• Banno
    23.1k
    I have another Anscombe article! Inevitability both a joy and a frustration. This one is Causality and Determination.

    So we procure for ourselves a Galton box:
    800px-Galton_box.jpg

    The standard philosophical prejudice is that given an accurate enough account of the position of the box and a given ball, a competent physicist will be able to tell us which of the bins across the bottom the ball will land in.

    And in this sense the path of the ball is determined.

    But of course no one could determine the final resting place of the ball. Even the smallest error in the initial positions will be magnified until it throws out the calculations.

    Anscombe wrote this in a time of only nascent chaos theory, which could only serve to amplify her point.

    The notion that the universe is determined fails.
  • Streetlight
    9.1k
    Damn you I'm trying to read other stuff look what you're making me do.
  • Banno
    23.1k
    :joke:

    It's sweet.

    I must have read it long ago, because it mentions Feynman's bomb, which I have used hereabouts previously. A bomb is attached to a geiger counter in such a way that it might explode; but it might not.

    What, if anything, caused the explosion? And hence we find that there are different causes.
  • Banno
    23.1k
    Anyway, the upshot is that the philosophical community is beset by a fantasy of determinism.
  • TheMadFool
    13.8k
    smallest error in the initial positionsBanno

    given an accurate enough account of the position of the box and a given ball, a competent physicist will be able to tell us which of the bins across the bottom the ball will land in.Banno

    The notion that the universe is determined fails.Banno

    You're basing your argument on error. Do you mean that if there were no error, the physicist would know where each ball will land? If yes, then determinism is true. If no, then that's real indeterminism. Can you prove it?
  • Wheatley
    2.3k
    Anscombe wrote this in a time of only nascent chaos theory, which could only serve to amplify her point.Banno
    That word never really appealed to me anyhow. Neither 'indeterminism.'
  • TheMadFool
    13.8k
    I suppose determinism/indeterminism has significance only to the extent that it bears on human freedom (free will). One problem with the notion of free will is the nature of this freedom, particularly that aspect of its nature as relates to causality. A claim that we possess free will is one that attributes to us the power to make choices that are not the effects of things that are beyond our control. Put differently, if we possess free will, the act of making a choice is not an effect in the causal web.

    So far so good, I must say but, what exactly does this freedom, this free will, provide, by way of benefit, to us? At this juncture what must be mentioned is the idea of responsibility - if we possess free will then, it's claimed, we're completely accountable for our actions. Now, what exactly are actions? What is an action? Isn't an action just a link in the causal web? Actions are actions to the extent that they're part of causality. Without causality, there can be no such thing as an action. Free will is important only so far as we wish to control causality via our actions.

    In other words the idea of free will involves being both not a part of the causal web (choices are not effects) and also a part of the causal web (our actions have effects).. To want to have it both ways is never a good idea unless.. :chin:
  • Banno
    23.1k
    There can never be no error.
  • Banno
    23.1k
    if we possess free will, the act of making a choice is not an effect in the causal web.TheMadFool

    So... our choices are accidental? Random?

    ...responsibility...TheMadFool
    The argument would be something like "I want someone to blame, and I can't do that if you did not freely choose, hence you have free will"? That's, as you imply, a poor argument. Nor, I think, is it cogent.

    Yes, free will theorists seem to want it both ways.
  • Kenosha Kid
    3.2k
    The notion that the universe is determined fails.Banno

    The notion that a competent physicist could predict the outcome of a single ball drop fails if there is any uncertainty, however small, of the initial state of the ball. But chaotic systems are nonlinear (sensitive to initial conditions) deterministic systems. The outcome is knowable in principle but not in practise (technological limit).

    The same distribution above is also seen in the electron single-slit experiment. In that instance, even if the initial wavefunction were known perfectly, the outcome of where the electron ends up is unpredictable. Depending on your interpretation, that is non-deterministic.
  • Banno
    23.1k
    Yep.

    Are you aware of Physics without determinism: Alternative interpretations of classical physics? Mentioned in Has physics ever been deterministic?.

    The upshot seems to be that determinism is a metaphysical assumption from which the classical determinist view of physics follows, and that this assumption can be removed with suitable mathematical alterations.

