Date: Tuesday, 24 May 1988, 14:24-EDT From: AIList Moderator Nick Papadakis Sender: nick@MIT-AI Reply-To: AIList@AI.AI.MIT.EDU Subject: AIList Digest V7 #4 To: ailist-outgoing@mc Status: R AIList Digest Wednesday, 25 May 1988 Volume 7 : Issue 4 Today's Topics: Philosophy - Free Will (First of at least three digests on this topic) ---------------------------------------------------------------------- Date: 7 May 88 17:51:22 GMT From: trwrb!aero!venera.isi.edu!smoliar@ucbvax.Berkeley.EDU (Stephen Smoliar) Subject: Re: Free Will & Self Awareness In article <31024@linus.UUCP> bwk@mbunix (Barry Kort) writes: > >It is not uncommon for a child to "spank" a machine which misbehaves. >But as adults, we know that when a machine fails to carry out its >function, it needs to be repaired or possibly redesigned. But we >do not punish the machine or incarcerate it. > >Why then, when a human engages in undesirable behavior, do we resort >to such unenlightened corrective measures as yelling, hitting, or >deprivation of life-affirming resources? > This discussion seems to be drifting from the issue of intelligence to that of aggression. I do not know whether or not such theses have gone out of fashion, but I still subscribe to the hypothesis that aggression is "natural" to almost all animal life forms, including man. Is your adult self so rational and mature that you have not so much as banged your fist on the table when your software does something which particularly frustrates you (or do you feel that adults also transcend frustration)? ------------------------------ Date: 8 May 88 09:52:18 GMT From: TAURUS.BITNET!shani@ucbvax.Berkeley.EDU Subject: Re: Free Will & Self-Awareness In article <2070013@otter.hple.hp.com>, cwp@otter.BITNET writes: > Did my value system exist before my conception? I doubt it. I learnt it, > through interaction with the environment and other people. Similarly, a > (possibly deterministic) program MAY be able to learn a value system, as > well as what an arch looks like. Simply because we have values, does not > mean we are free. Try to look at it this way: even assumeing that you did learn your values from other people (pearents, teachers, TV etc.) and did not make anything up, how did you decide, what to adupt from who? randomly? or by some pre-determent factors? doesn't it make values worthless, if you can always blame chance, heritage or some teachers in your school, for your decisions? There is one mistake (In my opinion, ofcourse), which repeat in many of the postings on this subject. You must make difference, between values as themselfes (Like which is your favorite color, whether you belive in god or not), from the practicing of your system-of-values (or alignment as I prefer to machine), is the given realm, you 'play' on, because 'real' things (in the manner of things that exist in the given realm), are the only things which have a common (more or less...) meaning to all of us. Now, if you will think of values, as theyre pure self, and not as theyre practice in realety, you will see that they are not 'real' in this manner, and therfore cannot be learnt or 'given'. Maybe one day we will be able to create machines that will have this uniqe human abielty to give a personal meaning to things... Infact, we can do it already, and could for thousands of years - we create new human beings. O.S. ----------------------------------------------------------------------------- I think that I think, and therfore I think that I am... ------------------------------ Date: 8 May 88 20:19:03 GMT From: oliveb!felix!dhw68k!feedme!doug@ames.arc.nasa.gov (Doug Salot) Subject: Re: Free Will & Self Awareness In article <1179@bingvaxu.cc.binghamton.edu> Cliff Joslyn writes: >In article <4543@super.upenn.edu> Lloyd Greenwald writes: >>This is a good point. It seems that some people are associating free will >>closely with randomness. > >Yes, I do so. I think this is a necessary definition. > >[good points about QM vs Classical vs Ignorance as views of Freedom deleted] I don't believe randomness (in the quantum mechanical sense) is important to a sense of free will. The illusion of free will is what's important, and when dealing with a computing machine (the brain) which makes state changes on the time order of milliseconds, it is simply impossible for that machine, self-aware or not, to view its state changes as deterministic when its state changes are based on finer-grained state changes that occur at or near the speed of light. It seems to me that it would be straight-forward to give a computer program with the ability to monitor its *behavior* without giving it the ability to find causal relations between its holistic state and its behavior. -- Doug Salot || doug@feedme.UUCP || {trwrb,hplabs}!felix!dhw68k!feedme!doug Feedme Microsystems:Inventors of the Snarf->Grok->Munge Development Cycle ------------------------------ Date: 9 May 88 01:32:40 GMT From: pdn!ard@uunet.uu.net (Akash Deshpande) Subject: Re: Free Will & Self-Awareness Consider a vending machine that for $.50 vends pepsi, coke or oj. After inserting the money you make a selection and get it. You are happy. Now consider a vending machine that has choices pepsi, coke and oj, but always gives you only oj for $.50. After inserting the money you make a selection, but irrespective of your selection