Avignon Day 3: Reductionism

Every academic who attends conferences knows that the best parts are not the formal presentations, but the informal interactions in between. Roughly speaking, the perfect conference would consist of about 10% talks and 90% coffee breaks; an explanation for why the ratio is reversed for almost every real conference is left as an exercise for the reader.

Yesterday’s talks here in Avignon constituted a great overview of issues in cosmological structure formation. But my favorite part was the conversation at our table at the conference banquet, fueled by a pretty darn good Côtes du Rhône. After a long day of hardcore data-driven science, our attention wandered to deep issues about fundamental physics: is the entire history of the universe determined by the exact physical state at any one moment in time?

The answer, by the way, is “yes.” At least I think so. This certainly would be the case is classical Newtonian physics, and it’s also the case in the many-worlds interpretation of quantum mechanics, which is how we got onto the topic. In MWI, the entirety of dynamics is encapsulated in the Schrodinger equation, a first-order differential equation that uniquely determines the quantum state in the past and future from the state at the present time. If you believe that wave functions really collapse, determinism is obviously lost; prediction is necessarily probabilistic, and retrodiction is effectively impossible.

But there was a contingent of physicists at our table who were willing to believe in MWI, but nevertheless didn’t believe that the laws of microscopic quantum mechanics were sufficient to describe the evolution of the universe. They were taking an anti-reductionist line: complex systems like people and proteins and planets couldn’t be described simply by the Standard Model of particle physics applied to a large number of particles, but instead called for some sort of autonomous description appropriate at macroscopic scales.

No one denies that in practice we can never describe human beings as collections of electrons, protons, and neutrons obeying the Schrodinger equation. But many of us think that this is clearly an issue of practice vs. principle; the ability of our finite minds to collect the relevant data and solve the relevant equations shouldn’t be taken as evidence that the universe isn’t fully capable of doing so.

Yet, that is what they were arguing — that there was no useful sense in which something as complicated as a person could, even in principle, be described as a collection of elementary particles obeying the laws of microscopic physics. This is an extremely dramatic ontological claim, and I have almost no doubt whatsoever that it’s incorrect — but I have to admit that I can’t put my objections into a compact and persuasive form. I’m trying to rise above responding with a blank stare and “you can’t be serious.”

So, that’s a shortcoming on my part, and I need to clean up my act. Why shouldn’t we expect truly new laws of behavior at different scales? (Note: not just that we can’t derive the higher-level laws from the lower-level ones, but that the higher-level laws aren’t even necessarily consistent with the lower-level ones.) My best argument is simply that: (1) that’s an incredibly complicated and inelegant way to run a universe, and (2) there’s absolutely no evidence for it. (Either argument separately wouldn’t be that persuasive, but together they carry some weight.) Of course it’s difficult to describe people using Schrodinger’s equation, but that’s not evidence that our behavior is actually incompatible with a reductionist description. To believe otherwise you have to believe that somewhere along the progression from particles to atoms to molecules to proteins to cells to organisms, physical systems begin to violate the microscopic laws of physics. At what point is that supposed to happen? And what evidence is there supposed to be?

But I don’t think my incredulity will suffice to sway the opinion of anyone who is otherwise inclined, so I have to polish up the justification for my side of the argument. My banquet table was full of particle physicists and cosmologists — pretty much the most sympathetic audience for reductionism one can possibly imagine. If I can’t convince them, there’s not much hope for the rest of the world.

92 Comments

92 thoughts on “Avignon Day 3: Reductionism”

  1. I think it is a false dichotomy to separate philosophies into “reductionist” and “anti-reductionist” camps. It seems to rely on the assumption that each layer of the “fundamental laws” become “simpler”. Historically this may be true, but I think extrapolation is unjustified.

    There is the whole thing about current known laws being ’emergent’ — I don’t have a strong opinion either way, but I am ready to keep an open mind about this…

  2. “If you believe that wave functions really collapse, determinism is obviously lost; prediction is necessarily probabilistic, and retrodiction is effectively impossible.”

