Friday, March 21, 2014

This be the multiverse

(Totally stole that title from Sarang.)

So, I've been having a three-day debate with a Twitter pal (James) about morals in the multiverse. It started with these tweets:
I was actually referring to the gravitational waves detected at BICEP2, which lend support to an inflationary model of the universe, where our universe is just one bubble in a larger multiverse. But this quickly segued into a discussion about the possibility of ethics in a multiverse where every possible universe exists. (There are viable models of reality where this could be the case: one a way of interpreting quantum mechanics, another the result of a truly infinite universe.)

James was/is worried that in such a multiverse, morals would be impossible:
For example, if you see a drowning child, why save it, because in some universe, you don't save it, and it's all the same anyway (or something to that effect.)

This reminds me a lot of the philosophical argument that comes up from time to time in regards to free will, i.e., if there's no free will, how do we punish criminals? This argument drives me bananas! This is basically philosophers thinking "there's no free will" means "criminals don't have free will, but I do."

I think I finally came up with the right metaphor to describe what's wrong with this kind of thinking, and it has to do with levels of accuracy — significant digits, if you will. Basically, moral decisions (any kind of decision, really) happen at the level of our immediate, experienced reality. You might need Newtonian physics to make some of your human decisions, but this isn't a use case where you're going to need to involve quantum mechanics or general relativity. (Unless you're literally a rocket scientist and doing the math on satellite orbits.) You don't need to get down to the level of accuracy of an atomic clock to decide whether or not you're going to save a drowning baby.

When we talk about whether or not we truly have "free will" in the pure sense, or if we're living in some version of a multiverse, we're speculating about a very fine-grained level of the "underlying reality." We're going to an extreme, ultra-high-resolution level of accuracy, way way over to the right in terms of decimal points of accuracy. And the thing is, when you get that accurate, human concepts like "free will" and "morals" just don't exist. We're reduced to probability clouds. So worrying about morals in an infinite multiverse is like worrying about how your complexion looks under an electron microscope.

I see this kind of wrong-headed thinking a lot, so I kind of want to coin a term for it  — something like "The Resolution Fallacy." Anyone know if someone else has already formulated a statement of this kind?

18 comments:

  1. These ideas come up in Dennett's writings on free will and on consciousness. When we take the "intentional stance" towards something in the world, we are employing a particular level of description -- certain features of the world are efficiently represented as agents, because of certain regularities in their behavior over time -- even if in principle they can be represented more exactly as e.g. "probability clouds." We represent ourselves and other people this way. The descriptions really fit, more or less, or the strategy wouldn't work. This allows us to ignore all the low-level implementation details in thinking about ourselves and other people. A good starting point might be: "True Believers: The Intentional Strategy and Why It Works." http://mind.ucsd.edu/syllabi/06-07/Phil285/readings/true-believers.pdf

    Some of this analysis can be traced back at least to Kant. Kant understood that, on a scientific view of the universe, all of our actions are determinate. We can't really be said to be making decisions at that level of resolution. But we are forced to take ourselves to be free choosers anyway: "every being which cannot act in any way other than under the idea of freedom is for this very reason free from a practical point of view" (Ak. 4:448). We take a perspective of freedom that is apart from the scientific perspective of universal cause and effect.

    But maybe there is a valid worry about how the low resolution where we live can be built up out of the high resolution of microphysics (or whatever). It seems like the larger scales should be reducible to the smaller in some sense. In that case, we might want to know where, in the move between small and large scale, the secret freedom juice comes in.

    ReplyDelete
    Replies
    1. Yes, I'm totally down with Dennett's intentional strategy stuff. Semi-related to the idea that there's no coherent definition of "life" but it's a useful concept for us anyway. (And before ever reading any Kant, I had the same theory intuitively.)

      Secret freedom juice! I guess I think what's interesting there is why we have the illusion of freedom juice, from inside ourselves. Theoretically, as an outside and highly accurate observer, we don't need freedom to explain a person's actions anymore than we need it to explain an asteroid's. We just don't currently understand humans as well as we understand asteroids.

      Delete
    2. Yes, some reduction problems are a lot harder than others but it does not seem that this ought to pose any philosophical difficulties. (A more general point, I think, is that it is usually wrong to say "X is an illusion": our knowledge of the external world is resolution-limited, and there are ontologically utterly different theories that agree to within experimental resolution. What seem like coherent objects at a given resolution are not so at higher resolution, and the relation between the two scales needn't be simple. All scientific objects are illusions, but some illusions are more useful than others. And I guess this is why I am neither persuaded nor disturbed by the "many-worlds" aspects of quantum mechanics...)

      Delete
    3. Yes it makes sense to put it that way, perhaps more accurate to say it's a construction than an illusion, wherein everything we perceive is a construction based on our limited abilities of perception.

      Delete
    4. Yes, I have been reading Pierre Duhem, an OG physicist who stubbornly denied the reality of molecules (of course this was not a quirky position until late in his life). He makes the point that whatever the underlying "metaphysics" of the world are, they are permanently inaccessible to us -- they are always below "experimental resolution." Accordingly, scientists should never try to offer "explanations" -- stories about the world in terms of supposedly fundamental objects.

      Delete
    5. I wonder if he was also against explanations of poems.

      Delete
    6. I think one should think of scientific stories as Hilbert did about mathematics: as complicated but hopefully internally consistent structures that agree with reality where the two intersect. Or as Woolf did about fiction: as a spider's web that is largely free-floating but "attached to life at all four corners."

      Delete
  2. Incidentally, I find Eliezer Yudkowsky close to unreadable, but I kind of wish I didn't. I believe "How An Algorithm Feels From The Inside" is broadly about these issues. Some of his analysis of decision theory problems is also in a similar vein, I think.

    ReplyDelete
    Replies
    1. I used to have kind of a crush on Eliezer Yudkowsky but haven't read anything he's written in years.

      The "what if a tree falls" example from the beginning of that post is very much like most free will arguments; people don't agree on what free will means before they start arguing.

      Delete
  3. Comment from my buddy Jeff Eaton who does not have a Google account (!) and therefore cannot comment here:

    You are taking the exact right approach to dealing with the rabbit hole James has fallen down. I would have made a comment exactly like Grobstein, including a Dennett reference, had s/he not beaten me to it and said it even better than I would have (dat Kant quote!).

    This concept comes up a lot in debates about "reductionism" and the "hard reductionists." They aren't debates so much as pedantic nitpick fights about the appropriate resolution to view a problem. "It's all physics!" "But that's not relevant!" Dennett is in favor of using the level of resolution that yields the most predictive models. Sometimes you to make a leap of faith, e.g. leaping from atoms to neurons, from individual neurons to networks carrying out computations, all the way up to I WANT THE COOKIE. But if the leap yields awesome, useful, replicable results then it was an ok move.

    James is also making a separate bad move in arguing for/against moral theory based on physics. Physics deals with what is and is not; moral theories deal with what you ought and ought not do. Many have tried, but no one has done a good job arguing from an is to an ought pretty much ever.

    http://en.wikipedia.org/wiki/Is%E2%80%93ought_problem

    ReplyDelete
  4. I like that: "you cannot argue from is to ought" nor from aught to naught (we are here, to argue from nowhereness is disingenuous and, more importantly, inaccurate).

    I agree wholeheartedly with but base my choice of microscopy on my body.

    Embodiment as I understand it does a pretty good job defining the limit... (in this case, to use your metaphor, the degree of resolution). My consciousness is bounded-by/related-to this universe, more specifically, my body in this universe.

    The bodies of other universes may exist but I do not belong to them ("I am my body") and what I do/should do/will do is tethered to what is proximal/available to my senses... because all else is conjecture (and contra kant--I believe my ability to reason was developed in inextricable concert with those senses and thus *always* tethered--to recognize the formative relationship is key).

    I can extrapolate but only as far as I have examples that serve as hypothetically relevant analogic sources. I cannot extrapolate infinitely because *I* am not infinite. I must start with proximal moral action (founded on observable results) and ripple out from there using all faculties available to me. When encountering what is foreign, I can and do make mistakes that become part of my lived experience and thus can be learned from.

    Lived experience as the rule. Spooky action at a distance: caring minute-ly does not have to mean caring small-mindedly or short-termedly. Technology and reason may extend the body but cannot transcend it, nor erase it. Only occasionally blind us to its finite persistence.

    ReplyDelete
    Replies
    1. Yes, I think scope is crucial when thinking about morality. I think ethics are more complicated now than they were, say, 1000 years ago, because we know so much, and economies are more connected...we like to think we can make perfect moral decisions but the consequences of our actions are unpredictable. Like when we formulate policies to benefit the environment and end up making things worse. Basically the systems are already SO complex, we're already bad at morality, why complicate things further by worrying about stuff that can't possibly affect "us" or vice versa.

      Delete
  5. One of the most significant limits on how we (human beings) perceive the universe and make choices (to whatever extent we can be said to make them) is that we, including the consciousness of each of us individually, exist in three dimensions, not four. We appear to be "moving" or developing through time. This even though, considered from a 4th-dimensional standpoint, time and space are the same thing and everything happens, has happened, will happened, all at once.

    So in the 4th-dimensional sense, everything is already determined, because everything has already happened (or is happening, or will happen). But from the standpoint of our third-dimension consciousness (putting aside for the moment questions of what consciousness is), we don't know yet what's going to happen in the "future."

    What may or may not happen (or has already happened or not) in one of the other possible universes in the multiverse probably isn't relevant to any choices we make in this universe where we apparently exist. I say "probably" because, following some of the threads of quantum physics, it's always at least theoretically possible that something that happens in another separate universe may in some way affect what happens in this universe, the possibility is always part of the "probability cloud;" but, the probabilites lean so heavily against this that it's likely that all possibility universes will live out their entire life spans and be gone, before any interaction between any of them occurs.

    Questions of morality aren't ever really purely theoretical. And even if you insist on a predetermined universe, where no one really has free will in any ultimate sense, okay, fine, but this "predetermined" universe then has developed in such a way that we find ourselves faced with choices about moral questions, whether real or only apparent.

    In fact, as I'm understanding it, if all possible universes exist, then it should also follow that all possible sequences of events within a single universe should also exist. So that may in fact mean that we do in fact have free will, within whatever limits surround it -- that a "predetermined" universe isn't really predetermined as accomplished fact, but only "predetermine" as a probability cloud, and various of the possibilities within the cloud will play out partly in response to moral choices we make as human beings. We ourselves, our conscious choices, are part of the probability cloud in all its possibilities.

    Coming at this from another direction: one of the examples I come to when I get into conversations about free will and moral choices and the nature of reality, etc., is: suppose you're bleeding badly from a cut on your arm. If you don't do something, you'll bleed to death before very long. Do you call 911? Or do you just wait and see what happens, because we exist in a multiverse?

    ReplyDelete
    Replies
    1. Just one note of clarification -- you don't have to believe that everything is "predetermined" in order to give up free will. Even if everything is completely random and unpredictable, that doesn't mean we have free will in the pure sense.

      Delete
    2. I ended up writing a book about this on my old blog, in which I hopefully don't come off as too big for my armchair: http://storyofalanslife.blogspot.com/2014/03/the-multiversal-sigh.html Would of course appreciate any thoughts.

      Delete
    3. I guess I'm of the opinion that plenty of philosophical worries ARE silly. You didn't really argue me out of my position that this kind of thinking is basically level-mixing -- taking some bits from one level of reality and mixing them in with bits from another level of reality and trying to make them play nice together.

      Delete
  6. Your fate is predetermined by what you do with your free will.

    ReplyDelete