A Tour Through Philosophy of Physics


Philosophy of physics is a massive field. Philpapers lists 15,000 papers in the philosophy of quantum mechanics and the philosophy of space and time alone. Most philosophy of physics is focused within these two areas. Yet much of that philosophy aims at solving problems that physicists no longer care about, never cared about, or waste little time on. Rickles notes that the topics covered in introductions to phil physics are "usually very old fashioned and limited in scope: 'spacetime' means the 'twin paradox'; 'statistical mechanics' means 'time assymetry'; and 'quantum theory' means 'the measurement problem'." Because such introductions are numerous, this tour provides a more modern state-of-the-art view.

We begin by asking what it is that physicists do and what sorts of treats they serve up for our philosophical study.

Following the Physicist

Physicists build theories, models, simulations, experiments, tools to facilitate experimental work, and so on. Some spend much of their time thinking about those theories and models and experiments. Others carry out the experiments, make measurements, examine and torture the data retrieved from experiments to get it into a workable format. Many spend much of their time doing mathematics, some pure and some applied. Sometimes the physicist outsources difficult work to the mathematics or statistics departments. As philosophers of physics we care both about the product of physics research and the processes used to manufacture the product. The philosopher is not—or should not bea mere mouthpiece for the physicist. All of the products and practices listed here fall under the philosopher's critical eye.

Interpretation

(Some) physical theories and models talk about the world, promising to reveal its secrets. But what do they say? How should we listen? Can we trust them? These questions belong to a philosophical practice called interpretation. Whereas the caricatured physicist shuts up and calculates, the caricatured philosopher of physics takes those calculations and tells us what they say the world is like. In truth, both physicists and philosophers are interpreters. Both care about describing the "furniture" or "hidden springs" of the world we live in, about uncovering what's real (whatever that means).

But what is interpretation? Most philosophers take for granted that interpretation looks like: assuming the theory is a true description of our world, what is the fundamental structure of the world we live in? The following quotes (stolen from Porter's Scientific Realism Made Effective) support this picture:
  • (Earman [2004], p. 1234): “Whatever else it means to interpret a scientific theory, it means saying what the world would have to be like if the theory is true.” 
  • (Van Fraassen [1991], p. 242): “Hence we come to the question of interpretation: Under what conditions is the theory true? What does it say the world is like? These two questions are the same.” 
  • (Belot [1998], p. 532): “To interpret a theory is to describe the possible worlds about which it is the literal truth. Thus, an interpretation of electromagnetism seems to tell us about a possible world which we know to be distinct from our own.”
  • (Fraser [2009], p. 558): “By ‘interpretation’ I mean the activity of giving an answer to the following hypothetical question: ‘If QFT were true, what would reality be like?’”
For example, assume that quantum field theories furnish an accurate description of the world. The philosopher subsequently asks "are particles or fields more fundamental?" Then they argue about this in highly technical and clever ways. There are alternative views of interpretation (Porter offers one that I'm personally sympathetic to) but the interpretation of theories, models, equations, data, etc. exhausts a large portion of the philosophy of physics literature.

Interpretation isn't all that spookydon't let it scare you. When someone tells you that "there is a tree over there with green leaves" you typically know what they mean. They've told you that something (a tree) exists and that it has certain (green, leafy) properties. Interpreting the statement comes naturally to you. The statements that physicists make often have less clear a meaning. We can readily imagine a world populated by trees, but it is not so easy to come to grips with how to interpret wavefunctions and lagrangian densities. And so the literature on interpretations of quantum mechanics was born.

Underdetermination and Theoretical Equivalence

A naive view of physics tells us that, once we have all the facts, the truth jumps out at us. Upon simply looking at the data we know reality. This is sadly not the case. Theories are cheap. Suppose you have a theory T1 that tells you the world only has particles. Suppose you have another theory T2 that tells you the world only has fields. Now suppose that T1 and T2 agree on all empirical matters, that they are empirically equivalent. In the naive view you would choose between competing theories by looking at the data, but T1 and T2 agree where data is concerned. What are you to do?


Enter the philosopher. Several moves are available (as described by Le Bihan and Read):
  • common core strategy: figure out where T1 and T2 fail to contradict each other, committing to their common ontology (what they say the world is like). If they both agree that atoms exist, for example, believe that atoms exist. Yet because they disagree on whether particles or fields exist, you have little choice but to suspend judgement and hope for a future resolution.
  • discrimination strategy: use some sort of selection principle to choose T1 or T2's ontology. You might use a coherence selection principle: choose whichever theory better coheres with what we already know about the world. You might also use an epistemic selection principle: choose whichever theory is better understood at the moment. 
  • unification: look for a theory more fundamental than T1 or T2 that unifies them. 
I'm guessing that these moves don't sound too satisfying to you, and they shouldn't. Underdetermination is a hard problem. It gets even worse. 

