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

Free will (reprise)

  1. Sep 4, 2003 #1


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    Since the original thread on fate vs. free will has been lost, I'll try to duplicate my contribution to the discussion here. Basically, I'm not sure that the concept of free will is actually meaningful in any special sense, if we define a free choice in the conventional sense, as follows: given a choice made by an individual under a certain set of circumstances, the choice could have been made differently by that individual, even under precisely the same set of circumstances. More simply put, free choices are not fixed or determined insofar as they could have been made differently.

    Let's consider for the sake of argument two ideal beings, one whose actions are fixed in the deterministic sense (D) and one who possesses free will (F). Let's further presume D and F live in independent parallel universes and that up to this particular point in time, D and F have lived absolutely identical lives in absolutely identical universes-- they have had the same experiences, memories and thoughts, had identical interactions with identical people, and made the same choices. (Note that D could not have made his choices any other way, whereas F just happens to have made his 'free' choices such that, to this present moment in time, they perfectly overlap with every decision D has made.) Now let's pick up the story in the present, when D and F are about to choose whether to eat chocolate or vanilla ice cream. Let's call the moment when D and F have made their respective decisions t0. Leading up to t0, we know that D and F both favor (say) chocolate to the same degree. That is, the sum total of their thought, actions, and experiences to this point have acted on both D and F such that they both experience the same inclination to choose chocolate. As a consequence of D's nature, D will inevitably choose the chocolate. However, given that F's choice is not fixed, we cannot predict what F will choose. For instance, if we could enact this scenario arbitrarily many times, we might find that D chooses chocolate 100% of the time whereas F chooses chocolate 70% of the time and vanilla 30% of the time. (Assume that F, like D, has already made up his mind that he will indeed eat one ice cream or the other.)

    We note that F's inclination to choose chocolate is expressed by his 70% likelihood of choosing chocolate over vanilla. But what accounts for the 30% of the time when F chooses vanilla? It CANNOT be that F has suddenly changed his mind on a whim, since if that were the case, D would have also changed his mind on a whim at the last instant and chosen vanilla (remember that D and F are identical up to t0). In fact, any meaningful explanation for why F sometimes chooses vanilla can be ruled out by the same logic. If by "meaningful explanation" we mean an explanation made sensible in terms of events, thoughts, feelings, etc. leading up to t0, then any meaningful explanation for why F chooses vanilla in a particular scenario must necessarily also apply to D. A free will advocate might say that at precisely t0, F has a spontaneous, non-deterministic thought that (in some scenarios) changes his mind and compels him to choose vanilla. But we have seen that this spontaneous thought cannot have any meaningful explanation; if it did, this meaningful explanation would also have applied to D. It therefore seems that there is no explanation other than that F's decision making has an inherent element of randomness to it, as if (for instance) some part of F's decision making depended on the outcome of quantum events in his brain. What good is F's free will if its agency is random and essentially meaningless?

    Now let's focus on D. A free will advocate might say that although we have been speaking of D as if he has a choice, he actually has no such choice since his decisions are already determined. I say this is nonsense, the result of a common misconception in the free will discussion. Let's first firmly establish what it means to "choose."

    From Merriam-Webster:
    1 a. to select freely and after consideration <choose a career>
    1 b. to decide on especially by vote; elect <chose her as captain>
    2 a. to have a preference for <choose one car over another>
    2 b. decide <chose to go by train>

    in a free manner: as a : of one's own accord <left home freely> b : with freedom from external control <a freely elected government> c : without restraint or reservation <spent freely on clothes> d : without hindrance <a gate swinging freely> <currencies are freely convertible> e : not strictly following a model, convention, or rule <freely translated>

    From dictionary.com:
    1. To select from a number of possible alternatives; decide on and pick out.
    2 a. To prefer above others <chooses the supermarket over the neighborhood grocery store>
    2 b. To determine or decide <chose to fly rather than drive>

    I see nothing in the above definitions that implies that D has no choice over whether to choose chocolate or vanilla. He is selecting from a number of possible alternatives: chocolate or vanilla. After some consideration, he decides to act on his preference, and he selects the chocolate. This choice is made freely, since it is made of D's own accord and with freedom from external control (there is no one holding a gun to D's head telling him he'd better pick the chocolate). I believe the crux of the common misconception lies with this last point, namely that D's choice is indeed made freely. Now, D's decision is in some sense determined, i.e. by the laws of physics. But the agency through which the laws of physics determine D's choice is precisely D's brain itself, in other words, that which is responsible for his rationality and sense of identity. D's choice is determined, not externally but internally, in other words, D's choice is determined by D himself. It is not as if D exists outside of the laws of physics and that these laws impose their determination on him from the outside. Rather, it would be more correct to say that the laws of physics simply describe that agency through which D himself makes his own choices.

