Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Susskind and Hawking on hard determinism

  1. Aug 23, 2010 #1
    Does anyone know the stances of prominent physicists (Susskind, Hawking, others) on hard determinsm?
  2. jcsd
  3. Aug 23, 2010 #2
    I think they do. :devil:
  4. Aug 24, 2010 #3


    User Avatar
    Science Advisor

    I'm pretty sure any scientist worth their salt agrees with hard determinism, that free will is just an illusion. However, I don't think many have made their stances public on this matter.

    The only way out of hard determinism, after all, is some sort of supernatural entity, and most scientists don't go for that.
  5. Aug 29, 2010 #4
    How is hard determinism justified in light of spontaneous generation? Is there believed to be some unidentified cause for this phenomanon?
  6. Aug 29, 2010 #5


    User Avatar
    Science Advisor

    Huh? Spontaneous generation is an old, falsified theory regarding the origin of life, specifically that life was produced from non-living matter spontaneously all the time. This does not happen. What are you talking about?
  7. Aug 30, 2010 #6
    My bad. It must be clear by now I am no physicist; I am more informed about philosophy, which I admit is not as interesting. I think I meant to use the term "spontaneous particle creation."

    So... my revised question is: How is hard determinism defensible when we consider spontaneous particle creation?

    Is the creation of these particles truly spontaneous? Or are there specific conditions necessary for their creation?

    Because clearly, if spontaneous particle creation can occur without a cause, hard determinism cannot be valid.

    Also, the double slits experiments seem to invalidate hard determinism. Hard determinism demands there must be a singular inevitable effect for every cause. The double slits experiments produce multiple effects for a single cause (and the experimenters insist the conditions are constant). How then are hard determinism and quantum physics combatible?
  8. Aug 30, 2010 #7


    User Avatar
    Science Advisor

    Let's consider the two different aspects of hard determinism separately:

    1. Logical determinism: All propositions are either true or false. This is not contradicted by quantum mechanics in any form.
    2. Causal determinism: the statement that all effects have causes. This one gets a little bit complicated, because in physics there is often no clear distinction between "cause" and "effect". However, we can recover the same meaning through invoking what is called "unitarity". If a set of physical laws are unitary, then this is an indication that if we know the entire state of the system at a given time, then we can, given sufficient computer power, calculate the state of the system at any other time. So far, all discovered laws of physics hold to this statement. There was some thought that quantum mechanics was non-unitary, due to its apparent random nature under wavefunction collapse, but this was basically solved by Everett back in the 50's.

    In this view, the vacuum itself has a dynamic wavefunction, and if we knew that wavefunction perfectly, we could, in principle, calculate each and every particle that fluctuated out of the vacuum. So yes, it is the underlying behavior of the vacuum itself that would be the "cause" in this sense.

    This gets a little bit more complicated in terms of wavefunction collapse. As Everett showed, the appearance of collapse happens without inserting an assumption of collapse into quantum mechanics. So there is no reason to believe that there is any such thing as real wavefunction collapse. But even if we postulate that wavefunction collapse is real, instead of each "effect" being a result of fully-deterministic evolution, some "effects" are random results of the underlying behavior, and we could, if we knew everything, compute the probability distribution of each result. I don't think this changes any of the philosophical implications in any meaningful sense.

    The only way I've seen that people have attempted to change this picture in a meaningful sense is to attempt to invoke the effects of observers in determining the underlying behavior. But that's magical thinking that almost nobody within physics takes seriously.
  9. Aug 30, 2010 #8


    User Avatar
    Science Advisor
    Gold Member

    In this universe just about anything possible must eventually happen, given enough time. This is not so much deterministic as it is inevitable under the rules of quantum theory. Does that mean we are unique? Almost certainly. Does it mean there are no other sentient intelligent life forms in the universe? Almost certainly not.
  10. Aug 31, 2010 #9
    1. This is definitely not true in multi-history universe
    2. Truly infinite worlds (where initial conditions are not prepared artificially in a very specific way) have the same properties as MWI, the only difference that 'alternatives' are far from each other in space while in MWI they coexist in the same one.
    3. In both cases I have no idea how to define 'FREE WILL' when ALL alternatives occur:

    You can chose between alternatives A and B ==> you have free will
    Choice between A and B is predetermined ==> no free will
    Both A and B are guaranteed to occur ==> ????
  11. Aug 31, 2010 #10


    User Avatar
    Science Advisor

    MWI doesn't imply that all alternatives occur. It only implies that all possible alternatives occur. It is not necessarily the case that anything we could imagine is actually a possible alternative.