Underdetermination by duality

We talked about theories that are empirically equivalent to each other. There are lots of other equivalences to worry and philosophize about: theories that are mathematically equivalent but inequivalent in other ways (a problem for structural realists, usually), theories that are ontologically equivalent but that disagree on empirical data, etc. Another popular kind of equivalence (or symmetry) in physics is that of a duality.

We say that two theories are dual to each other when they determine the same physics. No matter which theory we use, physics remains unchanged. This might sound similar to empirical equivalence, but the discovery of dualities is generally surprising. Dualities often hold between theories that are far-separated from one another in some respect. For example, a theory defined at high energies (to describe high energy phenomena) might be the dual to a theory defined at low energies. This turns out to be very productive for physicists as it allows us to use the more tractable lower energy theory to investigate properties of higher energy physics where its dual is defined. But what does this mean? How can two theories that tell us very different things about the world be equivalent? What the hell should we say the world is like in such cases? Philosophers of physics worry about these things.

Physics Avoidance

Physicists do not walk through the steps of the scientific method (taught in high-schools as a toy model of scientific inquiry) and directly compare their theories to reality. Their work is far messier. Much of their work involves physics avoidancea phrase I borrow from Mark Wilsonor clever techniques by which very hard problems can be avoided and easier problems tackled in their place. So, the study of nature is not direct: we do not stroll leisurely, but fight against nearly insurmountable forces to gain even an indirect glimpse, in poor light and with terrible vision, of the truth.

The following are physics avoidance strategies and have individually received different amounts of attention from philosophers of physics. In each of these cases I will use the example of one particular target system, nuclear structure, to make things concrete:
  • Idealization: while there are different kinds of idealizations, typically an idealization is a completely different system than the target system where some distortions are introduced. The idealized system resembles the target system in some ways but is distinct. For example, nuclear physicists have modeled nuclear structure by analogy with a liquid drop-like incompressible fluid. Of course, nuclei are not actually fluids. So in the liquid drop model we study a system that is strictly distinct from the actual target system of nuclear structure.
  • Approximation: the inexact description of a target system. Most mathematical descriptions of nuclear structure are merely approximate. For example, suppose that an exponential function exactly represents the behavior of some target nuclei as it decays. You might approximate that function with a power series. Note that the physicist could choose an idealized system to study such that the power series is an exact description of it (even though it inexactly describes the actual system).  
  • Simulation: most physical computations are done with the help of some sort of computer. In general, simulations are models of some system in the world that are run on a computer (e.g., by way of executing an algorithm to solve some set of differential equations). Sometimes these simulations are used when there is no data. Sometimes data is put into the simulation and used to fix some set of parameters experimentally. Sometimes the simulations help us figure out what to look for when we're doing experiments. In nuclear structure, for example, we might use a lattice simulation on a computer to study how multiple nuclei interact with each other over time. If you're familiar with LIGO and gravitational waves, simulations are used to construct waveform template banks generated by the solutions to Einstein's field equations. When we get data from the detectors, those templates are matched to the data such that we can infer whether the signal in the data matches the prediction of general relativity. 
  • Analogue modeling: a species of idealization in which an analogy is constructed between the target system and a more tractable system. The liquid drop model is one such example. There exists an entire field called "analogue gravity" where physicists try to study gravitational systems by way of a comparison to different physical systems, such as fluids, electromagnetism, etc. 
  • Model reduction: the reduction of a mathematical model's complexity. So, if the dimensionality of your model of protons and neutrons is in the thousands, you might coarse grain the system and only study the bulk/collective behavior of the protons and neutrons instead. Similarly, rather than try and track the trajectories of trillions of particles in a room, we talk about bulk properties like "temperature" to kill off unnecessary degrees of freedom.
protons and neutrons on a lattice with lattice spacing a.      

As I mentioned, these are all indirect routes to characterizing nature. The epistemic and ontic questions one might ask here are endless, and philosophers of physicists explore them restlessly. Besides wanting to know what models themselves are and what good modeling looks like, we want to know what these physics avoidance strategies tell us about the world (if they do tell us about the world) and how exactly they do so. Can we use a model of a fluid to test theories of gravity? How are false models so good at prediction; why do they work so well? 

The Intersection with Phil Sci

More general issues in philosophy of science tend to crossover in some way with issues in physics and the philosophy of physics. These vary from realism and theory change to emergence and reduction. One such debate happened between Batterman and BelotSome philosophers and physicists hope to reduce everything else to physics: 

Briefly, a reductionist will tell you that we can't model a lion's behavior with quantum mechanics simply in virtue of computational expense. If we had extremely powerful computers, everything can in principle be reduced to the most fundamental physics. Batterman argues that more fundamental theories are often explanatorily deficient by necessity. For example, he notes that quantum mechanics itself makes reference to classical concepts and that it could not be otherwise. Dirac says something similar in his 1957 Principles of Quantum Mechanics text: 
Quantum mechanics occupies a very unusual place among physical theories: it contains classical mechanics as a limiting case, yet at the same time requires this limiting case for its own formulation. 
Like Batterman, Belot tackles several case studies in physics and makes use of the mathematical formalism; philosophers of physics need to know the physics. Belot argues that the less fundamental stuff (like the lion or classical mechanics) is, contra Batterman, eliminable. It can be reduced. At the level of mathematical formalism, we can still define the less fundamental in terms of the more fundamental: classical mechanics (semi-classical mechanics, really) can be derived from the formalism of quantum mechanics alone. While I have simplified the argument, this is a paradigmatic case of a philosophy of science argument that intersects with physics and the philosophy of physics. More generally, philosophers of physics often insert themselves into philosophy of science arguments or borrow tools from philosophy of science to use when looking at physics.