    This might make D's existential situation seem demeaning or puppet-like to some, but I see no reason for this sentiment. Even though the agency through which D makes his decisions is ultimately the same as everyone else's, D's distinct set of genetics, brain structure, and personal experience ensure that he is a unique and special guy whose agency (decision making) is essentially his and his alone, even if it is built upon the common ground of the laws of physics.

    So we are left with this simple proposition: D has indeed made his own free choice. The fact that he could not have chosen otherwise is irrelevant.
  2. jcsd
  3. Sep 4, 2003 #2
    Yes, it is irrelevant to him (part of the reason that we can't prove or disprove predestination), however it is not irrelevant altogether. You see, you quoted the dictionary as defining "choose" as "to select from a number of possible alternatives". In the case of the pre-determined one, there were no possible alternatives. Of course, he will never know this, so - again - it is irrelevant to all parties involved, but the fact remains that he did not really "choose" anything (because there were no possible alternatives).
  4. Sep 4, 2003 #3


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    This is indeed a sticky point, but I would contend that vanilla exists as a possible alternative to chocolate before D's brain has settled into the state where it has definitively made its choice to select chocolate. That is, if we consider the condition of D's choice before it has been made, it exists in a sort of nebulous state that has not settled on one alternative or the other, even though it will inevitably settle on chocolate. If we have perfect deterministic knowledge of D's brain and can predict that he must choose chocolate, all we are doing is projecting the resolution of this nebulous state into some point in the future. In the moments when D is still actively considering his choice, there is not yet a definitive answer. In this sense, vanilla is still a 'possible' alternative.
  5. Sep 4, 2003 #4
    To say that it is a "possible alternative", is to say that it was possible for D to choose vanilla. This is not the case.

    Also, in a less semantic light, we haven't really defined what it is that predetermines D, we have merely taken for granted that he is predetermined. I think that the nature of the "predeterminer" has much to do with the nature of the "predetermination" itself.

    For example, if some omnipotent God were the one that predetermined D, then D would have no choice but to choose chocolate, even though the possibility would appear to present itself to him - but, in this case, an outside force is involved, which inclines D in the way that it wants. However, if it is just an intrinsic nature of the Universe, that D is predetermined to choose chocolate, then D will choose chocolate, and there was no force necessary to incline him toward it and away from vanilla, since vanilla will never even present itself as an apparent alternative.

    You're right, it is rather sticky.
  6. Sep 4, 2003 #5
    Excellent thought experiment Hypna, but a bit too much. I think you might wish to simplify it more for the sake of clarity. Tackle one issue at a time.

    For example, your original assertion concerns two simple parallel universes, but then you wander off into a discussion of Quantum Indeterminacy. It's best in my opinion to seperate the discussion on the logic train from that of speculation on the reality of our existence.

    As far as the logic train goes, the issue arises that neither D nor F can tell whether or not they have free will. Although your scenario does not mention it, for all we know both are ardent fatalists. The real trick in attempting such thought experiments is to think outside our own personal habitual boxes, to do a bit of lateral thinking. Implicitely what your argument assumes is that we somehow have the God-like ability to just know all these things.

    As for Indeterminacy and Relativity, neither has been reconciled fully to date and we do not have that God-like ability to decide if our existence is fatalistic, deterministic, or if we possess free will. Hence, the importance of addressing the thought experiments clearly and succinctly.
  7. Sep 4, 2003 #6
    Ok let me clarify, and if I'm just reiterating what mentat's saying in a different context, sorry.

    You do have a point, HOWEVER, you're going on the assumption that free will exists, which I propose it does not. In fact both (D) and (F) will make the same choice inevitably, no matter how many times the scenario is run. Because (D)=(F). They are both one and the same. Thier thought patterns are equal. Thier experiences, everything up to the point of selection is equal. Therefore each time they would follow the same behavioral patterns and decision based on the static congruity. The only way to change the selection is to introduce another variable into the equation, thus skewing the sampling.

