Date: Mon 9 May 1988 00:07-PDT From: AIList Moderator Kenneth Laws Reply-To: AIList@KL.SRI.COM Us-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList V6 #96 - Philosophy To: AIList@KL.SRI.COM Status: RO AIList Digest Monday, 9 May 1988 Volume 6 : Issue 96 Today's Topics: Philosophy - Free Will & Randomness ---------------------------------------------------------------------- Date: 3 May 88 19:23:33 GMT From: ulysses!sfmag!sfsup!glg@ucbvax.Berkeley.EDU (G.Gleason) Subject: Re: The future of AI [was Re: Time Magazine -- Computers of the Future] In article <1053@crete.cs.glasgow.ac.uk> gilbert@cs.glasgow.ac.uk (Gilbert Cockton) writes: >The main objection to AI is when it claims to approach our humanity. > It cannot. That's a pretty strong claim to make without backing it up. I'm not saying that I disagree with you, and I also object to all the hype which makes this claim for current AI, or anything that is likely to come out of current research. I'm also not saying your claim is wrong, only that it is unjustified; there is more to learn before we can really say. There are new ideas in biology that build upon "systems theory," and probably can be tied in with the physical symbol systems theory (I hope I got that right) that suggest that information or "linguistic interaction" is fundamental to living organisms. In the May/June issue of "The Sciences," I found an article called "The Life of Meaning." It was in a regular column (The Information Age). I won't summarize the whole article, but it does present some compelling examples, and arguments for extending the language of language to talking about cellular mechanisms. One is how cyclic AMP acts as an internal message in E. coli. When an E. coli lands in an environment without food, cyclic AMP binds to the DNA, and switches the cell over to a "motion" program. Cyclic AMP in this role has all the attributes of a symbolic (or linguistic) message: the choice of symbol is arbitrary, and the "meaning" is context dependant. This becomes even more clear with the example of human adrenaline response in liver cells. The hormone binds to sites on the outside of the cell which causes an internal message to be generated, which just happens to be cyclic AMP. The cell responds to the cyclic AMP (not by a DNA based mechanism as in E. coli) by producing more glucose. The composition of the message has nothing to do with the trigger or the response, it is symbolic. So, how is this relevant to the original discussion. I don't see any fundamental difference between exchanging chemical messages or electronic ones. Although this does not imply that configurations of electronic and electromechanical components that we would call "alive" are possible or that it is possible to design and build one, it doesn't rule it out, and more importantly it suggests a fundamental similarity between living organisms and "information processors." The only difference is how they arise. Possibly an important difference, but we have no way to prove this now. Gerry Gleason ------------------------------ Date: 2 May 88 16:11:11 GMT From: clyde!mtunx!whuts!homxb!houdi!marty1@bellcore.com (M.BRILLIANT) Subject: Re: Free Will & Self-Awareness In article <717@taurus.BITNET>, shani@TAURUS.BITNET writes: > In article <30502@linus.UUCP>, bwk@mbunix.BITNET writes: > > > > I would like to learn how to imbue silicon with consciousness, > > awareness, free will, and a value system. > > .... free will and value systems - such things cannot > be 'given', they simply exist..... > .... You can write 'moral' programs, even in BASIC, if you want, > because they will have YOUR value system.... It has been suggested that intelligence cannot be "given" to a machine either. That is, an "expert system" using only expertise "given" to it out of the experience of human experts is not exhibiting full "artificial intelligence." BWK suggested "artificial awareness" as a complement to "artificial intelligence," but apparently that is not enough. You need artificial learning. My value system was not "given" to me, nor was my professional expertise; both were learned. At its ultimate, AI research is really devoted to the invention of artificial learning. For full artificial intelligence, the machine must derive its expertise from its own experience. For full artificial awareness, the machine must derive its values from its own experience. Not much different. Achieve artificial learning, and you will get both. I hate to rehash the old "Turing test" again, but a machine cannot pass for human longer than a few hours, or days at most, unless it has the capacity for "agonizing reappraisal": the ability to "reevalueate its basic assumptions." That would be learning as humans do it. M. B. Brilliant Marty AT&T-BL HO 3D-520 (201)-949-1858 Holmdel, NJ 07733 ihnp4!houdi!marty1 Disclaimer: Opinions stated herein are mine unless and until my employer explicitly claims them; then I lose all rights to them. ------------------------------ Date: 6 May 88 13:31:25 GMT From: eniac.seas.upenn.edu!lloyd@super.upenn.edu (Lloyd Greenwald) Subject: Re: Free Will & Self Awareness In article <10942@sunybcs.UUCP> sher@wolf.UUCP (David Sher) writes: >It seems that people are discussing free will and determinism by >trying to distinguish true free will from random behavior. There is a >fundamental problem with this topic. Randomness itself is not well >understood. If you could get a good definition of random behavior you >may have a better handle on free will. > This is a good point. It seems that some people are associating free will closely with randomness. To me true randomness is as difficult to comprehend as true free will. 