you get oj. You may feel cheated. Thus, the willed result through exercise of freedom of choice may not be related to the actual result. The basic question of freewill is - "Is it enough to maintain an illusion of freedom of choice, or should the willed results be made effective?". The latter, I suppose. Further consider the first (good) vending machine. While it was being built, the designer really had 5 brands, but chose (freely, for whatever reasons) to vend only the three mentioned. As long as I (as a user of the vending machine) don't know of my unavailable choice space, I have the illusion of a full freedom of choice. This is where awareness comes in. Awareness expands my choices, or equivalently, lack of awareness creates an illusion of freewill (since you cannot choose that which you do not know of). Note that the designer of the vending machine controls the freewill of the user. Indian philosophy contends that awareness (=consciousness) is fundamental. Freewill always exists and is commensurate with awareness. But freewill is also an illusion when examined in the perspective of greater awareness. Akash -- Akash Deshpande Paradyne Corporation {gatech,rutgers,attmail}!codas!pdn!ard Mail stop LF-207 (813) 530-8307 o Largo, Florida 34649-2826 Like certain orifices, every one has opinions. I haven't seen my employer's! ------------------------------ Date: 9 May 88 02:36:41 GMT From: quintus!ok@unix.sri.com (Richard A. O'Keefe) Subject: Re: Free Will & Self Awareness In article <1182@bingvaxu.cc.binghamton.edu>, Cliff Joslyn writes: > Again, we're talking at different levels (probably a > subjective/objective problem). Let's try this: if you are free, that > means it is possible for you to make a choice. That is, you are free to > scrap your value system. At each choice you make, there is a small > chance that you will do something different, something unpredictible > given your past behavior/current value system. If, on the other hand, > you *always* adhere to that value system, then from my perspective, that > value system (as an *external cause*) is determining your behavior, and > you are not free. I'm not sure that there is any point in continuing this, our basic presuppositions seem to be so alien. Orthodox Christianity holds that - God is able to do anything that is doable - God is not constrained by anything other than His own nature - it is impossible for God to sin >From my perspective, such a God is maximally free. >From Chris Joslyn's perspective, such a God is minimally free. I think the problem lies in my disagreement with Joslyn's definition quoted above. He defines freedom as the ability to make choices, and seems to regard unmotivated (random) ``choices'' as the freest kind. I also take exception with his view that a value system is an external cause. My "value system" is as much a part of me as my memories. Or are my memories to be regarded as an external cause too? Note that behaving consistently according to a particular value system does not mean that said value system is immune from revision. There are some interesting logical problems involved: in order to move _rationally_ from one state of your value system to another, you have to believe that the new state is _better_, which is to say that your existing value system has to endorse the new one. I would regard "scrapping" one's value system as irrational (and it is not clear to me that it is possible far a human being to do it), and if being able to do it is freedom, that's not a kind of freedom worth having. My objection to randomness as a significant component of freedom is that a random act is not an act that **I** have *willed*. If I were to randomly put my fist through this screen, it wouldn't be _my_ act any more than Chris Joslyn's or J.S.Bach's. It is sheer good luck if a random act happens to be in accord with my wishes. Unfortunately, we have living proof that randomness as such is not freedom: consider the people who suffer from Tourette's syndrome. It might be objected that behaving in a way consistent with one's goals, beliefs, and wishes is too much like behaving in a way consistent with a program to be counted as freedom. It would be, if we were not capable of revising said goals and beliefs. To be free, you have to _check_ your beliefs. I propose as a tentative definition that a robot can be said to possess free will provided that its actions are in accord with its own mental models and provided that it has sufficient learning capacity to be able to almost wholly replace the mental models it is initially provided with. ------------------------------ Date: 9 May 88 08:50:08 GMT From: mcvax!ukc!strath-cs!glasgow!gilbert@uunet.uu.net (Gilbert Cockton) Subject: Re: Free Will & Self-Awareness In art. <5404@venera.isi.edu> smoliar@vaxa.isi.edu.UUCP (Stephen Smoliar) writes >a rule-based or other mechanical account of cognition and decision making >is at odds with the doctrine of free will which underpins most Western morality >Cockton should either prepare a brief substantiation or relegate it to the >cellar of outrageous vacuities crafted solely to attract attention! Hey, that vacuity's sparked off a really interesting debate, from which I'm learning a lot. Don't put it in the cellar yet. Apologies to anyone who doesn't like polemic, but I've always found