    We may PERCIEVE our lives as free will, but in fact they are based on patterns established through our experiences, our intelligence, personality, interactions with other people and our environment, and a myriad of other variables that when combined, bring together the solution of chocolate ice cream. The variables could appear infinite, but if one could calculate them, it would all boil down to something that could have no other outcome without changing the equation. Logically it makes sense, and may one day be proven. That is when we have enough computing power to bring all the variables together to a solution.

    By the way, I think we've split up the post in half..
  8. Sep 4, 2003 #7


    User Avatar

    Staff: Mentor

    I like the thought experiment except for one thing: probability. Probability is based on how well you know the situation at hand. Sure, based on "F" eating chocolate 70% of the time in the past that there is a 70% chance that he'll eat it this time. Sounds nice.

    But wait, if you dig a little deeper, you might find that he never eats chocolate more than 5 times in a row. Or maybe he's never eaten vanilla twice in a row. Or maybe he eats chocolate in the afternoon and vanilla in the evening.

    There is ALWAYS more information that can be added to your understanding of F's decision making process to tighten up your probability. What you find is that the more information you add, the closer and closer (asymptotically) to 100% you get. So then the difference becomes 100% for D who THINKS he has a choice and 99.9999....% for F who KNOWS he has a choice.

    Mathematically, the two are equivalent.

    So what does this mean? You have freewill.

    One possible problem to my thesis: randomness. I don't beleive there is anything random about human behavior (I think psychologists would tend to agree). If for example you believe humans are 30% completely random, then you'd asymptotically approach 70% and not 100%.
    Heh. Interesting. The substance of our arguements was identical but the conclusions exactly opposite. I think we mathematically proved that freewill exists and you think we mathematically proved that freewill does not exist. So what's the difference?

    The choice.

    Freewill means having a choice. If there is chocolate and vanilla ice cream in your freezer, then you DO have a choice. The fact that your choice is predictable does NOT change the fact that the choice exists.
    Last edited: Sep 4, 2003
  9. Sep 4, 2003 #8
    Possession is ninetenths of the law only in a court of law. Among physicists, it is nothing special.
  10. Sep 4, 2003 #9


    User Avatar

    It's more likly that D chooses both vanilla and chocolate considring that we all make the right and wrong choices before we act wheather we are aware of it or not. You can choose without action. I can freely choose to become a superhero but actually becoming one is unlikly.

    Freewill. I am willing to do many things freely. But my actions have limits.
  11. Sep 4, 2003 #10
    Mathmatically it can be proven, but in the physical world it is not an absolute. I'm not speaking purely in mathmatical terms. If we were to break down his decision into variables(and there would be billions I'm sure) then we would discover that yes he might choose a different flavor when face with decision multiple times, but ultimately it would be based on the changing of a variable, however minute. It could be something as small as him glancing over at someone eating a vanilla ice cream just prior to choice, but ultimately the variation in his selection would be based on variables, not congruity.

    Let's try a different example. Dice. Basic representation of probability. One might say in a sense the dice have "free will" because of randomness. But ultimately the outcome of the throw CAN be calculated if you were able to ascertain things such as the position of the dice when they left the hand, the gravity, the degree of force when they left the hand, the vector, the angle at which they first contacted the surface thrown on, etc, etc. Granted those calculations would be extremely difficult and need to be done extremely fast to predict the outcome prior to the dice landing, but ultimately it could be predicted.

    The same principal can be applied to the human side of the equation. Yes it would be infinitely difficult to predict a person's choice, because you would have to have all the person's experiences in order to determine outcome of a situation, but it is theoretically possible.

    I've heard that the total sum of a human being's memories could be contained in 15 petabytes. Perhaps when we reach that level (not too far off) we can begin to calculate such things. It really comes down to perception. The scales on which these predictions are based, is a scale beyond our current ability to perhaps even understand, but it can't be discounted. What it comes down to is that when you break it down far enough, there cease to be any variables. Because with each specific item of data known, there are no variables, and that leads to a finite and unchangable solution.

    I would also like to point out that you're describing trying to repeat the ice cream experiment over and over again. Well if time were in an infinite loop, it would be the same repetition over and over again. That moment would repeat in exactly the same way again and again on to infinity, unless a variable was introduced. Some kind of external influence would have to interact with the time loop in order to change the sequence of events. Otherwise, he always chooses chocolate, because that is the only conclusion he can come to in that loop.