    I don't have access to the actual paper.
  • Metaphysician Undercover
    12.3k
    The upshot seems to be that determinism is a metaphysical assumption form which the classical determinist view of physics follows, and that this assumption can be removed with suitable mathematical alterations.Banno

    The determinist metaphysical assumption is very well expressed as Newton's first law. This law assumes as a given, the temporal continuity of any describable set of conditions, requiring a "force" to change anything. The prerequisite "force" is the cause of change, which the determinist latches on to. How do you expect to remove the requirement of a force as the cause of change, without denying Newton's first law?
  • Banno
    23.1k
    Given the deep understanding of mathematics and physics that you have demonstrated elsewhere, you might forgive me for not paying your posts here much attention.
  • Kenosha Kid
    3.2k
    I don't have access to the actual paper.Banno

    Cheers, I'll take a look around. Although I guess there's no guarantee of finding it without determinism. :wink:
  • Banno
    23.1k
    Ah, sweet. Thanks.
  • fdrake
    5.8k
    The upshot seems to be that determinism is a metaphysical assumption from which the classical determinist view of physics follows, and that this assumption can be removed with suitable mathematical alterations.Banno

    I agree that that conclusion fits with Anscombe's paper.

    It ought not to have mattered whether the laws of nature were or were not deterministic. For them to be deterministic is for them, together with the description of the situation, to entail unique results in
    situations defined by certain relevant objects and measures, and where no part is played by inconstant factors external to such definition. If that is right, the laws’ being deterministic does not tell us whether ‘determinism’ is true. It is the total coverage of every motion that happens, that is a fanciful claim.

    If you write down a bunch of equations involving a time variable, take an initial state at t=0, and evolve the system forward to t=t, determinism there would just be that there is a single state at t=t and that the complete specification of the t=0 state entails that there is a single possible state at t=t. "That" state is logically entailed by "this one".

    Determinism as Anscombe seems to characterise has a much broader scope; it doesn't matter whether if a completely specified t=0 state entails a unique t=t state within a model, determinism has to guarantee this complete specification outside the model too. Every state must be completely specified by a trajectory taken in accord with a hypothetical physical law. The determinism spoken about is the panoptic vision of Laplace's demon, if we know all now, then we know all.

    We can never be in the position of Laplace's demon; and that's a fact of life. So we can't recourse to Laplace's demon and begin conjuring up hypothetical scenarios that take a hypothetical completely known input state and propagate it through our knowledge to a completely known output state.

    Nevertheless, if we have a good model that is deterministic, it works predictively. Such a model doesn't care if reality is watched over by Laplace's demon or not.
  • Banno
    23.1k
    Nevertheless, if we have a good model that is deterministic, it works predictively. Doesn't matter if reality is watched over by Laplace's demon or not.fdrake

    We can also drop that the model is deterministic: if we have a good model, it works predictively.

    Isn't that right?
  • fdrake
    5.8k
    Isn't that right?Banno

    Think so! I was writing in the context of the claim: "If there were no Laplace's demon, then our deterministic models wouldn't work", the point being it doesn't matter for deterministic model functioning if there is a Laplace's demon or not.
  • Olivier5
    6.2k
    the philosophical community is beset by a fantasy of determinism.Banno

    The belief in determinism is not universal among philosophers. Ever read Karl Popper?
  • Christoffer
    1.8k
    But of course no one could determine the final resting place of the ball. Even the smallest error in the initial positions will be magnified until it throws out the calculations.Banno

    How is this premise concluded with the conclusion below?

    The notion that the universe is determined fails.Banno

    Just because we cannot measure all parameters necessary for predicting the future, doesn't mean that all parameters aren't there to determine the outcome.

    We can argue for quantum randomness affecting the larger scaled world, but it is probabilistic and the probable outcomes become infinitely more deterministic as soon as we leave a planck scale. Saying that something is random when we don't have enough ability to calculate all parameters doesn't mean it is random.

    And if we include ourselves into the equation, we are far too big in scale to be free of deterministic randomness.

    Take the idea of a quantum dice. The illusion we have is that if we make choices based on this, we are neither free nor determined, but actually acting out of randomness. However, this is also an illusion since the quantum randomness gets determined by measuring it. So surrounding parameters still determine the outcome, combined with the apparent choice of throwing the dice and making choices based on an outcome.