    In either case, what does occur is everything that does occur is due to purely deterministic evolution of the wavefunction, so there is no choice to be had, so free will is just an illusion.
  12. Aug 31, 2010 #11
    The most active in determinism of the prominent physicists is Gerard 't Hooft. From his web site you can see that one of his research interest is: "Fundamental aspects of quantum physics. I have deviating views on the physical interpretation of quantum theory, and its implications for Big Bang theories of the Universe. This topic has been expanded upon in recent publications, cf. "The mathematical basis for deterministic quantum mechanics"; The Free-Will Postulate in Quantum Mechanics; Entangled quantum states in a local deterministic theory, arXiv:0908.3408 (see my publication list)."

    The 't Hooft idea is that "there is a close relationship between the quantum harmonic oscillator and a particle moving along a circle". From this idea follows arXiv:0903.3680 (Compact time and determinism: foundation) which I believe is the most promising and powerful approach to determinism in relativistic quantum mechanics. The "particle along a circle" in this case is the "de Broglie internal clock", that is a periodic phenomenon (a periodic field) with periodicity in time T=h/E. The de Broglie internal clock of an electron is faster that 10^-20 s(<h/m_e c^2). As for a "dice rolling to fast" w.r.t the resolution on time it is impossible to predict the outcomes. This is why quantum mechanics appear to be indeterministic. But as for a dice, if you time resolution is sufficiently good (modern experimental techniques are far to reach a time resolution of 10^-20 s) you would be able to describe the underlying deterministic dynamics.
    Last edited: Aug 31, 2010
  13. Aug 31, 2010 #12


    User Avatar
    Gold Member

    No, these are the rules of probability, that everything which is possible must eventually happen.

    Monkeys as we speak are trying to crack Shakspeare's Romeo and Juliet.
  14. Aug 31, 2010 #13


    User Avatar
    Science Advisor

    Everything that we can imagine, however, is not necessarily possible. We don't necessarily know what the outer bounds of possibility are. In fact, we actually can't know because we can't have the full information of the underlying wavefunction available to us (even if we did have the computer power to process such a massive beast).

    In fact, in a rather deep sense, our universe hides from us so much of its information, that the argument over whether the universe is purely deterministic or deterministic + random is largely academic. As far as every observation we will ever make is concerned, the universe is deterministic + random. But, I claim, this difference is not a departure in any sense from the philosophical implications of hard determinism.
  15. Sep 2, 2010 #14
    I don't believe in hard determinism. I don't think that free will is an illusion.

    There are a number of other ways. One is that collections of objects exhibit behavior that is independent of those objects, and "will" comes from the behavior of those objects.

    Personally, when someone says that the *ONLY* way something is possible, I'm not going to accept that statement unless they show me a formal mathematical proof. There is a big difference between saying *I've proved mathematically that the only answer you can get when you add 2 and 2 is 4* and *I think this is only possible because I can't imagine it*.

    "Proof by lack of imagination" is something I expect from creationists, and I'm getting worried that people I used to respect like Dawkins are showing the sort of habits I've learned to dislike from creationists.

    Of course the problem with trying to prove the non-existence of free will by mathematical methods is that then you have to make the statement in formal mathematical terms, which as far as I can tell hasn't been done.
  16. Sep 2, 2010 #15
    But it is quite strongly contradicted by Goedel's incompleteness theorem which says that any consistent formal system which is stronger than mathematics contains statements that are unprovable.

    Also one very interesting field of mathematics is non-standard logic.

    But quantum mechanics kills this because there is behavior that is not obviously specified in the state of the system. I have two uranium atoms with identical state, one decays now, one decays a million years from now.

    Also, you run into problems when the computer is part of the system rather than something external. If you have the computer as being external to the system, then you run into thermodynamic and Maxwell daemon type problems. One you put in thermodynamics, you find that it is impossible to build a computer with infinite computer power, at which point things get interesting.

    What you end up with are states that are not observed and *cannot be observed*. If we just label those states "free will" then free will exists.

    But we can't. In fact I'm pretty sure that you can mathematically show that you can't know the state of a wavefunction perfectly. What happens is that if you *could* know the state of a wavefunction, it turns out that you run into huge causality problems.

    This is actually a major problem in quantum computing.

    Again, if we could know everything and if we could compute everything, then we'd have huge thermodynamical problems. There are some fundamental limits in how much you can compute and how fast you can compute. The basic issue is that if you had infinite compute power, then you could construct Maxwell's daemon.

    The fact that you are limited in what you can observe and what you can compute, and those limitation are *fundamental* has a lot of philosophical implications.

    But any case, statements about the universe that assume the existence of infinite compute power are flawed, because you can show that these contradict thermodynamics.
  17. Sep 2, 2010 #16
    Define "illusion" (seriously).
  18. Sep 3, 2010 #17
    Actually we do. You can use formal mathematical methods to show that certain things are not possible.

