impossible, just beyond comprehension through logic. >Now, if you dive into Philosophy of Logic, you'll find that many other >far more mundane phenomena aren't capturable within FOPC, hence all >this work on non-standard logics. Slow progress here though. Mr Cockton is confusing one fairly restricted logic with the whole plethora I was referring to. There are logics specifically designed for dealing with problems of self reference (cf Craig Smory\'nski in Handbook of philosophical logic Vol2 `modal logic and self-reference') and they place very clear restrictions on what is possible in terms of self-referential systems and what is not; there has not been `Slow progress here'. > anyone seriously hold with certainty that logical impossibility >is equivalent to commonsense notions of falsehood and impossibility? I freely admit that I don't understand what he means here, unless he is making some sort of appeal to metaphysical concepts of truth apart from demonstrability and divorced from the concept of even analytic falsehood in any way. There are Western philosophers (even good ones) who invoke metaphysics to prove such things as `God exists' (I feel that God exists, therefore God exists---Rousseau), or even `God does not exist' (I feel that God does not exist, therefore God does not exist---Nietztche). Certainly facts may be `true' irrespective of whether we can `prove' them (the classical example is `this statement is not provable') though this again depends on what your idea of `truth' is. And there are different types of `truth' as he points out; any synthetic `truth' is always tentative, a black sheep can be discovered at any time, disposing of the previous ``truth'' (two sets of quotation marks) that all sheep were a sort of muddy light grey, whereas analytic `truth' is `true' for all time (cf Euclids `Elements'). But introspective `truth's are analytic, being purely mental; we have a finite base of knowledge (what we know about ourselves), and a set of rules that we apply to get new knowledge about the system; if the rules or the knowledge change then the deductions change, but the change is like changing Euclid's fifth postulate; the conclusions differ but the conclusions from the original system, though they may contradict the new conclusions, are still true, since they are prefixed with different axioms, and any system that posits perfect introspection is going to contain contradictions (cf Donald Perlis: `Meta in logic' in `Meta-level reasoning and reflection', North Holland for a quick survey). What happens in formal logic is that we take a subset of possible concepts (modus ponens, substitution, a few tautologies, some modal operators perhaps) and see what happens; if we can generate a contradiction in this (tiny) subset of accepted `truth's, then we can generate a contradiction in the set of all accepted `truth's using rational arguments this should lead us to reevaluate what we hold as axioms. These arguments could be carried out in natural language, the symbols, which perhaps seem to divorce the whole enterprise from reality, are not necessary, they only make things easier; after all Aristotle studied logic fairly successfully without them. Se\'an Matthews Dept. of Artificial Intelligence JANET:sean%sin@uk.ac.ed.aiva University of Edinburgh ARPA: sean%uk.ac.ed.aiva@nss.cs.ucl.ac.uk 80 South Bridge UUCP: ...!mcvax!ukc!aiva!sean Edinburgh, EH1 1HN, Scotland PS I apologise beforehand for any little liberties I may have taken with the finer points of particular philosophies mentioned above.
shani@TAURUS.BITNET (05/24/88)
In article <38@aipna.ed.ac.uk>, sean@aipna.BITNET writes: > 1. does anyone have the slightest idea of what `free will' looks like. It doesn't look like anything! It is just the knowing that you are a player character. I know I am! do you? O.S. ------------------------------------------------------------------------------- I am what I think, therefore I am what I am -------------------------------------------------------------------------------