    The likely truth of the universe is that it's based on probability on a scale so small that the order created out of it becomes deterministic.

    I.e the universe as perceived and measurable for definitions on how it works becomes deterministic and anything outside of that is neither perceivable nor measurable. It becomes rather a reality vs unreality. To describe our universe is to describe it with and through reality. Any attempt to describe it through unreality fails to be relevant as it's not part of reality both in perception and in terms of ability to be measured.

    What are we then talking about when talking about quantum randomness if we can't talk about it, perceive or measure it without destroying it into a deterministic reality? Unreality is unable to be compatible with reality, we cannot define it and it cannot define reality. Therefore quantum randomness can be concluded to exist through math, but it cannot be part of reality as it's not a component of what makes up the universe, it's outside of it.
  • Banno
    23.1k
    Cool. I'm still working my way through the Del Santo article. Anscombe no doubt has one eye on the ethics of free will, so yes, her account is broader.
  • Banno
    23.1k
    Ever read Karl Popper?Olivier5

    A bit. What do you have in mind?
  • Banno
    23.1k
    Not too sure where you are going. In this thread I'm more interested in uncertainty in classical physics.
  • Kenosha Kid
    3.2k



    I'm not sure if it's that profound, although I confess I haven't finished it yet. But looking at his definition of infinite precision, it disallows infinite epistemological precision. When a classical physicist says the initial conditions can be known in principle, she means zero epistemological uncertainty in principle. This is a practically unattainable limit, but that's what "in principle" is understood to mean.

    Del Santo's definition, all his own afaik, is that infinite epistemological precision means that the number of of decimal points has no finite lower bound, but may not be infinite. This is equivalent to saying that the error may be arbitrarily small, but not zero, which, to me, says it cannot be arbitrarily small. His infinite precision is that attained by infinite technological progress which always approaches, but never reaches zero uncertainty. It is in itself a reasonable definition, but he is using different language to cast doubt on determinism rather than using argumentation within the same language.

    If you couple this arbitrarily small but nonzero uncertainty to a chaotic time-evolution, it is true you cannot predict the outcome of an event. But it is true because you could not specify the initial conditions exactly. I'm at a loss as to what can we learn from this?
  • Christoffer
    1.8k


    In context, I mean that quantum randomness doesn't affect classical physics since classical physics deals with large determined sizes. Quantum randomness boils down to predictability as soon as the scale of the system is above that of scales required for random operation. I.e we have a random system that defines the properties of an object and those properties don't change since the probability of them being broken is so low it could be considered infinitely improbable. Classical physics can't define this randomness since classical physics breaks down at those scales, but that doesn't mean they can't measure deterministic outcomes if the ability to measure is powerful enough, since classical physics calculate the actual reality of our universe. Quantum randomness is outside of this reality since it exists on a scale where our reality does not work and cannot work because of it.

    It's the same as with black holes. Spacetime breaks down within it and while we can measure, calculate, and speculate about the reality within a black hole, it is impossible to define it with measurements of our reality. Inside a black hole, there's unreality. The same goes for quantum randomness, it is outside reality. Both the black hole and quantum randomness affect reality, but they both become deterministic when "entering" our reality in the form of causal events.
  • Banno
    23.1k
    pp 5-6, there is an interesting arguent that concludes
    One thus ought to consider alternative interpretations of classical physics that do not enforce the principle of infinite precision. Namely, interpretations that do not assume that physical quantities take values in the real numbers.

    When we make a measurement, say the velocity of a falling object at t=1, need we assume that there is indeed a rational number that represents that velocity with infinite precision? Are we better off, epistemologically and ontologicaly, making no such assumption?

    Edit: That is, why assume that we could have infinite information about our measurement...
  • Banno
    23.1k
    Sure. The present conversation is not about quantum randomness. It's about indeterminacy in classical physics.
  • TheMadFool
    13.8k
    There can never be no errorBanno

    Is it possible that as the error is reduced, the accuracy of predictions also increases commensurably?

    If yes then doesn't it imply, at least in a theoretical sense, that with zero error we can make perfect predictions? This then goes to show that what we think of as indetermnism is a product of our ignorance rather than something that inheres in nature.