    The problem with using that to prove things about "free will" is that you can't put the term "free will" into formal mathematical language.

    This is a problem because if you have a discussion in which the terms are *not* rigorously defined. Then you have problems because people take results from definition A and apply them to definition B.

    I believe that it is possible to show that if you assume the rules of quantum mechanics and thermodynamics are correct that any complex system will exhibit behavior at time t+deltat which cannot be calculated from the observable state of the system t. (If anyone disagrees we need to have that discussion.) This is a statement that can be mathematically proved correct or incorrect.

    Now if I *define* "free will" as quantum states of a system that cannot be observed or computed. It's a definition, and if you accept that definition, then mathematically free will exists. Now you can challenge that definition, but then it's not clear what the rules are for defining which definition is correct or not. It's also dangerous when people use the same word to mean two different things.

    Or you can argue that there is some external supernatural force that is pulling the strings. That also fits the observations. Now you may not like the idea of some supernatural force, but you can't exclude that explanation from the observations.

    But it is. In law and ethics, "free will" is an rather important concept. If I shoot someone, it matters a lot if I "intentionally" did it or I "accidentally" did it. Now it may in fact be that when a lawyer, ethicist, or philosopher utters the term "free will" he is really talking about something other than quantum states, and it's not my definition of free will.

    But if it is the case that we are talking about different things, then it because dangerous for physicists to make statements about "free will not existing."

    Also I do think that when physicists make statements, that they need to be really careful that any scientific authority that they have isn't used when they make non-scientific statements.
  19. Sep 3, 2010 #18
    First of all, it's not clear that there isn't an observational way of distinguishing different interpretations of quantum mechanics. As far as I am aware, no one rigorous demonstrated that the different interpretations of QM are observationally equivalent.

    Second of all, what we can know and what we can't know about the universe is not an academic exercise. A lot of the research I'm interested in comes from trying to make faster and smaller computers, and we are hitting some fundamental mathematical and physical limits. What can be computed and what *can't* be computed is a very active area of research with a lot of practical applications. For example, whether it is mathematically possible to break encryption or whether the laws of the universe mean that you can't decode the messages the bank and the ATM are sending each other makes a lot of difference.
  20. Sep 3, 2010 #19
    Not only this, but also a probalem can be well-defined and have a solution, but it can be non-calculable

    Also, as I suggested before, it is possible that things like 'free will', 'consciousness' have some weird mathematical properties so they cant be, in principle, understood. Check the first post here:


    Finally, Goedels theorem is in fact very important. For example, the well-known Game of life on infinite board has very simple, deterministic and calculable physical laws, however, Goedel theorem is applicable to it (because you can build real, infinite Turing Machine in that universe) so some statements about gave of life cant be proven or rejected.
    Last edited: Sep 3, 2010
  21. Sep 3, 2010 #20


    User Avatar
    Science Advisor

    That's, um, impossible. Because the behavior isn't independent of the components. The behavior may be radically different and not easily computable from the components, but it is still dependent upon them.

    I don't see how it's that hard. If the fundamental laws of physics are unitary, then the physical configuration at any one specific time has a one-to-one correspondence to the physical configuration at any other time. Yes, you might be right that the actual configuration could conceivably be uncomputable, even with infinite computer power. But it doesn't actually matter, because if unitarity is true, then the system before and after some passage of time, due to the one-to-one correspondence, are actually just the same system. You can't "sneak in" free will without breaking that one-to-one correspondence.

    Of course, we know that actually calculating the behavior of a brain is a practical impossibility, and you're right that it may even be a fundamental impossibility. However, this just means that the illusion of free will is a really really good illusion. It doesn't mean that free will actually exists, just that the illusion is good enough that we don't need to worry about it. Not knowing or even being unable to know the result doesn't change this, as long as the result is necessarily a consequence of the previous configuration.

    The only way around this, as near as I can see, is to redefine free will so that it is consistent with determinism. But classically free will has been considered to be in opposition to determinism.
  22. Sep 3, 2010 #21
    And actually those are called "emergent properties or behaviours" and are considered dependent of the components in a very, very far sense from that of "hard" determinism, so you should check again what "hard" determinism is, because otherwise you are admitting twofish-quant is right.

    You again have softened your determinism quite a bit, which is not bad. As soon as you realize that Hume proved centuries ago that the idea of the one to one correspondence in time of systems(causalty) is what actually turns out to be a darn good illusion (and of real practical utility indeed), you will also have rejected soft determinism.
    Last edited: Sep 3, 2010
  23. Sep 3, 2010 #22


    User Avatar
    Science Advisor

    Huh? Hard determinism, as near as I can tell, makes no assumptions whatsoever about whether or not we do know or even whether or not we can know our future choices. It's merely a statement that those choices, whatever they may turn out to be, are a consequence of physical law and wholly determined by the previous state of the universe.