    Sorry for the noob question here, but is it possible that wave functions really collapse, and what we see as probabilistic is actually just an incomplete picture of laws we don’t know yet?

  3. Thanks Mike. So you say its random which reality I observe? I think there is quite a big gap here which it doesn’t seem like there is an answer for. How do you know it really is random, and not directed by some unobserved control mechanism?

    Fascinating stuff.

  4. Dave,

    You could always say it was directed by some unobserved control mechanism, or by spirits or a God, or by a programmer and that we live in a matrix.

    The main problem with these types of explanations is that they can be used for explaining any crackpot theory one could possibly posit. My favorite non-explanation explanation is: “the wizard did it”. Take a look at this brief video from a recent TED conference where Deutsch addresses just this point:

    http://www.youtube.com/watch?v=folTvNDL08A

    Let me know what you think.

  5. When discussing issues like this, maybe physicists need the language of Kolmogorov complexity, computability and algorithmic information theory, or what I like to call the “computational cosmos” models. According to these models, a theory of everything is the algorithm with the shortest description length (smallest Kolmogorov complexity) which outputs the observed state of the universe. Essentially it is a mathematical formalization of induction using Ockham’s Razor. Marcus Hutter has an interesting paper about TOE’s using this approach: http://arxiv.org/abs/0912.5434

  6. If one accepts MWI, then there must be many distinct ways for a universe to branch from any given state to an identical subsequent state (with higher entropy, so there’s no contradication there). But that seems to contradict the idea that the wave function of the universe (assuming it exists) could be tracked back by the Schrodinger equation linearly, and hence uniquely, from the later state to the earlier state.

    I suppose the question becomes what constitutes “the universe” in this context? I’d say it’s the union of all causally connected patches bounded by an anti-de-Sitter region, which of course from our standpoint must be incomparably larger than the observable universe (if that makes sense).

    Regarding the wave function, has anyone tried going to the opposite extreme of Everett and postulating that it’s apparent smoothness is merely the net effect of a host of constantly varying “spikes” for want of a better word. That would be somewhat analogous to an amplifier’s sound level LEDs which, despite being discrete oscillating bars close up, collectively resemble from a distance a smoothly evolving curve.

    The temporary vanishing of a spike would represent a measurement of the system, and this would be either “internal”, while the system remained closed, or a conventional measurement by some other interacting system in the vicinity. Either way, the wave function could continue on its merry way, maintaining the same basic nature, without having to wear two hats so to speak (either smoothly evolving or momentarily vanishing)

  7. fco– We don’t think that’s possible, although of course it’s hard to be sure. You would need a “hidden variable theory,” and those seem unsatisfactory for various reasons. But it’s okay to keep an open mind.

  8. I don’t think that reductionism ‘in principle’ is really consequential. For sciences like biology and economics. I think there is quite a gap here. I agree with Sean also.

  9. Roughly speaking, the perfect conference would consist of about 10% talks and 90% coffee breaks; an explanation for why the ratio is reversed for almost every real conference is left as an exercise for the reader.

    Bladder control?

  10. Am I allowed to observe myself and collapse my own wavefunction?
    Is a proton allowed to observe itself and collapse its wavefunction?
    If the answer to question 1 is yes, and 2 is no, do we not have a difference at macroscopic compared to quantum levels?

  11. I would think that laws are laws, and the reductive vs. emergent distinction is irrelevant.

    If the fundamental laws are deterministic and the higher level laws are deterministic, then the entire history of the universe is *still* determined by its exact physical state at any one moment in time.

    This is true because the deterministic emergent laws are going to “emerge” predictably from the state of the more fundamental lawyer. So they will always act similarly in similar situations.

    Even if you switch from deterministic laws to probabilistic laws, the reductionist vs. emergent distinction is still irrelevant. The emergent laws just emerge “probabilistically” from the state of the more fundamental layer.

    Emergent laws are just another layer of what you already have…not something qualitatively different.

    I’d go further an argue that determinism vs. “probabilism” really isn’t that important either. In fact, determinism is a special case of probabilism…the case where all probabilities are either 0% or 100%.

    Maybe?

  12. On a different but interesting subject — there is word out of the LHC regarding the Higgs. Still internal, but interesting (via Peter Woit’s blog):

    Internal Note
    Report number ATL-COM-PHYS-2011-415
    Title Observation of a γγ resonance at a mass in the vicinity of 115 GeV/c2 at ATLAS and its Higgs interpretation
    Author(s) Fang, Y (-) ; Flores Castillo, L R (-) ; Wang, H (-) ; Wu, S L (University of Wisconsin-Madison)
    Imprint 21 Apr 2011. – mult. p.
    Subject category Detectors and Experimental Techniques
    Accelerator/Facility, Experiment CERN LHC ; ATLAS
    Free keywords Diphoton ; Resonance ; EWEAK ; HIGGS ; SUSY ; EXOTICS ; EGAMMA
    Abstract Motivated by the result of the Higgs boson candidates at LEP with a mass of about 115~GeV/c2, the observation given in ATLAS note ATL-COM-PHYS-2010-935 (November 18, 2010) and the publication “Production of isolated Higgs particle at the Large Hadron Collider Physics” (Letters B 683 2010 354-357), we studied the γγ invariant mass distribution over the range of 80 to 150 GeV/c2. With 37.5~pb−1 data from 2010 and 26.0~pb−1 from 2011, we observe a γγ resonance around 115~GeV/c2 with a significance of 4σ. The event rate for this resonance is about thirty times larger than the expectation from Higgs to γγ in the standard model. This channel H→γγ is of great importance because the presence of new heavy particles can enhance strongly both the Higgs production cross section and the decay branching ratio. This large enhancement over the standard model rate implies that the present result is the first definitive observation of physics beyond the standard model. Exciting new physics, including new particles, may be expected to be found in the very near future.

    See: http://cdsweb.cern.ch/record/1346326?

    Peters take is as follows: “A commenter on the previous posting has helpfully given us the abstract of an internal ATLAS note claiming observation of a resonance at 115 GeV. It’s the sort of thing you would expect to see if there were a Higgs at that mass, but the number of events seen is about 30 times more than the standard model would predict. Best guess seems to be that this is either a hoax, or something that will disappear on further analysis.”

  13. Jason Dick (#10), I find your argument somewhat persuasive. However, it seems to rely on the assumption that it is always meaningful to speak about the individual behaviors of the particles that make up a macroscopic system, and moreover that the combined descriptions of the behaviors of all of these particles describe all of the meaningful facts about the system as a whole.

    What about the case of an entangled state? If I have even two particles in a state like |0>|0> + |1>|1>, I can’t tell you whether either particle is in state |0> or state |1>, or any other state vector for that matter. I suppose I could describe the state of one of the particles by its reduced density matrix, but the phase difference between the two terms in the two-particle state isn’t a property of either particle, but rather of the system as a whole.

    I’m still sympathetic to reductionism, but I think it would be a lot easier to make the argument if we lived in a world of classical particles.

  14. If you forced me at gunpoint to decide between reductionism and non-reductionism, I would probably choose reductionism. That said, I think the debate is fairly pointless if the central concern is about what is possible in principle. Because we’re talking about the empirical/phenomenal world, here. In principle, in theory, hypothetically—these phrases don’t get us anywhere. You still have to do the work to show that one side I right and the other is wrong.

    Consider Boyle and the air pump. Was it possible to create a true vacuum? This was a question people debated in Britain ad nauseum. People would get into bar fights about it, posing the most sophisticated objections both against and in favor of each side. Nature abhors a vacuum, they said. The concept of a vacuum was philosophically nonsensical. But the fact of the matter was that people just didn’t know, at least not until Boyle went out and did the work necessary to prove it.

    Can human activity be reduced to the laws of quantum dynamics? I doubt it. Can it be reduced to some form of physical laws? That seems much more likely. But certain? No. Who knows what kind of curveballs the phenomenal world might throw at us? The fact of the matter is that we won’t know until someone goes out there and does the work to prove it one way or the other, empirically—i.e. we won’t know for a long, long time. I guess, what I’m saying is “talk away if you want,” because it’s going to be a while before this debate is actually settled.

  15. In discussing reductionism and emergence, I think it is crucial to consider both the underlying laws and the particular configuration on which those laws act. As Sean and others have pointed out, whatever macroscopic laws exist should be fully consistent with the underlying microscopic ones (otherwise it is hard to see how one could meaningfully distinguish the microscopic and macroscopic laws — they would all just be “laws” with equivalent standing).

    However, microscopic laws need not imply a particular set of macroscopic laws. A good example is biology: the biological laws we know all govern organisms on Earth, but those organisms (indeed the existence of DNA) are products of both the microscopic laws and the initial conditions (Earth and its properties) plus boundary conditions (sun, moon and their relationship to Earth). I doubt many would argue that the microscopic laws imply that we have our particular Earth, sun and moon; otherwise, every star would have a solar system that looks just like ours. Hence, we must consider that our biology, and by extension our biological laws, largely depend on chance events (initial configuration of matter) that cannot be deduced from the laws alone.

    (Sure, one could find fault with the biology example and note that certain laws seem “universal” — natural selection and random mutation are obvious examples if one requires life to be self replicating in order to call it “life.” But in our ignorance we cannot arbitrarily rule out the possibility of other kinds of “organisms” that display some of the characteristics we would ascribe to life, such that random mutation and natural selection aren’t the primary laws that govern their existence. )

    I think Sean and others are thinking of reductionism primarily in terms of the underlying laws, i.e., QM and GR. Perhaps those who argue in favor of a different set of macroscopic laws are really arguing for emergent laws, i.e., laws that are consistent with, but not implied by, microscopic laws. That is, the disagreement may stem more from misunderstandings, possibly because the parties involved have not have carefully stated exactly what they mean by “macroscopic laws.”

    On the other hand, if there are serious physicists and cosmologists who think there are macroscopic laws that are inconsistent with microscopic ones, then I really wouldn’t know what to say to those people… It certainly seems weird to think that a law could appear out of nowhere which governs a large collection of particles whose microscopic behaviors are determined by microscopic laws, and yet is not itself constrained by microscopic laws.

  16. “Note: not just that we can’t derive the higher-level laws from the lower-level ones, but that the higher-level laws aren’t even necessarily consistent with the lower-level ones.”

    Sean,

    If your interlocutors were arguing that higher-order processes are not consistent with lower-order processes (that is, that emergents do not strictly supervene upon their basal realizers) then they are not arguing for non-reductive physicalism in the sense excepted by the majority of modern philosophers.

    On the other hand, if you are suggesting, as you seem to be, that emergents are not derivable (that is, neither fully explainable nor predictable) without remainder from their basal realizers, then you are arguing for non-reductive physicalism (that is, for emergence) in its modern sense.

    This is a highly contentious area of philosophy at the moment with incredibly sophisticated arguments on both sides relating to causation, individuation, functionalism, basic ontology, multiple realizability, etc.

    The general intuition is that the subjects of the special sciences (biology, psychology, economics, etc) are not reducible without remainder to descriptions of elementary particles and their relations. If this were the case, then we would seem to be forced into some form of eliminativism regarding most of our common sense beliefs.

  17. Arguments such as these are why so many people, often ESPECIALLY other academics, hate physicists (especially theoreticians) so much. You could have just the same discussion at home, come to the same conclusions (read “none”) except that a) your lunch mates might not be as famous b) the wine and architecture probably not as interesting. You could probably have the same disucssion in the uni cafeteria with any randomly chosen first year philosophy major….

  18. Rob Knop wrote:

    …the MWI vs. Copenhagen thing becomes something of a red herring.

    Just like Communism.

