philosophy and physics
Jonathan Bain interviewed by Richard Marshall.
Jonathan Bain is an ice cool philosopher of physics. He broods on how intuitions are challenged by new theories, about Newtonian gravity, about whether physicists are running blind and the implications for philosophers, of whether we should be realists or anti-realists, about the nature of space and time, about the nature of Lee Smolin’s doubts about physics, about what the electron teaches us, about particles and fields, about structural realism and whether there can be different empty universes, about why Albert won the Krauss vs Albert spat and what to make of the multiverse. That’s one hell of a motherload! Keep it 100 yawl!
3:AM: How did you become a philosopher?
Jonathan Bain: I got interested in philosophy as an undergraduate physics major. I would get burned out on physics courses and then compensate by taking philosophy courses. My original plan was to go to graduate school in physics, but during my senior year I found a copy of Michael Friedman’s ‘Foundations of Spacetime Theories’, which seemed incredibly interesting despite the fact that I could barely understand it. This led me to graduate school in history and philosophy of science, where I discovered that philosophy of physics was alot more interesting than physics itself, and that you could actually do philosophy of physics as a living.
3:AM: You’re interested in working out how we should understand and interpret theories of physics. So I guess we’re going to have to see if we can get a grip on a whole host of stuff. But before looking at details, do you think that with quantum mechanics and relativity, contemporary theories about physics are more difficult to understand than previous physics, or has it always been a problem? It just seems a tad easier to relate experienced reality to a Newtonian universe, but maybe that’s just my ignorance talking?
JB: There’s a sense in which this has always been a problem, at least to the extent that intuitions informed by past theories may sometimes conflict with proposals about how theories that supersede them should be interpreted. One might claim that there’s a “lag time” between intuitions and theories in physics.
One way to see this is to note that for everyday, macroscopic, slow-moving objects (like cats, cars, and asteroids), Newtonian physics, relativity, and quantum mechanics all make exactly the same predictions (for all practical purposes). They only disagree about microscopic objects (like electrons and photons, say), and fast-moving and/or large-scale, megascopic objects (like particles moving in a particle accelerator, or black holes). Newtonian physics predicts that these latter sorts of objects behave, more or less, just like cats, cars and asteroids, whereas quantum theory and relativity predict that they behave differently; and there’s alot of evidence that confirms the latter predictions. So the problem may be in reconciling one’s intuitions about how macroscopic, slow-moving objects behave with what the evidence indicates about the behavior of microscopic, and/or fast-moving, and/or megascopic objects. I’d agree that there are aspects of relativity and quantum theory that make this reconciliation difficult, but this doesn’t seem like a new problem in the history of science: new theories tend to extent the reach of old theories into domains that aren’t immediately directly experientiable.
(For instance, most people’s experience of falling objects would suggest that a hammer would hit the ground before a feather if both were dropped simultaneously from the same height, and they would be right, under typical scenarios. And perhaps most people would say this would happen even in non-typical scenarios in which one imagines nothing existing in the region surrounding the falling objects. But evidence (i.e., YouTube videos of astronauts on the moon dropping hammers and feathers) indicates that there are scenarios in which both hit the ground at the same time, regardless of the height. Reconciling everyday experience with the evidential record in this case amounts to replacing Aristotelian intuitions about falling objects with Gallilean/Newtonian intuitions.)
3:AM: You have talked about different theories of Newtonian gravity as being empirically indistinguishable and you ask whether they really are different theories or just different ways of expressing the same thing? How could they be different if they are empirically indistinguishable? Aren’t some fundamental ideas about identity being violated if they are different?
JB: There are versions of Newtonian gravity that (under a literal interpretation) describe spacetime as being curved, and describe gravity as a manifestation of this curvature (just like in general relativity); and there are other versions that describe spacetime as being flat, and describe gravity as an ontologically real field. And one can mount an argument that these curved and flat spacetime versions make all the same empirically testable predictions. One thing this argument assumes, perhaps, is some sort of distinction between in-principle observable objects, and in-principle unobservable objects, with the former being capable of detection by means of empirical tests, and the latter not. So two theories can be empirically indistinguishable in the sense that they agree on the predictions they make concerning the in-principle observable objects they posit, but disagree on the nature of at least one in-principle unobservable object they posit.
3:AM: Reading some of the things Tim Maudlin says it sounds like physics has got itself into a position where it knows how to solve engineering problems but is clueless about what any of the theories actually mean. Not only that, physicists since Bohr don’t see it as their problem. I think he also says that physics is using the wrong maths. So that just blows my head off! How far off am I with this impression that we’re running blind in physics?
JB: In the early years of quantum mechanics, researchers were really very interested in philosophical issues, Bohr included! Historically I think we can probably blame World War II and the Cold War for current attitudes among many physicists towards philosophy (David Kaiser has argued that the enrollment glut in physics graduate programs in the 50s and 60s motivated the pedagogical turn in teaching quantum mechanics away from philosophy and towards more of a “shut-up-and-calculate” mentality). It would be great if contemporary physics were running blind: more work for us philosophers! But there are alot of contemporary physicists who are deeply interested in foundational issues. Many of them work in quantum gravity and quantum information theory. I think these areas of physics are where the cutting edge stuff is being done. This is where new mathematical techniques (for instance, category theory) are being developed and deployed. And it’s also where there seems to be the most overlap between physics and philosophy of physics.
3:AM: A key issue is whether contemporary physics is best understood as scientific realism or anti-realism. It seems that every theory is underdetermined, and this threatens any claim to realism. You’ve looked at an argument from Steven Weinberg to investigate this haven’t you? So what does the argument do and what does it conclude? Should we be scientific realists or not?
JB: Weinberg has an argument to the effect that quantum field theory is the way it is because it’s the only way to reconcile certain basic principles of quantum mechanics with certain basic principles of relativity. In other words, if you assume these principles, then you’re inevitably led to a unique theory (i.e., Weinberg’s version of quantum field theory). On the surface, this seems like a foil to claims that anti-realists make about theories in physics being underdetermined. In particular, one can agree that, for instance, theories A and B are underdetermined by empirical evidence, but that nevertheless, there are other reasons to pick theory A over theory B. Theory A, say, may be (in some sense) uniquely entailed by some set of basic principles, while theory B may not be. This isn’t necessarily an argument in support of realism; rather, it suggests that naive claims about underdetermination don’t necessarily hold up under scrutiny. I think we should be scientific realists of a certain type, but for different reasons.
3:AM: One of the big issues you grapple with is the question about what space and time are. Modern spacetime physics fuses them into a single thing doesn’t it? But there’s an argument, ‘the hole argument’ that suggests general relativity has things in its mathematical structure that don’t correlate with physical reality. Is that right? How do you go about sorting out the issues around all this?
JB: The standard interpretation of special and general relativity entails that space and time are fused into a single thing, spacetime; but there are other ways to interpret these theories that attempt to uphold a more traditional ontological distinction between space and time (and the standard interpretation itself comes in many different versions). However, the typical way questions about the nature of space and time are cashed out is in terms of two historical positions: substantivalism and relationalism. Substantivalism claims that space, and/or time (and/or spacetime) exist as physical things (in some sense), independently of other physical things. Relationalism claims that space, and/or time (and/or spacetime) exist, but consist merely in the relations between physical things. The hole argument is an argument against a particular substantivalist interpretation of general relativity (what’s called “manifold substantivalism” by afficianados). It’s basically a reformulation, in the context of general relativity, of Leibniz’s shift argument against Newton’s concept of absolute space. Leibniz claimed that the existence of absolute space would violate his principle of the identity of indiscernibles and his principle of sufficient reason. An essential part of his argument involves imagining a “shift” of states of affairs (i.e., the world) in absolute space, and then comparing shifted and unshifted states of affairs.
The hole argument claims that if spacetime exists in the manner in which manifold substantivalists say it does, then general relativity would violate a particular notion of determinism. And an essential part of the hole argument involves imagining an analogous shift of states of affairs (the only difference is that the shift and the states of affairs in the general relativistic context are defined in a different way than they are in Leibniz’s argument). In both arguments (from a contemporary point of view), the shifts are defined by symmetry transformations (spatial translations and velocity boosts in Leibniz’s argument, which are symmetries of Newton’s laws of motion; and diffeomorphisms in the hole argument, which are symmetries of Einstein’s equations). The general idea in both cases is to demonstrate that an interpretation of a theory that awards distinct ontological status to two objects related by a symmetry transformation may lead to problems. One conclusion is that one should restrict one’s ontology to equivalence classes of objects related by symmetry transformations (in some theories, these are called gauge-invariant objects).
A huge industry of responses to the hole argument grew and flourished in the 90s, but it’s since died down a bit. Interestingly enough, just as it was dying down among philosophers, the hole argument was picked up by physicists working in quantum gravity. It turns out (unsurprisingly) that how one interprets general relativity may influence how one attempts to extend it to a theory of quantum gravity.
JB: I think to understand these concerns, and the extent to which they’re well-motivated, it may help to first consider the cultural history of quantum gravity. Lee Smolin plays no small role in this history. It’s a history of two opposing camps: the relativitists, who approach the elusive quantum theory of gravity from the point of view of general relativity; and the particle physicists, who approach quantum gravity from the point of view of quantum field theory. (The task of constructing a quantum theory of gravity is the task of reconciling general relativity with quantum theory; this is hard since these two theories are conceptually, logically, and mathematically incompatible with each other.)