    The difference between "hard" determinism and "soft" determinism isn't an actual difference in discussions about the behavior of the physical universe. It appears to be, instead, a difference in the definition of free will, where for soft determinism one weakens the definition of free will so that it is compatible with a completely deterministic universe. I don't see this as a meaningful distinction, especially because one reaches the exact same conclusions on moral behavior whether through hard determinism or soft determinism.
  24. Sep 3, 2010 #23
    Yes it is. Temperature is an example of an emergent phenomena that is independent of the behavior of the components. Again, we may be using different definitions of what a "dependent component" is.

    To give to other examples, computer programs and money are two things whose behavior cannot be derived or described in terms of their components, because both computer programs and money cannot be broken down into physical components.

    Here is an example. I have two $20 bills that are physically exactly identical. Since they have the same serial numbers, one is counterfeit and one is not. This is an example of whether something is real or not has nothing to do with physical properties. Also, how money behaves has nothing to do with the physical manifestation of that money. $20 is $20 whether I represent it in paper money or ledger entries. If I take an bank ledger entry marked $20, and then I copy that ledger entry into something made of stone, and I burn the original paper book, it's still $20.

    Also, there are things that are uncomputable as a matter of either fundamental physical or mathematical limits.

    The problem is that if the configuration of system cannot be determined from observable information (and it can't be) then you have system behavior that is independent and non-determinable from external observation. You may have a one-to-one correspondence between behavior and physical state, but there is no one-to-one correspondence between physical state and observables, then you have a state that allows for behavior that cannot be determined from what is observed.

    Define illusion. I don't have a good sense of what you would consider an illusion what what you would consider real. It's not immediately obvious to me why you'd consider free will to be an illusion, and electrons and atoms to be non-illusionary. What about marriage? Is marriage an illusion? Money? Is money an illusion?

    I don't see how you can come up with a definition of reality by which free will is an illusion but money, marriages, and Newtonian forces are "real." I'd be interested in seeing you try.

    Except that using scientific arguments you can't force me to accept the existence of the hidden states that you need to get determinism to work. If the result is *unknowable* and *unobservable*, I can come up with perfectly good rules of physics in which the configurations you talk about don't exist at all. If you believe that those states exist, that's fine, but you can't through scientific argumentation force me to conclude that those hidden variables exist.

    At this point we are in the realm of theology.

    Or to come up a non-deterministic model of the universe. If you are a determinist that is fine, and I'll treat this as a quasi-religious belief just as if you are Catholic. However what I object strongly to is your original statement that which is that no scientist worth his salt would reject hard determinism. I think I'm a scientist. I reject hard determinism. Statement False QED.

    The fact that people have believed things for thousands of years doesn't make it true.
  25. Sep 3, 2010 #24
    The moment you show that those states are unobservable and uncomputable, then this is a matter of theology and not science, which you can reject if you don't like it.

    I say those states that you talk of don't exist. Since they are not directly unobservable and cannot be computable, then you can't show scientifically that I'm wrong. I can come up with non-deterministic models of the universe in which those hypothetical states that you speak of don't exist, and those models are as good as the ones you come up with.

    Not clear. I don't know if ethicists and physicists are talking about the same thing when we talk about free will. The problem is that formal mathematical logic is not useful in formulating ethical rules. One problem is that most people are not that good at formal logic, and so a system of ethical rules that is based on formal logic turns out to be unusable.

    Personally, I don't think that physicists have anything useful to say about moral behavior.
  26. Sep 4, 2010 #25


    User Avatar
    Science Advisor

    So what? That doesn't change anything.

    Free will is defined as the ability to make choices free from constraints: that one can, in principle, make any choice. One may not be able to act on any choice, but one can make any choice. In principle. But the laws of physics guarantee that whatever choice we do make is actually wholly and completely determined by the past physical state of our own brains combined with whatever our brains interact with in the mean time. The idea that we could make any choice at all is an illusion.

    One way in which the illusion is maintained is with our own ability to consider hypothetical scenarios. We often consider hypothetical scenarios, and try to consider what their outcomes will be. We then select among the alternatives based upon the perceived desirability of those outcomes. This whole process feels, to us, like we have the option of choosing any of the hypothetical scenarios, when in reality the choice we do make is determined by a combination of our own preferences and the outcome of this hypothetical mind game.

    Er, we're talking about language here. As in how words are used. Definitions are only decided through consensus. Believing a definition is a certain way, for thousands of years, certainly does make it so. If you can come up with a new definition, and successfully get others to accept it, then fine. But then we're just arguing semantics and I'm not really interested any longer.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook