On 5/1/2010 9:33 PM, Jesse Mazer wrote:

On Sat, May 1, 2010 at 11:07 PM, Rex Allen <rexallen...@gmail.com <mailto:rexallen...@gmail.com>> wrote:

    On Sat, May 1, 2010 at 8:40 PM, Jesse Mazer <laserma...@gmail.com
    <mailto:laserma...@gmail.com>> wrote:
    > On Sat, May 1, 2010 at 1:43 PM, Rex Allen <rexallen...@gmail.com
    <mailto:rexallen...@gmail.com>> wrote:
    >> On Thu, Apr 29, 2010 at 10:58 PM, Jesse Mazer
    <laserma...@gmail.com <mailto:laserma...@gmail.com>>
    >> wrote:
    >> >
    >> > I think you've got the argument wrong.
    >> I think you're wrong about my getting the argument wrong.  :)
    > I suppose it depends what you mean by "the argument". It is
    possible you
    > could find *some* mainstream scientist who seriously considers the
    > possibility that all our historical records of a low-entropy
    past are wrong
    > or that we are actually Boltzmann brains with false memories,
    but for any of
    > the physicists I have read who have brought up these ideas, like
    > Carroll and Brian Greene, it is completely clear to me that they
    > consider these to be reductio ad absurdum arguments, not that
    they actually
    > think these are likely to be true. If you disagree, I suggest
    you haven't
    > actually read these authors very carefully, or haven't really
    > what you read.

    Well, I think the passage I quoted pretty much stands on it's own.
    Without the extra assumption of the Past Hypothesis, the data we have
    available leads to a conclusion that isn't cognitively stable when
    combined with the assumption of physicalism.

    I would take this as a reductio ad absurdum argument against

"Physicalism" is much too vague--we can imagine a wide variety of *possible* laws of physics, including ones where the universe would be predicted to maintain a consistent low entropy (I think the old Steady State cosmology would be an example, with new low-entropy matter and energy being continually created in empty space). Carroll's reductio ad absurdum relied on more specific features of the laws of physics in *our* universe, like time-symmetry. But if we're talking about the laws of physics in our universe, it's not as if we already know that these laws would naturally lead us to predict Boltzmann brains would be more common than ordinary observers and so we have to tack on the "Past Hypothesis" as an extra assumption. The point is that since we don't yet know the ultimate laws of physics, we can't yet calculate a probability distribution for histories of the universe/multiverse that would allow us to decide whether ordinary observers are likely to be more common than Boltzmann brains or vice versa. The "Past Hypothesis" basically amounts to the idea that if we *did* know these ultimate laws, they would indeed imply that in an "average" history of the universe/multiverse, ordinary observers will be more common than Boltzmann brains. And Carroll proposes a specific way the laws of physics might work that could make this plausible.

    The eternal recurrence problem is a related, but not identical,
    problem than the issue introduced by the principle of indifference.
    Here Sean invokes probabilistic reasoning on infinite sets, which
    Brent and I are still discussing.  Though I just noticed that we
    accidently wandered off the main list into a private email exchange.

    Anyway.  Onwards:

    > Then on p. 223 he explains in more detail why we can be
    confident we aren't
    > Boltzmann brains: because the level of order we experience is
    far greater
    > than what the vast majority of possible Boltzmann brains should
    be predicted
    > to experience (though he does bring up the possibility that our
    > of an orderly environment could just be a hallucination).

    This was one of the points of my "The 'no miracles' argument against
    scientific realism" thread...which died an untimely death.

    So how does he rule out this hallucination possibility?  Or the
    Boltzmann brain simulator possibility?  What facts do we have about
    the nature of reality that rules it out?

    Another extra assumption.  The "we can trust our observations, even
    though our observations imply that we can't trust our observations"

He doesn't specifically address the hallucination possibility, but I think one could naturally extend his argument to deal with it. His initial argument assumes that a Boltzmann brain would have a properly-functioning sensory system, and that it would be overwhelmingly likely that such a brain would perceive high-entropy surroundings rather than low-entropy surroundings...so, the fact that we *do* see low-entropy surroundings can itself be seen as a falsification of the Boltzmann-brain-with-functional-sensory-system hypothesis. It is true that a Boltzmann brain in high-entropy surroundings might not have a functional sensory system and might instead hallucinate a low-entropy world. But I'd argue that if we consider all possible arrangements of a set of particles that would be sufficiently brainlike to be conscious, we should expect the average observer produced by a random fluctuation would have an arrangement with the *bare minimum* of order needed for consciousness, which I think would be likely to involve much more chaotic perceptions than we actually have (for example, I can still be quite conscious while hearing and seeing only white noise). So I would say there's a good qualitative argument for thinking the Boltzmann brain hypothesis should predict that our perceptions would be hightly chaotic, and this prediction is not borne out by my experience.

