Quantum Darwinism!

Starting yesterday afternoon, I decided to begin doing some research into quantum physics and the Many-Worlds Interpretation thereof in preparation for Star Trek: DTI — since a book about time travel in the Trek universe has to deal with parallel universes to some extent.  True, the portrayal of temporal physics and alternate realities in ST is quite fanciful, but as my readers know, I like to do what I can to ground ST’s fanciful science in principles from real science.  I’m often surprised at how feasible it is.

I took quantum mechanics in my final year as a physics major in college, but it never really clicked with me.  Part of it was that the calculus was just too complicated for me, but I think part was that it was all so abstract.  I was interested in the concepts, the meaning, but all the class covered was the math.  I recall describing it to a friend as “variables performing unnatural acts on each other.” Nowadays, I still can’t follow the math in any detail, but I find it easier to study the subject now that I’m motivated to use it in fiction.  In fact, I got so caught up in it last night that I barely got any sleep.  It helps that I’ve found some good articles, starting on Wikipedia and including sites linked to in its articles on quantum theory.

So anyway, the basic idea of quantum physics is that a particle or other system can exist in multiple states at once — a superposition, it’s called.  The question is, how does that produce the classical world we observe where  everything appears to have one definite state?  The old idea, the Copenhagen interpretation, was that the wavefunction “collapses” into a single state when it’s measured.  Copenhagen didn’t explain how or why this happened; it was an ad hoc postulate that nobody was really happy with.  (The Schroedinger’s Cat thought experiment, often cited as an illustration of this phenomenon, was actually a critique of it, because it argued that there could be a scenario wherein a macroscopic object like a cat was forced into an impossible dual state because of the superposition of a single atom — the decay or nondecay of that atom determines whether or not the poison is released.)

These days, we understand it as a process called quantum decoherence, which is nicely discussed at this site.  The key is that we don’t really observe a particle directly; we observe its effect on the environment it interacts with.  If you measure the state of a particle, you’re actually observing the state it’s induced in the measuring device.  So the wavefunction’s multiple states never actually collapse into one; rather, they’re all still there, but we, as part of the environment interacting with the particle, only observe one state at a time.

So the idea that Schroedinger’s Cat is both alive and dead until the box is opened by an observer is wrong.  When the atom’s two states (decayed and undecayed) interact with the poison trigger, the interactions with the billions of particles in that trigger cause decoherence, isolating the states’ effects from each other, so the trigger will only react to one state at a time.  So the behavior becomes “classical” through the interaction with the trigger, and when the trigger interacts with the poison and the poison interacts with the cat, they all become part of that same combined quantum system and share the same single state.  The cat survives or dies before the observer opens the box, because the decoherence happens at the trigger.

This is why it’s a mistake to do what so many laypeople do and interpret wavefunction collapse mystically, as the result of observation by a thinking being.  In fact, observation is just one type of interaction between the local system and its environment.  Any such interaction causes decoherence, once the effects of the interaction propagating through the environment become thermodynamically irreversible (like a glass shattering — it’s all but impossible to make the atoms revert to being an intact glass and jump off the floor onto the table).

This idea was proposed in 1957 by Hugh Everett, in what he called the relative state formulation: when a system interacts with its environment, its state can no longer be described in isolation, but only as it exists relative to the environment it interacts with.  So the environment becomes correlated — or quantum entangled, in modern terminology — with the state of the particle.  But each of the multiple states of the particle is separately entangled with its environment, describing a separate system.  Instead of just the particle having a wavefunction that’s a superposition of states, the combined system of the particle and environment has a superposed wavefunction, the whole schmeer existing in two or more states at once.  And once the difference between those states becomes irreversible, they stop interfering and have no more interaction.  From that point on, the system has multiple independent histories.

This has become known as the Many Worlds Interpretation, and that’s often taken literally: each independent measurement history represents a parallel reality.  One world where the cat lives, another where the cat dies.  Naturally, this is the interpretation that applies in the Star Trek universe, with all its parallel realities.  Here’s a really good discussion of MWI, what it means, and what it doesn’t mean.  But although MWI is increasingly accepted among physicists as a mathematically valid approach to quantum physics, not too many of them believe that the other “worlds” literally exist as parallel universes; rather, they consider the alternative measurement histories to be simply alternative possibilities that are present but swamped within our singular reality, states that exist in a mathematical sense but don’t really split off into alternate universes.  This is the view I favor when I’m not writing a work of fiction that requires using MWI.