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. Can anyone give us a better view of randomness then this? I've heard some mention of true randomness at the quantum level. Does anyone have any information on this? Given that current theories of free will tie it so closely to randomness, it seems necessary to get a handle on true randomness. Lloyd Greenwald lloyd@eniac.seas.upenn.edu ------------------------------ Date: 4 May 88 07:19:52 GMT From: TAURUS.BITNET!shani@ucbvax.Berkeley.EDU Subject: Re: AIList V6 #86 - Philosophy In article <1579@pt.cs.cmu.edu>, yamauchi@speech2.cs.cmu.edu.BITNET writes: > Actually, I have read The Society of Mind, where Minsky writes: >[A quote of Minsky] > I would agree with this. In fact, unless one believes in some form of > supernatural forces, this seems like the only rational alternative. You are touching the very core of the problem. The point in which, this 'only random and determination exist' is getting into problems is the question of responseability i.e., if everything is pre-determened or random, how can you asume responsebility to what you are doing? and if responsability does not exist, the whole matter of free will and value system has no content, so, if free will and value system can be given to a machine, it is meaningless, and if it has meaning, it is depended on a third, irational factor (Free will), which cannot (Menwhile?) be given to a machine... O.S. ------------------------------ Date: 6 May 88 09:20:18 GMT From: otter!cwp@hplabs.hp.com (Chris Preist) Subject: Re: Free Will & Self-Awareness R. O'Keefe replies to me... > > Did my value system exist before my conception? I doubt it. >This is rather like asking whether some specific number existed before >anyone calculated. Numbers and value systems are symbolic/abstract >things, not material objects. I have often wondered what philosophy >would have been like if it had arisen in a Polynesian community rather >than an Indo-European one (in Polynesian languages, numbers are _verbs_). >---------- Oh no! Looks like my intuitionist sympathies are creeping out!!! Seriously though, there IS a big difference between numbers and value systems - Empirical evidence for this is given by the fact that (most of) society agrees on a number system, but the debate about which value system is 'correct' leads to factionism, terrorism, war, etc etc. Value systems are unique to each individual, a product of his/her nature and nurture. While they may be able to be expressed abstractly, this does not mean they 'exist' in abstraction (Intuitionist aside: The same could be said of numbers). They are obviously not material objects, but this does not mean they have Platonic Ideal existance. We are not imbued with them at birth, but aquire them. This aquisition is perfectly compatible with determinism. So what does this mean for AI? Earlier, in my reply to O.S., I was arguing that our SUBJECTIVE experience of freedom is perfectly compatible with our existance within a deterministic system, hence AI is not necessarily fruitless. You have drawn me out on another metaphysical point - I believe that our intelligence (rather than our capacity for intelligence), our value systems, and also our 'semantics' stem from our existance within the world, rather than our essential nature. Sensation and experience are primary. The brain is a product of the spinal chord, rather than vice-versa. For this reason, I believe that the goals of strong AI can only be accomplished by techniques which accept the importance of sensation. Connectionism is the only such technique I know of at the moment. Chris Preist ------------------------------ Date: 6 May 88 17:06:34 GMT From: vu0112@bingvaxu.cc.binghamton.edu (Cliff Joslyn) Subject: Re: Free Will & Self Awareness In article <4543@super.upenn.edu> lloyd@eniac.seas.upenn.edu.UUCP (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. Consider the concept of Freedom in the most general sense. It is opposed by the concept of Determinism. We can say of anything, either it is absolutely determined (it will always do one thing and only one thing), or it is somwhat free (sometimes it will do one thing, other times another). This is so whether we talk of molecules in a box or the actions of an organism. >To me true randomness is as difficult to comprehend >as true free will. I agree. That's because both psychological free will and randomness are cases of my general sense of Freedom. Freedom is a very difficult things to understand. >We can't demonstrate true randomness in present day >computers; von Neumann machines are highly Determined systems. They posess so little Freedom that it is essentially null. This is what they have been designed to do. However, it is easy to demonstrate that von Neumann machines are slightly free. Consider the distribution of bit errors in a cpu or RAM, or of read errors on a disk drive. These are random events. To that extent the computer is Free. This is not especially useful or interesting Freedom, nevertheless it is there. >the closest we can come (to my knowledge) is to generate a string >of numbers which does not repeat itself. This is not possible in a von Neumann machine. >I've heard some mention of true randomness at the >quantum level. See recent (last two years) articles in _Scientific American_ concerning hidden variables theories in QM. As I described in a brevious article, we can think of two cases of randomness, subjective and objective. Subjective randomness is usually equated with ignorance. For example, in Newtonian physics if I had sufficient information about initial conditions I could predict the roll of a die. Objective randomness is your "true", or irreducible, or inherent, or unavoidable randomness. There has been a great debate as to