    Free will is only perception. We always try and 2nd guess our choices. We always say if we could go back we'd make a different choice. But If you COULD rewind time, you would still make the exact same decision, unless you knew of the future. Because the exact same events would occur, along with your exact same thought process and interactions, and history would repeat itsself... UNLESS you introduce a variable to change it, but that isn't free will- it's a decision based on a change in perception of events.

    See where I'm going with this ?

    The fact that you have a choice does not infer free will because the choice is already made for you. You just can't percieve it:wink: You say that there is a .0001 percent different. I disagree, because if all variables and information that are possible are known, then the variance is 0. I can see a computer one day running a scenario. Remember the movie war games" and how the computer kept running all possible outcomes? he was calculating variables. It's the exact same thing on a much larger scale. If all variables are known, then outcome can be calculated. If you put a deck of cards face down, then flip them over 1 by 1, noting each card, until you reach the last card, then you know what that card is. You don't need to turn it over because it is the only remaining card, and it is known because all others are already known. This is the concept I'm referring to, only on a much larger scale.
    Last edited: Sep 4, 2003
  12. Sep 4, 2003 #11
    And yet with human beings, you have a "choice" in the way that you handle the dice. Of course the guys running the test may "ask" you to handle them a certain way, but who says you "have" to? And if they try to "coerce" you into doing it, then you're probably better off not participating in their tests.

    Ever consider that free choice exists out of the equilibrium which exists between two opposing extremes? And, that once an equilibrium is achieved, and functionality exists (in a holistic sense), that it's pretty much up to the individual to decide?
  13. Sep 4, 2003 #12


    User Avatar

    Staff: Mentor

    That is contradictory. If the choice is already made for you then the choice isn't yours. If you have a choice, then you can make a choice. Thats axiomatic. You can't have it both ways.
    Sounds good except for one thing: Dice are inanimate objects governed purely by probability. They don't make choices either real or imagined. You have consciousness. You CAN make choices. The fact that you chose chocolate doesn't mean that you couldn't have chosen vanilla if you wanted to.

    No actually, I was saying that its 100% freewill. That .0000....1% variance is randomness going to zero.

    Zantra, how many times have you seen The Matrix and The Matrix Reloaded? They are the best discussions of fate and freewill since Sophoclese:

    "as you adequately put it, the problem is choice."

    Gawd, I can't wait for The Matrix III.
    Last edited: Sep 4, 2003
  14. Sep 4, 2003 #13


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    I just threw the quantum indeterminancy thing in there to make it comprehendable in physical terms how F's decision making could have an element of randomness to it, although it's not a key component of my argument.

    Well, you are right in saying that if we actively consider our choices to be pre-determined or free, this will ultimately have an effect on how we make our decisions, regardless of whether we actually do have free will or not. However, from a theoretical standpoint, even if F considers himself to be a fatalist, there is still a nonzero chance that he will choose vanilla; if there was not, that would imply that he could not have chosen otherwise-- so he would effectively be devoid of free will, contradicting our initial definition. Likewise, even if D considers himself to have free will, he will always choose chocolate due to his deterministic nature.

    edit: Also, I'm not saying we have the ability to discern whether a given being has free will or not. I take it as axioms for my argument that F has free will and D does not, for the purpose of exploring precisely what it means to have 'free will.' My argument basically amounts to a proof by contradiction that the 'free' component of F's decision making cannot be meaningful in any physical sense, only an element of randomness.
    Last edited: Sep 4, 2003
  15. Sep 4, 2003 #14
    No you're misunderstanding me. You're saying that we can control the person throwing the dice. We cannot change that. He will inevitably throw the dice in the manner he was meant to. various factors such as his mood, alertness, physical makeup, agility, nerve response and other things will contribute to the outcome of the roll. All of those factors can be traced back and calculated to predict the outcome. Instead of trying to control the throwers actions, we must view all the factors which contribute to the way he throws the dice.
  16. Sep 4, 2003 #15
    Another problem here is that we can't go on doing things "habitually" before we expire. Or, let's say as a result of eating so much ice cream we eventually develop diabetes, and our doctor tells us to moderate "our habits" or else? That kind of throws the whole experiment out the window as well doesn't it? The fact that the variables needed in order to keep the tests predictable have changed? And, that with change comes a "new choice?"
    Last edited: Sep 4, 2003
  17. Sep 4, 2003 #16