But either way, there are still questions about the actual physical mechanism behind decoherence.  What causes one state to dominate in the macroscopic system while the others fade away (or get shunted off into other realities)?  This is where Wojciech Zurek and his theory of Quantum Darwinism come in.  It’s pretty much just what it sounds like.  A Darwinian evolutionary process can happen — indeed, must happen — in any system that meets three conditions: 1) It has reproducing entities; 2) the reproductions are not exact; and 3) the environment favors the reproduction of some traits over others.  In such a system, some entities will have traits that let them reproduce more successfully than others.  That means there are more and more of them with each generation, until they inevitably overwhelm the competition.  (This is one of the many reasons why creationism is such BS.  The basic mechanism of evolution is so simple as to be inevitable.  Not only does it happen, but there’s no way to prevent it from happening in any system that meets those three simple conditions.)

In Quantum Darwinism, what’s being “reproduced” is information about quantum states.  The information about the original particle is encoded in the states it induces in the other particles/systems it interacts with, so each particle’s state becomes an “offspring” of the original state.  Basically, the process selects for states that can survive the decoherence process.  The states that survive are ones that are stable enough to survive interaction with other particles and thus get “copied” over and over by multiple interactions, so that they’re encoded redundantly throughout the environment.  Unstable states may be copied once or a few times before being destroyed, so their information doesn’t propagate as far.  The larger the environment becomes (i.e. the further the information spreads out into the universe), the more dominant the redundant information gets, swamping out the alternate states.

This is why reality looks classical.  We measure an object by measuring the environment it’s interacted with, and different observers measuring different parts of the environment wil see the same redundant information and agree on the reality they observe.  The original particle is still in multiple states, but the information about those other states has been swamped because it didn’t get reproduced redundantly enough.

So instead of a vast number of parallel realities, all carrying equal weight, what you have instead is a “signal” of classical reality on top of a faint “background noise” consisting of the unfulfilled potentials of all the other possible outcomes.  Which is how the universe can appear classical even while being entirely quantum-mechanical.  It’s not perfectly classical, which would require infinite redundancy, but it’s close enough to look that way for the most part.

Which doesn’t mean that Quantum Darwinism is incompatible with Many-Worlds.  It’s actually derived directly from Everett’s assumptions.  But as I said, it’s an open question whether MWI can be taken literally, whether the “worlds” are objectively real or just mathematical constructs.  Zurek doesn’t take sides on the question, and he says the following:

Objective existence can be acquired (via quantum Darwinism)
only by a relatively small fraction of all degrees
of freedom within the quantum Universe: The rest is
needed to “keep records”. Clearly, there is only a limited
(if large) memory space available for this at any time.
This limitation on the total memory available means that
not all quantum states that exist or quantum events that
happen now “really happens” — only a small fraction of
what occurs will be still in the records in the future.

If I’m reading this right, Zurek seems to be saying that Quantum Darwinism allows for there to be more than one “real” history to the universe, but rules out the interpretation of MWI stating that every possible outcome must be equally real.  So there could be a finite number of parallel timelines — maybe just those robust enough to stand out from the noise.  That fits the Darwinian paradigm, since a successful species can split apart into multiple coexisting species.  Not every genetic variation or mutation spawns a whole new species, but the most reproductively successful ones generally do.

That’s an interpretation I think I can live with.  The idea that every possible reality is real, that I’m splitting off at every instant into thousands of alternate selves, is one I find inelegant and kind of disturbing.  And from a dramatic standpoint it’s undesirable; if every decision actually happens more than one way, then any story’s outcome is arbitrary and meaningless.  However, if the number of realities is greater than one but still limited, it’s not so bad.  The number could still be quite large, though, large enough to accommodate the myriad universes (to coin a phrase) seen in Star Trek.

Note also that Zurek seems to be saying that a “real” alternate history won’t necessarily remain “real” forever.  The information could be lost, that state of the universe wiped from the cosmic memory.  That has interesting ramifications from a fictional perspective, particularly where a book about time travel is concerned.  And that’s all I’m going to say on the subject for now.