it a great way of getting the ball rolling - I would use extreme statements as a classroom teacher to get discussion going, hope no-one's bothered by the transfer of this behaviour to the adult USENET. Anyway, the simplified, and thus inadeqaute argument is: machine intelligence => determinism determinism => lack of responsibility lack of responsibility => no moral blame no moral blame => do whatever your rulebase says. Now we could view morality as just another rulebase applied to output 1 of the decision-process, a pruning operator as it were. Unfortunately, all attempts to date to present a moral rule-base have failed, so the chances of morality being rule-based are slim. Note that in the study of humanity, we have few better tools now than we had in Classical times, so there are no good reasons for expecting major advances in our understanding of ourselves. Hence Skinner's dismay that while Physics had advanced much since classical times, Psychology has hardly advanced at all. Skinner accordingly stocked his lab with high-tech rats and pidgeons in an attempt to push back the frontiers of learning theory. At least you don't have to clean out the computer's cage :-) ------------------------------ Date: 9 May 88 10:52:56 GMT From: bwk@mitre-bedford.arpa (Barry W. Kort) Subject: Re: Free Will & Self Awareness I was doing fine reading Cliff's rejoinder to Lloyd's comments until I came to this part: >>We can't demonstrate true randomness in present day computers; >>the closest we can come (to my knowledge) is to generate a string >>of numbers which does not repeat itself. [Lloyd] >This is not possible in a von Neumann machine. [Cliff] I was under the impression that a simple recursion (or not-so-simple if one is a fan of Ramanujan) can emit the digits of pi (or e or SQRT(2)) and that such a string does not repeat itself. I think what Cliff meant is that a von Neumann machine cannot emit a string whose structure cannot be divined. If I wanted to give my von Neumann machine a *true* random number generator, I would connect it to an A/D converter driven by thermal noise (i.e. a toasty resister). --Barry Kort ------------------------------ Date: 9 May 88 13:11:04 GMT From: geb@cadre.dsl.pittsburgh.edu (Gordon E. Banks) Subject: Re: Free Will & Self Awareness In article <31024@linus.UUCP> bwk@mbunix (Barry Kort) writes: >It is not uncommon for a child to "spank" a machine which misbehaves. >But as adults, we know that when a machine fails to carry out its >function, it needs to be repaired or possibly redesigned. But we >do not punish the machine or incarcerate it. > >Why then, when a human engages in undesirable behavior, do we resort >to such unenlightened corrective measures as yelling, hitting, or >deprivation of life-affirming resources? > Pray tell, how do you repair, or redesign a human? Is "Clockwork Orange" the model we want to strive for? If you had a machine which was running amok, and you did not know how to repair it or redesign it, would not destroying it or isolating it from the objects of its aggression be a prudent course? Punishment can serve to "redesign" the human machine. If you have children, you will probably know this. Unfortunately, it doesn't work with everyone. ------------------------------ Date: 9 May 88 14:23:33 GMT From: sunybcs!nobody@rutgers.edu Reply-to: sunybcs!rapaport@rutgers.edu (William J. Rapaport) Subject: Re: Philosophy, free will Another good reference on free will is: Dennett, Daniel, _Elbow Room: The Varieties of Free Will Worth Having_ (MIT Press). ------------------------------ Date: 9 May 88 16:02:01 GMT From: bwk@mitre-bedford.arpa (Barry W. Kort) Subject: Re: Punishment of machines I was fascinated by John Nagle's rejoinder to my remarks about punishing a machine. John writes: > The concept of a machine which could be productively punished is >not totally unreasonable. It is, in fact, a useful property for some robots >to have. Robots that operate in the real world need mechanisms that implement >fear and pain to survive. Such machines will respond positively to punishment. > > I am working toward this end, am constructing suitable hardware and >software, and expect to demonstrate such robots in about a year. John's posting reminded me of the short story, "Soul of the Mark III Beast" which appears in _The Mind's I_. While I cannot dispute John's point that a game of engineered darwinism might produce a race of hardy robots, I must confess that I am troubled by the concept. Would not the survivors be liable to rising up against their creators in a titanic struggle for dominance and survival? Would we erect a new colliseum to enjoy the spectacle of intermachine warfare? Why am I both excited and horrified by the thought? --Barry Kort ------------------------------ Date: 9 May 88 16:28:40 GMT From: bwk@mitre-bedford.arpa (Barry W. Kort) Subject: Re: Free Will & Self-Awareness I was gratified to see Marty Brilliant's entry into the discussion. I certainly agree that an intelligent system must be able to evolve its knowledge over time, based information supplied partly by others, and partly by its own direct experience. Thomas Edison had a particularly rich and accurate knowledge base because he was a skeptic: he verified every piece of scientific knowledge before accepting it as part of his belief system. As