whether quantum uncertainty was subjective or objective. The subjectivists espoused "hidden variables" theories (i.e.: there are determining factors going on, we just don't know them yet, the variables are hidden). These theories can be tested. Recently they have been shown to be false. >Given that current >theories of free will tie it so closely to randomness, it seems necessary to >get a handle on true randomness. In my mind, the critical thing to understand about Freedom is that Freedom is always relative; Determinism is always absolute. What I mean is that when we talk about something being Free, we can always talk about degrees of freedom. A six sided die is more Free than a four sided, a twelve than a six. Or consider a probability distribution: it's Freedom is generally measured by it's entropy, which takes values in the interval [ 0, inf ). In order for the distribution to reach the infinite limit, it must be uniformly distributed over the whole positive real interval. This distribution is not well defined. In other words, we know what it means for something to be completely Determined. I submit that it is not possible for somethings to be completely Free. Absolute Freedom is an infinite limit; absolute determinism is a zero limit. This is obviously true in the realm of human affairs as well. It is easy for me to completely determine your actions: put you in a Skinner box, or straight jacket, or just kill you. And while I espouse Free Will, I do so only in this relative way. In no way can you tell me that you are absolutely free: drug delusions, dreams, illness, epilepsy, all kinds of physical/biological factors come into play which somewhat limit the Freedom of your mind. -- O----------------------------------------------------------------------> | Cliff Joslyn, Cybernetician at Large | Systems Science, SUNY Binghamton, vu0112@bingvaxu.cc.binghamton.edu V All the world is biscuit shaped. . . ------------------------------ Date: 5 May 88 09:37:54 GMT From: mcvax!ukc!strath-cs!glasgow!gilbert@uunet.uu.net (Gilbert Cockton) Subject: Re: Free Will & Self-Awareness In article <5100@pucc.Princeton.EDU> RLWALD@pucc.Princeton.EDU writes: > Are you saying that AI research will be stopped because when it ignores >free will, it is immoral and people will take action against it? 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. 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. > When has a 'doctrine' (which, by the way, is nothing of the sort with >respect to free will) any such relationship to what is possible? >From this, I can only conclude that your understanding of social processes is non-existent. Behaviour is not classified as deviant because it is impossible, but because it is undesirable. I know of NO rational theory of society, so arguments that a computational model of human behaviour MAY be possible are utterly irrelevant. This is a typical academic argument, and as you know, academics have a limited influence on society. The question is, do most people WANT a computational model of human behaviour? In these days of near 100% public funding of research, this is no longer a question that can be ducked in the name of academic freedom. Everyone is free to study what they want, but public funding of a distasteful and dubious activity does not follow from this freedom. 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. ------------------------------ Date: 5 May 88 19:35:39 GMT From: mcvax!ukc!its63b!aiva!jeff@uunet.uu.net (Jeff Dalton) Subject: Re: Free Will & Self-Awareness In article <717@taurus.BITNET> writes: - you are underastimating yourself as a free-willing - creature, and second, your request is self-contradicting ans shows - litle understanding of matters, like free will and value systems - - such things cannot be 'given', they simply exist. Is this an Ayn Rand point? It sure sounds like one. Note the use of `self-contradicting'. - You can write 'moral' programs, even in BASIC, if you want, - because they will have YOUR value system.... It is hard to see how this makes any sense whatsoever. ------------------------------ Date: 5 May 88 20:19:25 GMT From: mcvax!ukc!its63b!aiva!jeff@uunet.uu.net (Jeff Dalton) Subject: Re: Free Will & Self Awareness In article <770@onion.cs.reading.ac.uk> jadwa@henry.cs.reading.ac.uk (James Anderson) writes: >If the world is deterministic I am denied free will because I can >not determine the outcome of a decision. On the other hand, if >the world is random, I am denied free will because I can not >determine the outcome of a decision. Either element, determinancy >or randomness, denies me free will, so no mixture of a >deterministic world or a non-deterministic world will allow me >free will. Just so. Having one's actions determined randomly isn't much help. One of the problems with discussing free will here is that it's too easy to simply rehash arguments that have been handled in the philosophical literature. I thought it best to make this point in response to an article that I agreed with, though, because I'm not claiming that no one ever says anything valuable or that I am some kind of expert in these matters with no time to listen to the rest of you. Nonetheless, anyone who is seriously interested in such topics should be willing to do some reading. I would recommend Dennet's Elbow Room: The Varieties of Free Will Worth Wanting for its discussion of free will, for its relevance to AI, and for the interesting things that come up along the way. Jeff Dalton, JANET: J.Dalton@uk.ac.ed AI Applications Institute, ARPA: J.Dalton%uk.ac.ed@nss.cs.ucl.ac.uk Edinburgh University. UUCP: ...!ukc!ed.ac.uk!J.Dalton ------------------------------ End of AIList Digest ********************