The second camp (the particle physicists) generally now call themselves string theorists, and they’ve been the most market savvy: they get themselves on PBS science shows and write big glossy books about the nature of reality. The two camps were initially quite distinct, but alot of overlap between them has slowly developed. Smolin belongs to the relativitist camp. One problem that this camp faces is called the “problem of time”. This is a problem with how to interpret classical general relativity. The problem is that, under a typical way of understanding what counts as an observable of a theory, and what counts as representing change, (under which most theories in physics can be said to describe observables that change in time), the observables of general relativity do not change. This typical way (the “constrained Hamiltonian” formalism) is important for the relativitist camp since the standard way of turning a classical theory (like general relativity) into a quantum theory is based on it.
I don’t think string enthusiasts consider this to be that much of a problem, since this problem doesn’t affect other classical field theories, even when they’re formulated in the problematic way (this is actually one way of understanding the incompatibility of quantum versions of these theories and general relativity). I’m definitely not qualified to pass judgment on either camp. Personally, though, I’d like to think that the eventual quantum theory of gravity will be something Completely Different from either general relativity or string theory.
3:AM: It seems that there’s a lot of stuff in physics that doesn’t correlate with physical reality. You say we should look to the electron to learn navigational lessons about this kind of issue. So what do we learn from the history of the electron?
JB: We should learn, I think, that object-oriented ontologies are typically not robust under theory change. What seems to be more robust are structure-oriented ontologies (although what the latter amounts to is not all that clear, I admit). There’s an argument that’s sometimes mounted by anti-realists against scientific realism called the “pessimistic meta-induction”. In one form, it claims that there have been theories in the past which have warranted our belief (i.e., that have met our standards of epistemic warrant), although today we don’t think they should be interpreted literally (since they refer to things, like the ether, say, that contemporary theories don’t think exist). The conclusion then is that we should not literally interpret current theories that similarly warrant our belief. In this form, the argument concerns the stability of reference across theory change (although perhaps it’s better to characterize this as the stability of certain types of ontologies across theory change).
The history of the electron suggests that referential stability at the level of objects that possess properties may be lacking, but stability at the level of structure may not be: As an individual object, the electron went through alot of character redevelopment during the early part of the 20th century, arguably to the point where newer versions of it appeared unrecognizable from earlier versions; but the structure associated with it (perhaps as encoded in the Hamiltonian or Lagrangian associated with it) remained fairly recognizable during the same period, and still does.
3:AM: Another area where it seems there are difficulties for interpreting the theories is the relationship between particles and fields. Is this to do with the difficulty of accounting for the world as we experience it from the formalism of Relativistic Quantum Field theories?
[QCD Vakuum (source: wiki.arcs.org.au/bin/view/Main/ILDG)]
JB: This is certainly the view that most philosophers of physics hold (also many physicists). In particular, they think that intuitions about particles can’t be upheld in quantum field theory. The argument involves first identifying pre-theoretic intuitions about the nature of a particle, and then identifying the mathematical representations that support these intuitions, and finally arguing that these mathematical representations are not always well-defined in all quantum field theories. I think this argument at best demonstrates that intuitions about particles informed by previous theories may not be appropriate in the context of quantum field theory, but I don’t think this means that quantum field theory cannot be given a particle interpretation. In general, I don’t think we have built-in, pre-theoretic intuitions about experience; rather, our intuitions about experience are informed, implicitly or explicitly, by theories.
3:AM: I guess perhaps this question ought to be asked first – are fields real? Are is this a parade case of anti-realist mathematical formalism?
JB: The same argument that claims particle interpretations of quantum field theory aren’t viable is equally effective against field interpretations, too (the sorts of mathematical representations that support particle interpretations can be mapped onto mathematical representations that support field interpretations). Since I don’t think the argument works against particles, I don’t think it works against fields, either. Personally I think quantum field theory is a theory about both particles and fields, at least as it’s typically formulated by physicists who actually use it to derive and test predictions. If the sorts of things it refers to under the labels “particles” and “fields” don’t fully meet prior intuitions about particles and fields, then so much the worse for these intuitions.
3:AM: Structural realism seems to suggest that structure exists independently of objects that instantiate it. Doesn’t this mean that Roy Sorensen is right to say that you could have two totally empty universes that are nevertheless different?
JB: In the philosophy of science literature, there are two main variants of structural realism, one ontic (structures really exist) and the other epistemic (our theories at most give us knowledge of structure). The ontic version has many flavors. A “moderate ontic structural realist” claims that structures and objects co-exist and are mutually interdependent. A “radical ontic structural realist” claims that structures exist independently of objects. According to this latter creature, structure is supposed to be something; it’s just a different sort of something than objects. So two universes devoid of objects could still contain differing structures and thus be different.
3:AM: Now you defend structuralist interpretations of spacetime don’t you? So how do you do this and how do you deflect the idea discussed in the last question that you are committed to relations without relata?