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    This is basically what I was getting at. I assumed free will exists at the outset of my argument only to get a handle on exactly how we could distinguish a being with free will from a being without it. My conclusion is essentially the same as yours. Because D and F are identical in every way-- except for our assumption that for any decision F makes, he could have made it differenly-- we cannot trace the degrees of freedom in F's decision making to any physically meaningful variable. But this in itself does not imply that D = F. The ONLY room for variation in this thought experiment, ie the only way it can be that F chooses differently from D, is if part of F's decision making involves an element of randomness. Thus, I conclude that if free will exists, it is only by virtue of randomness. Hence, even if we do have free will, it is not really meaningful for us anyway, since we cannot control these degrees of freedom (the random processes) that exist in our decision making.

    edit: For the sake of consistency and clarity, I should say that D and F are precisely identical except for whatever difference exists in F that gives him his free will. If you accept my argument, this statement is equivalent to saying D and F are identical, except for the extra random processes that account for F's 'free will.'
    Last edited: Sep 4, 2003
  18. Sep 4, 2003 #17
    Yes, and we can look back and say, "That must have been the way it had to be, because what's done is done and it therefore couldn't have happened any other way." Hmm ... And yet, if we had chosen differently, it would have happened another way. In which case choice represents "the potential" that we might do things differently.
  19. Sep 4, 2003 #18


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    I think you are approaching this the wrong way. You're describing our inability to definitively predict F's behavior in terms of lack of perfect knowledge of F. But we do have perfect physical knowledge of F, insofar as we know that F has lead a life identical to D's. We also know that D always picks chocolate. So why is it that F does not always pick chocolate? The difference between F's and D's behavior cannot be accounted for by any deterministic physical variable (such as, "F has a preference for never eating vanilla twice in a row"), since by definition D's decision already demonstrates the sum total influence of all such deterministic variables. Thus, the only possible way F can make a different decision is if some part of his decision making is not deterministic, i.e. random.

    Anyway, if we did have the ability to predict F's choice with 99.999999...% precision, that would imply that F really could not have made his choice differently, and this contradicts our initial assumption that F has free will.

    This point I agree with, although the crux of the matter is what you think of as choice (as illustrated by Mentat's objection).
  20. Sep 4, 2003 #19
    What's that Dirty Harry used to say? "A man's got ta understand his limitations." :wink:

    And, so long as he works within the boundaries of his limitations, then he's free to choose.

    While people don't just do stupid things like throwing themselves off of buildings if they want to live. :smile:
  21. Sep 4, 2003 #20


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    From Merriam-Webster:
    1 a. being within the limits of ability, capacity, or realization
    1 b. being what may be done or may occur according to nature, custom, or manners

    2 a. being something that may or may not occur
    2 b. being something that may or may not be true or actual <possible explanation>

    I believe you are using "possible" in the sense expressed in 1a and 1b, and I am using it in the sense expressed in 2a and 2b. It is indeed beyond the limits of D's capacity to choose vanilla. However, before we know what D's decision is, we can say that vanilla is a possible choice insofar as it may or may not be the case that D chooses vanilla; we're just not sure yet.

    Anyway, even though I concede that vanilla is not a possible alternative in the sense you're using the word, I don't think this tears down my whole argument by implying that D has no choice. The situation is simply that D must decide between vanilla and chocolate, and D makes his decision based on his own thought processes before ultimately selecting chocolate. If you contend that D has no choice in this situation, you are essentially saying that D has no decision to make either. But clearly he must roll the question around in his mind and finally decide on chocolate; this is nothing but the process of deciding, or choosing. Again, the fact that it could not have been otherwise is irrelevant.

    This is basically the internal/external distinction. When I said that D has a choice even though he could not have chosen otherwise, I have assumed that D is indeed self-determined, i.e. his actions are controlled by his brain/mind/self (albeit through the agency of the laws of physics). It could be that there is some demiurge determining D's actions for him, in which case he really wouldn't have a choice-- but that's besides the point. Most people automatically equate pre-determination with lack of choice, regardless of whether the determining is done internally or externally. I simply want to establish that even a deterministic being has a choice, so long as the outcome of his choice is determined internally (ie, by that being itself).
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Similar Threads - Free reprise Date
Can an AI have free will? Yesterday at 11:41 PM
How much free time do you guys get? May 2, 2017
Awesome free way to learn physics through experiments Mar 25, 2017