a result, he was able to envision devices that actually worked when he built them. I think Minsky would agree that our values are derived partly from inheritance, partly from direct experience, and partly from internal reasoning. While the state of AI today may be closer to Competent Systems rather than Expert Systems, I see no reason why the field of AI cannot someday graduate to AW (Artificial Wisdom), in which an intelligent system not only knows something useful, it senses that which is worth knowing. --Barry Kort ------------------------------ Date: 9 May 88 17:41:09 GMT From: COYOTE.STANFORD.EDU!eyal@ucbvax.Berkeley.EDU (Eyal Mozes) Subject: Re: Free Will and Self-Awareness In article <726@taurus.BITNET>, shani@TAURUS.BITNET writes: >In article <402@aiva.ed.ac.uk>, jeff@aiva.BITNET writes: >> Is this an Ayn Rand point? It sure sounds like one. Note the use >> of `self-contradicting'. > >I bet you will not belive me, but I'm not sure I know who Ayn Rand is... I didn't see anything in Shani's message that looks like it's based on Ayn Rand (she certainly wasn't the only philosopher to oppose self-contradiction), but I do agree with Jeff that Ayn Rand's writings can shed light on the free will issue. To those who haven't heard of her, Ayn Rand was a novelist and a philosopher, and both her novels and her philosophy books are highly recommended. I think there are two main ways in which Ayn Rand was relevant to the free will issue: 1. Her basic approach, of basing philosophical theories on observation of facts rather than on assumptions about what the world should be like, is an approach that all those who discuss the issue of free will should learn. Specifically, we have to realize that free will is a fact, perceived directly by introspection, and it is therefore wrong to reject it just because we would like all natural processes to conform to the model of physics. 2. Ayn Rand has identified the exact nature of free will, in a way that is consistent with the facts, does not suffer from the philosophical problems of other free will theories, and demonstrates why free will is not connected to, and is actually incompatible with, randomness. Man's free will lies in the act of focusing his consciousness, in his choice to think or not and what to think about. This means that free will is consistent with having reasons that determine your thoughts and your actions, because, by your choice in focusing your consciousness, it is you who chose those reasons. Eyal Mozes BITNET: eyal%coyote@stanford ARPA: eyal@coyote.stanford.edu ------------------------------ Date: 9 May 88 18:22:01 GMT From: vu0112@bingvaxu.cc.binghamton.edu (Cliff Joslyn) Subject: Re: Free Will & Self Awareness In article <31337@linus.UUCP> bwk@mbunix (Kort) writes: >I was doing fine reading Cliff's rejoinder to Lloyd's comments until >I came to this part: > >>>We can't demonstrate true randomness in present day computers; >>>the closest we can come (to my knowledge) is to generate a string >>>of numbers which does not repeat itself. [Lloyd] > >>This is not possible in a von Neumann machine. [Cliff] > >I was under the impression that a simple recursion (or not-so-simple >if one is a fan of Ramanujan) can emit the digits of pi (or e or >SQRT(2)) and that such a string does not repeat itself. > >I think what Cliff meant is that a von Neumann machine cannot emit >a string whose structure cannot be divined. Hmm, I suppose you're right. I was thinking of your typical psuedo-random process whose cycle length was a function of the size of the seed. I forget the impact on the argument at this point: it seems it would rest on the epistemic grounds of determining a truly random string from a simply chaotic one. My impressions is that this is not always possible. Food for mailer Food for mailer Food for mailer Food for mailer Food for mailer -- O----------------------------------------------------------------------> | Cliff Joslyn, Cybernetician at Large | Systems Science, SUNY Binghamton, vu0112@bingvaxu.cc.binghamton.edu V All the world is biscuit shaped. . . ------------------------------ Date: 9 May 88 19:13:48 GMT From: mcvax!botter!klipper!biep@uunet.uu.net (J. A. "Biep" Durieux) Subject: Free Will, Quantum computers, determinism, randomness, modelling Warning: this is long! Question to Muslims at the end (but they should read the rest first before answering). In several newsgroups there are (independent) discussions on the nature of free will. It seems natural to merge these. Since these discussions mostly turn around physical, logical or mathematical notions, the natural place for this discussion seems to be sci.philosophy.tech (which newsgroup is devoted to technical philosophy, i.e. logic, philosophical mathematics (like intuitionism, formalism, etc.), philosophical physics (the time problem, interpretations of quantum mechanics, etc) and the like). Here follows a bit of the discussion in sci.math, plus some of my comments. - - - In article <5673@uwmcsd1.UUCP>, markh@csd4.milw.wisc.edu (Mark William Hopkins) writes: > This is probably the sharpest formulation of the widely held idea > that our free will may be ultimately derived from the Uncertainty Principle. > What's being argued here is that human beings are fundamentally random > (Turing Random) ... or less euphemistically, fundamentally insane. The > most succinct expression of these ideas is: > Human intelligence = Turing machine > + Quantum theoretical random signal generator - - - Some points (personal opinions): (1) The philosophical meaning of free will is mainly to form a basis for founding responsibility. To a lesser degree it is important with respect to the cluster of problems surrounding consciousness, experiencing, etc. It is in no way clear to me how "mechanical" explanations of the phenomenon can be of any help here. Is there any more reason for giving a victim a verdict of heavy duty instead of just life-long imprisonment (or easier still: death by destruction) because the victim came to its actions by a random decision process, or by trying in vain to model itself? Or does any such explanation give a hint as to the "why" of my sensation of existing? [I am not saying it doesn't: if you think it does, please explain!] [The same holds the other way too, of course: why take the trouble to give someone a "good old day" (especially if there is no motive of getting one yourself because of it).] (2) In answer to Drew McDermott's robot that can only predict the world by seeing itself as a black box that acts intentionally: that is exactly what there is to it: the robot has to model itself as an intentional being. And it *is* an intentional being. So is a glass with water and lots of salt. Any system with the right sort of feedback is intentional. The concept of intentionality is just a tool for solving simultaneous equations without getting into an infinite recursion. But what does intentionality have to do with free will? If I understand you well, you say that the robot would get the *illusion* of having free will because of having to model itself that way. Does that mean that until you "understood" the working of a magnet, since you had to model it intensionally, you were forced into the thought magnets had free will? Do you think matter has the free will to attract its like? (3) About explaining free will by introducing randomness: - I think those who say randomness is just as forcing as determinism: most of you have been talking about *external* randomness, and *external* determinism. Compatibilists see people make this error, and from that conclude that, since the objections to the theory don't hold, the theory is true after all. I think it isn't. Primo, the responsibility argument still holds, and secundo, the "explanation" is just shifting the problem, and therefore begging the (or another) question: - Randomness is a poorly (or even non-)understood phenomenon. Events may be random, but still have a probability distribution, which makes the thing even more complicated. Now free will might be an (unexplained in itself) explanation for randomness (the particle emerges where it wants to), at least as validly as the other way around. - Determinism, on the other hand, is not much better understood. It can easily be described, and so it *seems* better understood, but as far as I know, there is no valid explanation for why certain things happen always exactly the same way. (This is not even true in physics: QM.) The problem is as perplexing in logic: why are certain types of inference always valid? Even if I make one for the 100th time, again it is valid. - A widely accepted answer to both questions (on randomness and on determinism) is: because of free will (e.g. the free will of G-d to let the world be as it is). Free will, or at least will, is the "natural" explanation: primitive man supposes lots of spirits of all kinds to cause both regularities and irregularities. - Finally, several persons have argued that there is a division deterministic/random, and that there is nothing between. (Martin Gardner, in "The whys of a philosophical scrivener", argues that there *is*, or at least may be, something in between.) My impression is that these persons are looking for free will at the wrong place. To me, there is a spectrum (or dichotomy, rather), with probability (including determinism at both limits and randomness in between) at one end, and free will at the other. After all, determinism is just a special case of probability. A priori, it seems to me that it is easier to explain probability in terms of free will than free will in terms of probability. (4) I think free will and determinism don't have to clash, but for another reason than the ones I have read so far. Let me explain by way of an example: Suppose I am dreaming. Now I dream a world, and in that world someone proves that pi equals 7. What's more: he *correctly* proves so. He shows his proof to others, and the knowledge that pi is 7 becomes general knowledge in that world. That is possible. I will never be able to reproduce that proof as a correct proof in this world, but it that other world it is correct. Ask anybody in that dream-world of mine, and (s)he'll tell you that the laws of logic are as rigid as we think they are here. (By the way, what would the logical equivalent of the physical light speed be?) So that world is deterministic, but nevertherless my will made pi be 7. The only requirement is, that my will is *transcendent* to the world in question. Now anybody can imagine a world that is governed by *one* will, because we all have the dream example. But now imagine a world that is partially governed by lots of wills: every will calls the piece it governs "I". That may be somewhat harder to imagine, but to me it seems that that would produce a world in which the rules of logic and physics could be rigid and inviolable, but in which nevertheless free will could influence reality. I realise this too, is begging the question, for how would those free wills exist in their hyper-world, but one thing has changed: what first seemed impossible, now "only" seems ununderstandable (as we don't have any idea of what such a hyper-world would be like, and we only cannot imagine what sort of substratum they would have in stead of logic and causality). But let's stop now, before I introduce G-d, or even J-sus (ever been immanently present in your own dream-world?). One last question: how do Muslims explain the concept of responsibility, given the fact that they can't help what they do or don't do (kismet)? Or am I misunderstanding the concept of kismet? -- Biep. (biep@cs.vu.nl via mcvax) Some mazes (especially small ones) have no solutions. -- man 6 maze ------------------------------ Date: 9 May 88 20:27:16 GMT From: mcvax!ukc!its63b!aiva!tw@uunet.uu.net (Toby Walsh) Subject: Re: Free Will Drew McDermott's proposes a "cute" example of a robot R next to a bomb B, thinking about (thinking about (thinking about ..... its thinking) ....)); to avoid this infinite regress, he proposes "free will" = "ability to identify one's special status within one's model of the universe". This example immediately suggests to me the analogy with meta-level reasoning; reasoning about reasoning occurs at the meta-level, and reasoning about this meta-level reasoning at the meta-meta-level, .... To escape this infite regress of meta-meta-.... levels, we need to introduce the idea of (self-)reflection, where we reason about the meta^n-level in the meta^n-level. The notion of identifying one's special status within the model then becomes the analogous concept of naming between object- and meta-levels. But does this example/analogy tell us more about the annoying issue of free will ? No, I believe. It has much to say about consciousness but doesn't directly address what it is to have goals, desires, what it is to MAKE a decision when confronted with choice. Nevertheless, meta-level reasoning is an interesting model within which to formulate these concepts. ------------------------------------------------------------------------------- Toby Walsh JANET: T.Walsh@uk.ac.edinburgh Dept of AI ARPA: T.Walsh%uk.ac.edinburgh@nss.cs.ucl.ac.uk Edinburgh University Tel: (=44)-31-225-7774 ext 235 80 South Bridge, Edinburgh EH1 1HN ------------------------------ Date: 9 May 88 21:05:14 GMT From: bwk@mitre-bedford.arpa (Barry W. Kort) Subject: Re: Free Will & Self Awareness Perhaps it would help if I offered a straw proposal for invoking one's free will in a specific situation. Assume that I possess a value system which permits me to rank my personal preferences regarding the likely outcome of the courses of action open to me. Suppose, also, that I have a (possibly crude) estimate of your value system. If I were myopic (or maybe just stupid) I would choose my course of action to maximize my payoff without regard to you. But my knowledge of your value system creates an interesting opportunity for me. I can use my imagination to conceive a course of action which increases both of our utility functions. Free will empowers me to choose a Win-Win alternative. Without free will, I am predestined to engage in acts that hurt others. Since I disvalue hurting others, I thank God that I am endowed with free will. Is there a flaw in the above line of reasoning? If so, I would be grateful to someone for pointing it out to me. --Barry Kort ------------------------------ Date: 9 May 88 22:28:02 GMT From: turing.arc.nasa.gov!nienart@icarus.riacs.edu (john nienart) Subject: Re: Free Will and Self-Awareness In article <8805091739.AA27922@ucbvax.Berkeley.EDU> Eyal Mozes writes: > > Specifically, we have to realize that free will is a >fact, perceived directly by introspection, and it is therefore wrong to >reject it just because we would like all natural processes to conform >to the model of physics. What makes you so certain that _anything_ perceived be introspection is fact? Introspection provides me with the "fact" that there is, in fact, a "me" to do the introspecting, but there are a number of philosophies and religions (mostly of the Eastern variety) which insist that this is _not_ a fact, but rather simply an illusion we impose on ourselves, essentially through habit, and that through the proper discipline, we can train ourselves to note the absence of this self. After this process is complete (no claim is being made here as to personal success in this), introspection apparently confirms a hypothesis which directly contradicts our previous introspective belief. Which introspection is correct? > >Man's >free will lies in the act of focusing his consciousness, in his choice >to think or not and what to think about. This means that free will is >consistent with having reasons that determine your thoughts and your >actions, because, by your choice in focusing your consciousness, it is >you who chose those reasons. Maybe its just me, but I find rather frequently that I'm thinking about something that I'm _sure_ I'd rather not think about (or humming a trashy pop song I hate, etc.). It certainly feels at these times that I don't have complete control over that upon which my consciousness is focussed. So would you say that I _dont_ have free will (despite my introspective belief in it), or that the (introspectively motivated) "fact" of my lack of control is false? --John ------------------------------ Date: 9 May 88 23:55:11 GMT From: COYOTE.STANFORD.EDU!eyal@ucbvax.Berkeley.EDU (Eyal Mozes) Subject: Re: Free Will and Self-Awareness In article <796@hydra.riacs.edu>, john nienart writes: >What makes you so certain that _anything_ perceived be introspection is >fact? Introspection provides me with the "fact" that there is, in fact, a >"me" to do the introspecting, but there are a number of philosophies and >religions (mostly of the Eastern variety) which insist that this is _not_ a >fact, And that's exactly the point. Most philosophies, and all religions, make a lot of a priori assumptions about what the world should be like, some of them ridiculously contrary to fact. This approach must be rejected. >Maybe its just me, but I find rather frequently that I'm thinking about >something that I'm _sure_ I'd rather not think about (or humming a trashy >pop song I hate, etc.). It certainly feels at these times that I don't have >complete control over that upon which my consciousness is focussed. Are you saying that, at those times, you are making a deliberate, conscious effort to turn your thoughts to something else, and this effort fails? If so then, yes, it is just you; all the evidence I'm familiar with points to the fact that it's always possible for a human being to control his thoughts by a conscious effort. Eyal Mozes BITNET: eyal%coyote@stanford ARPA: eyal@coyote.stanford.edu ------------------------------ Date: 10 May 88 00:56:01 GMT From: trwrb!aero!venera.isi.edu!smoliar@bloom-beacon.MIT.EDU (Stephen Smoliar) Subject: Re: Free Will & Self-Awareness In article <1099@crete.cs.glasgow.ac.uk> gilbert@cs.glasgow.ac.uk (Gilbert Cockton) writes: >Research IS stopped for ethical reasons, especially in Medicine and >Psychology. I could envisage pressure on institutions to limit its AI >work to something which squares with our ideals of humanity. Just WHOSE ideals of humanity did you have in mind? I would not be surprised at the proposition that humanity, taken as a single collective, would not be able to agree on any single ideal; that would just strike me as another manifestation of human nature . . . a quality for which the study of artificial intelligence can develop great respect. Back when I was a callow freshman, I was taught to identify Socrates with the maxim, "Know thyself." As an individual who has always been concerned with matters of the mind, I can think of no higher ideal to which I might aspire than to know what it is that allows myself to know; and I regard artificial intelligence as an excellent scientific approach to the pursuit of this ideal . . . one which enables me to test flights of my imagination with concrete experimentation. Perhaps Gilbert Cockton would be kind enough to let us know what it is that he sees in artificial intelligence research that does not square with his personal ideals of humanity (whatever they may be); and I hope he does not confuse the sort of brute force engineering which goes into such endeavours as "smart weapons" with scientific research. >If the >US military were not using technology which was way beyond the >capability of its not-too-bright recruits, then most of the funding >would dry up anyway. With the Pentagon's reported concentration on >more short-term research, they may no longer be able to indulge their >belief in the possibility of intelligent weaponry. > Which do you want to debate, ethics or funding? The two have a long history of being immiscible. The attitude which our Department of Defense takes towards truly basic research is variable. Right now, times are hard (but then they don't appear to be prosperous in most of Europe either). We happen to have an administration that is more interested in guns than brains. We have survived such periods before, and I anticipate that we shall survive this one. However, a whole-scale condemnation of funding on grounds of ethics doesn't gain very much other than a lot of bad feeling. Fortunately, we have benefited from the fat years to the extent that the technology has become affordable to the extent that some of us can pursue more abstract studies of artificial intelligence with cheaper resources than ever before. Anyone who REALLY doesn't want to take what he feels is "dirty" money can function with much smaller grants from "cleaner" sources (or even, perhaps, work out his garage). > >The question is, do most people WANT a computational model of human >behaviour? Since when do "most people" determine the agenda of any scientific inquiry. Did "most people" care whether or not this planet was the center of the cosmos. The people who cared the most were navigators, and all they cared about was the accuracy of their charts. The people who seemed to care the most about Darwin were the ones who were most obsessed with the fundamental interpretation of scripture. This may offend sociological ideals; but science IS, by its very nature, an elite profession. A