JB: I’ve tried to motivate radical ontic structural realism with respect to spacetime by considering different formulations of certain applications of general relativity that disagree over the ontological status of objects. There are solutions to the Einstein equations that describe spacetimes with asymptotic boundary conditions (these solutions can be used to model spacetime singularities or arbitrarily curved spacetimes that behave more simply in the asymptotic limit). One way of formulating them (using tensor fields and a differentiable manifold) seems to support an ontology of objects and properties/relations. An alternative way of formulating them (using category theory) seems to support an ontology of structure devoid of objects.
I think the argument against radical ontic structural realism based on the slogan “no relations without relata” assumes a set-theoretic understanding of structure. In set theory, perhaps, it makes sense to say there can be no relations without relata (given typical extensional definitions of relations, at least). But there are other formalisms that you might consider in trying to develop a concept of structure. In particular, category theory seems to suggest a notion of structure that can be divorced from relata.
3:AM: Talking of nothing, what do you make of the recent spat between philosophers and physicists in the Kraus vs Albert affair? Was Albert right to say that Kraus misunderstood the philosopher’s question, and even if he did, was it a case where the philosopher’s question becomes redundant, as Dennett recently commented?
3:AM: I agree with Albert’s critique. I think the gist of it is that Krauss could have meant two things by the term “nothing” in the title of his book (“A Universe from Nothing: Why There is Something Rather Than Nothing“). If by “nothing”, he meant what philosophers (and theologians) have traditionally meant by “nothing”, then he fails to provide an explanation for the why question implied by his subtitle. I take it that what he explains in his book is how vacuum states in relativistic quantum field theory can, under certain initial conditions, evolve into states which describe non-trivial matter/energy content. This is an explanation of how something, maybe a bit more complex, can come from something, maybe a bit less complex. It’s not an explanation of how something can come from nothing in the traditional philosophical sense (although I personally find this traditional philosophical question a bit dull and uninteresting). If by “nothing” Krauss meant “a vacuum state in QFT”, then he does provide an explanation for the why question implied by his subtitle, but I think he’s being disingenuous in portraying this explanation as relevant to the traditional philosophical debate.
Whether physics makes such traditional debates redundant is an interesting question. Forget about the turf battle between physics and philosophy, there’s an even more bitter turf fight that occasionally flares up between philosophers of physics and philosophers who work in analytic metaphysics. Some of the former accuse some of the latter of relying too much on naive intuitions; and that the best-informed metaphysics comes from an (up-to-date) understanding of physics. I’m a bit sympathetic to this view. But in any event, you don’t need quantum field theory and cosmology to make this point. All you need to realize is that questions like “Why is there something rather than nothing?” are irrelevant to an understanding of physical phenomena, and that what’s more relevant are the sorts of questions that theories in physics can address (like “How can a vacuum state in relativistic quantum field theory evolve into a state with non-trivial matter/energy content?”).
3:AM: What are we to make of the multiverse?
JB: I think the multiverse is a very good way for cosmologists to get up to speed on topics in contemporary philosophy of science; in particular, contemporary analyses of the relation between theory and evidence (as opposed to outdated ones based on falsificationism), the nature of probabilistic inferences, and the nature of scientific explanations. One issue in the multiverse literature is how to define a probability measure over the set of multiverses. John Norton, for instance, has (I think persuasively) argued that certain features of typical measures are inadequate to represent some of the epistemic scenarios in which multiverses are posited (in particular, “fine-tuning” scenarios that seek to explain why certain fundamental constants have the values they do). These scenarios are characterized by claims that have completely neutral support from current evidence (current theories take the values of fundamental constants as brute facts), and using additive measures to represent how probable such claims are risks conflating this neutrality of evidential support with disfavoring evidence. On the other hand, if there are dynamical (or ontological) reasons, say, for positing multiverse scenarios, as opposed to epistemic reasons, then multiverse research becomes a bit more interesting. The many-worlds interpretation of quantum mechanics is an example of the latter.
3:AM: And finally, can you recommend five books that will help us go further into this weird but crucial philosophical world?
JB: Here’s a list of texts that address some of the issues above. (The last two are pretty technical; the last is actually a textbook on quantum field theory, but it’s remarkable for it’s in-depth coverage of many philosophical and conceptual topics.)
Quantum Mechanics and Experience, David Albert
Philosophy of Physics: Space and Time, Tim Maudlin
Everything Must Go, James Ladyman and Don Ross
Topics in the Foundations of General Relativity and Newtonian Gravitation Theory, David Malament
The Conceptual Framework of Quantum Field Theory, Tony Duncan
ABOUT THE INTERVIEWER
Richard Marshall is still biding his time.
First published in 3:AM Magazine: Monday, June 24th, 2013.