    I think one could be a consistent-historian like Gell-Mann or Hohenberg, a latter-day Copenhagener like Peres, a relationalist after Rovelli, a correlationalist after Mermin or a QBist in the tradition of Fuchs while still accepting the idea that emergent phenomena must be consistent with the underlying laws on which coarse-grained descriptions supervene.

  19. I am so out of my element 🙂
    Just wondering how Heisenberg’s Uncertainty Principle fits into the picture?

  20. I am definitely in agreement with Sean on reductionism being the ultimate way to understand reality. Braden, as far as the paper “More really is different” is concerned, it really doesn’t damage the case for reductionism at all. The fact that some certain things are considered undecidable in computation doesn’t mean that reductionism is false, and the fact that it is an infinite system automatically negates the results conclusions for our finite observable universe. Any scientist that thinks reductionism is wrong is going against a much more fundamental notion of causality, as reductionism is squarely tied to the notion that every cause precedes an effect. Theoretically, one could deduce the rest of the universes history from initial conditions and the laws of physics.

  21. Albert Zweistein

    Alright you Strict Reductionists – Time for a test.

    1. Explain in words, pictures and/or diagrams the well-observed phenomenon of human ontogeny (consult Wikipedia if you draw a blank).

    2. Now explain in full detail how ontogeny unfolds using only theoretical atomic and subatomic physics.

    Extra credit: Explain the remarkable observation that ontogeny recapitulates phylogeny using only QCD.

    Good luck you wild and crazy Platonists! You cannot possibly imagine how eagerly I await your blue books, or the 10^500 of them you will need for your attempted answer.

    High marks for those who realize early-on that they are on a fool’s errand.

    Best,
    Albert Z

  22. “every quantum event is inherently indeterministic”

    I have never been able to buy this. Everything is determined, though some events are unpredictable.
    In the case of rain, we can’t predict where each drop will fall; but if we had sensors that could detect each drop of rain and it’s trajectory, as well as information on the wind strength at every point in the air, and a detailed map of the ground, we could say exactly where they would all end up. Practically, it’s impossible. In principle, it’s unpredictable yet deterministic.

    The same goes for any event on the macro or microscopic scale. We would need technology that we can’t even conceive at this point in time to measure subatomic particles to a degree that we could predict what they would do at any future time. Maybe that technology is in practice impossible to achieve. But in principle, if we had the information, we could predict any event. The fact that we can’t predict it doesn’t make it indeterministic.

  23. Daniel,

    On trajectories, Bernard d’ Espagnat has an interesting discussion of the quantum mechanical treatment of “particle traces” seen in cloud chambers in “On Physics and Philosophy”, pg. 95:

    “At first sight, as we noted, these alignments of bubbles seem comparable to the white trails produced by a jet plane in a blue sky. Hence, when the ‘particle’ source is external to the chamber (or the emulsion), we not only attribute to each ‘particle’ inside this device a well-defined trajectory (coinciding with the trace) but also do not hesitate to continue the latter to the rear, by thought, up to the particle source. […] However, as we also noted, such a picture does not fit with quantum mechanics (nor, incidentally, with the Broglie-Bohm model, in which the corpuscles continually undergo deviations dictated by the whole-universe wave function). […]

    The true explanation for the observed alignments is not, therefore, to be looked for within the realm of such ideas, great as may be the force with which our intuition puts them forward. It essentially lies in the fact that, when the initial conditions are sufficiently known quantum mechanics makes it possible to predict what will be observed. It does this, as we know, by introducing mathematical symbols that were given names (wave function, state vector, etc.) and many of them evoke some picture. But, to repeat, the pictures thus called up are unreliable ones and play no role in the calculations.

    What quantum mechanics in fact yields are merely the probabilities that, for a given initial flow, microblobs will be observed at such and such places within the device. And, as already noted, the probabilities of concerning the cases of the blobs being aligned along the general direction of motion are considerably larger than those relative to any other configuration. In other words, what quantum mechanics predicts is just that, within the device, we shall see alignments (of microblobs or bubbles) consistent with what we actually observe and naively interpret as being ‘traces’.”

Comments are closed.

Scroll to Top