scientist who lets "most people" set the course of his inquiry might do well to consider the law or the church as an alternative profession. > Everyone is free to study what they want, but public >funding of a distasteful and dubious activity does not follow from >this freedom. And who is to be the arbiter of taste? I can imagine an ardent Zionist who might find the study of German history, literature, or music to be distasteful to an extreme. (I can remember when it was impossible to hear Richard Wagner or Richard Strauss in concert in Israel.) I can imagine political scientists who might find the study of hunter-gartherer cultures to be distasteful for having no impact on their personal view of the world. I have about as much respect for such tastes as I have for anyone who would classify artificial intelligence research as "a distasteful and dubious activity." > If funding were reduced, AI would join fringe areas such as >astrology, futorology and palmistry. Public funding and institutional support >for departments implies a legitimacy to AI which is not deserved. Of course, those "fringe areas" do not get their funding from the government. They get it through their own private enterprising, by which they convince those "most people" cited above to part with hard-earned dollars (after the taxman has taken his cut). Unfortunately, scientific research doesn't "sell" quite so well, because it is an arduous process with no quick delivery. Gilbert Cockton still has not made it clear, on scientific grounds at any rate, why AI does not deserve this so-called "legitimacy." In a subsequent article, he has attempted to fall back on what I like to call the what-a-piece-of-work-is-man line of argument. Unfortunately, this approach is emotional, not scientific. Why he has to draw upon emotions must only be because he cannot muster scientific arguments to make his case. Fortunately, those of us who wish to pursue a scientific research agenda need not be deterred by such thundering. We can devote our attention to the progress we make in our laboratories. ------------------------------ Date: Mon, 9 May 88 19:58 EST From: EBARNES%HAMPVMS.BITNET@MITVMA.MIT.EDU Subject: Free Will and Determinism OK Folks: There are a few developments that this arguement has undergone in the past hundred years that you should become familiar with. James Anderson was the first to point out the apparent paradox here on AIlist. To repeat what he said - 1) Determinism makes Free will impossible, because my action were determined from before I was born, and I therefore cannot have control over them; and 2) Free will is impossible without Determinism, because without strict determinism I do not have direct control over my actions (some random event could prevent my doing what I wanted to do). The first view was defended by Peter Van Inwagen in "An Essay on Free Will" and the second view was defended by Schopenhauer in "On the Freedom of the Will". The Arguement was settled by Dennett in his recent book "Elbow Room: Varieties of free will worth wanting". Dennett argues that when the idea of freedom is analyzed, there is no possible state of affairs that we could be refering to if we want to have complete control over our actions (i.e. - be able to do what we desire), and have our actions not be determined and so we must redefine what we mean by freedom. The arguement is involved but he uses analogies of what we mean by freedom in other senses, such as not being in prison to show that what we want when we want freedom is control over our lives and actions. What determinism claims is that our actions are determined by way of our desires being determined, not that our actions are controled in spite of our desires. His conclusion is that Schopenhauer was right, and that free will requires determinism rather than being prevented by it. Yes, this means that our desires are determined, but this is good. Our desires are determined by our enviornment, which is the best place for them to come from, since it is impossible for us to desire all of our desires (It would result in an infinite loop). The problem now is that Determinism is not in fact true, because of Quantum Mechanics. The probability 1 outcome of future events is a myth that has been thouroghly disproven. This raises the second point that has been missing in this discussion, a definition of random. I thank David Sher for pointing this out, and I offer the following clarification: there are two kinds of random - 1) Epistemic randomness means that we do not know what the outcome of an event will be; 2) Ontological randomness means that the outcome is not yet a fact of the matter (i.e. - it could turn out either way). Quantum Mechanics has shown that Ontological randomness exists (Read up on the Einstien, Pedalsky, Rosen [EPR] paradox). But the randomness occers primarily on the microscopic level, so that macroscopic events may still be almost totally determined. In conclusion, Free will is compatible with determinism, but determinism is not true. Free will is not only compatible with determinism, but dependant on it, so it appears that we have free will only to the extent that macroscopic events are determined. There are therefore degrees of freedom of the will. ------------------------------ End of AIList Digest ********************