    So... our choices are accidental? Random?Banno

    Is it an either-or choice between determinism and randomness? Is there no third, more palatable, alternative? Is it possible to retain determinism that applies to our actions but eliminate determinism as regards the making of choices, all the while avoiding the world of randomness. This is what I reckon I described as having the cake and wanting it too - we want to be causes, something that endorses causality but we don't want to be effects, something that rejects causality. If I carry this to its logical conclusion, we all want to be what in theological circles is known as a prime mover. In a sense then humans want to be god or, at the very least, god-like. If free will is real then we are all prime movers in our own right, capable of each initiating a causal chain that has no antecedent cause.

    I went off on a tangent there but returning to the question of whether randomness is entailed by rejecting determinism and how this randomness affects free will, I'd say that randomness here doesn't apply to the act of making choices but rather to how attractive or not the choices are. Determinism seems to operate on the desirability/undesirability of options and, within its framework, it's claimed that when a person makes a choice s/he does so because s/he has no control over her preferences, preferences that decide how appealing or unappealing a choice is. The act of choosing itself is not considered as part of this deterministic argument.

    If the above is correct then randomness implies that all available options become equally desirable/undesirable, no choice is either more appealing or less appealing than the other. If then a choice is made, it must be that this occurs in the complete absence of any and all influences that can affect making choices i.e. the notion of free will remains unaffected by randomness. In short, randomness doesn't mean our choices are random but that available alternatives in a given situation become all equally desirable.

    The situation here is analogous to the following:

    There's a person X who's made to choose between two cold drinks, say Pepsi and Coke.

    At one time, he's fed with information regarding how beneficial Coke is compared to Pepsi. This information plays the role of determinism by influencing the choice he'll make when offered a drink.

    At another time, no information is provided to him and both cold drinks become equally desirable/undesirable as the case may be. This is when determinism is false, and randomness is true, as there's nothing to influence his decision on which drink he'll choose.

    Note that the act of choosing itself is not part of determinism or randomness. All that determinism does is make choices appealing or unappealing. Ergo, randomness simply equalizes the appeal of the options given and this being done if a choice is made then that choice is free and free will exists.

    When we are faced with a choice, there is (seems to be) a pause in the causal chain as we consider the options available which, to me, suggests an interruption in the causal chain i.e. the act of choosing itself seems to be, in some way, not part of the causal chain. If that's true then determinism affects only how attractive/unattractive options are.

    If you don't mind, I'd like your views on this. Thanks. Something doesn't seem right.
  • Christoffer
    1.8k
    Sure. The present conversation is not about quantum randomness. It's about indeterminacy in classical physics.Banno

    Maybe I fail to see how they are not connected? Randomness in causality physics only exists when we have a lack of ability to measure all possible parameters. The only randomness that exist which breaks determinism is at a quantum level, a level that is incompatible with reality and therefore can't be counted as part of it, just like we don't count the breakdown of spacetime in a black hole as part of our reality, but instead exist outside it.

    So if our reality is measured with classical physics, determinism is not broken. If we lack the ability to spot errors, that doesn't mean determinism breaks, only that our tools did.
  • fdrake
    5.8k
    If you couple this arbitrarily small but nonzero uncertainty to a chaotic time-evolution, it is true you cannot predict the outcome of an event. But it is true because you could not specify the initial conditions exactly. I'm at a loss as to what can we learn from this?Kenosha Kid

    One way of phrasing Laplace's demon is "If the initial state within an arbitrary system was completely specified, it would evolve into a unique state at any given time point". It's an implication; complete specification of initial state => complete specification of trajectory.

    But we live in a world where complete specification of the initial state is practically impossible for almost all flows over time. That is, it is a fact that they are not completely specified for the most part.

    So the determinism of Laplace's demon is a hypothetical; if complete specification of input state, then complete specification of output state.

    That determinism of implication doesn't hold for almost every phenomenon because we know it's practically impossible to completely specify the input state that lead to its emergence. Its antecedent is false, so it is useless as an implication; it ceases to apply. What remains of Laplace's demon as an ontological thesis when it's rendered merely a hypothetical? We can't feed almost every system into its defining implication. So what systems are left?
bold
italic
underline
strike
code
quote
ulist
image
url
mention
reveal
youtube
tweet
Add a Comment