Once you take hallucinations as your hypothesis then there's no longer evidence for the existence of particles or brains or universes. Those are all inductive inferences from experiences - but if, ex hypothesi, the experiences are hallucinations then there is no remaining evidence for anything. So it seems to me that supposing our experiences may be solipistic hallucinations is self defeating - nothing follows from such an assumption.


    Quoting the book, page 363:

    "This version of the multiverse will feature both isolated Boltzmann
    brains lurking in the empty de Sitter regions, and ordinary observers
    found in the aftermath of the low-entropy beginnings of the baby
    universes.  Indeed, there will be an infinite number of both types.
    So which infinity wins?  The kinds of fluctuations that create freak
    observers in an equilibrium background are certainly rare, but the
    kinds of fluctuations that create baby universes are also very rare.
    Ultimately it's not enough to daw fun pictures of universes branching
    off in both directions of time; we need to understand things at a
    quantitative level well enough to make reliable predictions.  The
    state of the art, I have to admit, isn't up to that task just yet.
    But it's certainly plausible that a lot more observers arise as the
    baby universes grow and cool toward equilibrium than come about
    through random fluctuations in empty space."

    SO.  I think it's significant that *even with* all of his auxiliary
    hypothesis, he still judges it likely that Boltzmann brains do exist.
    And in such numbers that it's not clear whether they are more or less
    common than "normal" observers.

And again, if we knew the ultimate laws of physics, then with enough computing power and mathematical savvy, the question of which are more common could be *settled* one way or another without any extra tacked-on assumptions.

But the physics would have to predict some relative rate of occurence of things. The answer can't be provided by just finding that this or that is possible and therefore it happens infinitely often. The physics has to provide a measure, like the symmetry of a die or some other measure.

And Carroll gives a qualitative argument about how known physics could be extended in a way that could in fact naturally lead to the prediction that ordinary observers would be more common. So unless you can show that his speculations about future physics are likely to be false, you can't treat the Boltzmann brain possibility as a reductio ad absurdum against physicalism.

Smolin did this with his theory of Darwinian selection for black hole maximization. Although it is probably wrong for other reasons, it did provide an example of how a theory could show our kind of universe to be overwhelmingly probable.

    >> BUT these things all inevitably lead to more questions.  There
    seem to
    >> be only two possible "final" answers:
    >> 1)  Everything exists.
    >> 2)  Reality is essentially arbitrary.  There is no reason why
    >> existence is this way as opposed to some other way.  It just is.
    > Even if "everything exists", there is still the possibility of
    some definite
    > probability distribution on this "everything"--either a probability
    > distribution on all possible universes/computations/mathematical
    > or a probability distribution on all possible observer-moments.
    It's quite
    > possible that the probability distribution would be such that
    observers who
    > had *true* memories of a low-entropy past would be much more
    common than
    > random Boltzmann brains with no memories or false memories.

    Isn't it also quite possible that the opposite is true?

In the abstract, without considering anything about myself beyond the fact that I'm a conscious observer, I would say it's quite possible either is true (though I would say that the argument about the typical Boltzmann brain's experience being much more chaotic than my own should cause me to strongly favor the former possibility as more likely). The question could really only be settled if we knew the ultimate laws of physics and could compute the likelihood of various possible histories. You seemed to be saying that somehow we *already* knew enough about physics to conclude that entropy should have been higher in the past unless an extra logically independent "Past Hypothesis" saying the opposite was tacked on, so that's what I was disagreeing with.


It seems to me this argument keeps mixing two different kinds of problem. One is what kind of laws of physics and initial conditions are consistent with our observations. The other is Hume's problem of induction, how can we have knowledge of anything unless we assume (Hume would say without reason) the future will in some sense resemble the past.


You received this message because you are subscribed to the Google Groups 
"Everything List" group.
To post to this group, send email to everything-l...@googlegroups.com.
To unsubscribe from this group, send email to 
For more options, visit this group at 

Reply via email to