What Should Physicists Even Do? 

Physicists sometimes disagree about what physics is. Is string theory physics? Are interpretations of quantum mechanics just spooky metaphysics that physicists should ignore? Is it okay to use non-empirical methods when empirical methods are not forthcoming? If not, should we just give up on investigating really hard problems for now? Which methods are legitimate methods that physicists should use; which methods are not; and how do we make such decisions? What kinds of evidence are admissible? Philosophers have long been interested in such questions, questions that physicists themselves tend to ask only when something is amiss. Consider a few observations:
  • Physicists have historically and still use non-empirical considerations when doing physics. They invoke things like simplicity, beauty, naturalness, physical intuition, cognitive understanding, and so on when evaluating theories, thinking about which problems we ought tackle and how we ought tackle them, and so on. Philosophers of physics love this; it keeps them in business. A recent conference between physicists and philosophers of physics output an entire text, Why Trust a Theory, focused on problems of this sort.
  • Physicists tend to take their discipline to offer constraints on other disciplines and on good-thinking. If a result in biology or chemistry violates what physicists know, it must be thrown out. If the beliefs of a philosopher doing metaphysics runs afoul of what the standard model of particle physics tells us, it cannot stand. Thus physicists do not only run experiments, build theories, and tell us what the world is like, but they perform some other duties too. They limit what is possible and what is impossible. They tell us what makes physical sense (e.g., where some explanations for the origin of life are ruled out for physical reason alone). Naturally, the philosopher cares about this too.

Some Flavor From Yours Truly

Here I'll briefly summarize some research directions that I am currently exploring in the hopes that you can get a more clear picture of what philosophers of physics care about:

Scale Separation

In nature, we often find that scales are separated from each other. You can study the trajectory of a baseball without having a theory of quantum gravity. Physics at the baseball scale and at the scale of quantum gravity decouple from one another such that we can study them separately. See my EFT thread for a better explanation of this. Many physicists take this scale separation to be a necessary precondition for doing science:
    • (Duncan [2012], p. 570) "if this [scale separation] did not occur, the correct interpretation of the results of experiments carried out in some localized region of spacetime would be contingent on a specification of the state of the universe faar beyond the boundaries of the laboratory---a state of knowledge about our surroundings which we can clearly never possess." 
    • (Zeidler [2000] p. 192) "this [separation of scales] is the reason why we are able to do physics." 
Unfortunately, scales often fail to be separated and our effective field theories don't work. So what can physicists possibly mean when they appeal to scale separation as a necessary precondition (as not only is it not necessary, it isn't even actual)?

Knowing Things

Newton thought his theory of universal gravitation was, well, universal. (Most) physicists have learned their lesson. Theory change was violent to our beliefs, and it's hard to be a realist about what our theories say. Anti-realists use a pessimistic meta-induction (over the graveyard of discarded scientific theories) to show that, just in case modern theories are like past theories in the right way and modern scientists are like past scientists in the right way, we should doubt what science says about the world.

But now we have tools, such as renormalization, that allow us to precisely delineate the domain of applicability of our theories and characterize to what extent some theory depends on unknown physics (e.g., how sensitive physics at the scale of hadrons is on physics at the quark scale, or how sensitive physics at the scale of baseballs is to physics at the Planck scale). Put very simply, this appears to be one way in which we have gotten better at knowing things. That is, when theory change comes back around to hit us, we won't be caught with our pants down. Our theories are effective and the ways in which we do physics to study their limits differentiates contemporary research from that which was lost in the past (more technically: we're in a better epistemic position than we once were). ∴ realism is true. 

There's More

What I've detailed is only a small window into the contemporary philosophy of physics, but there is much more. Below I will post a few resources that can work to guide your future study if I have captured your interest:

Introductions/Anthologies
Primary Texts




























Comments

  1. This comment has been removed by the author.

    ReplyDelete
  2. People overrate the "success"of physical science and mathematics in explaining the "reality" of matter.

    From a philosophy of science point of view the deeper issue is that physical science and mathematics "entirely forget" the reality of spacetime manifold itself, let alone the matter, because we are only using charts & transition maps to study the phenomenon of spacetime and not the "noumenon" of spacetime.

    Technical explanation here: https://t.co/IJr7OrXknH?amp=1


    The physicist actually says that in physics we shamefully and entirely forget the reality of spacetime in itself. Yet we explain it fully by differential geometry.

    ReplyDelete

Post a Comment

Popular posts from this blog

The Realist Quest