Epistemology, Not Metaphysics, Justifies Experiments

While I was visiting the IAS a few weeks back, they had a workshop on Quantum Information and Black Holes. I didn’t see many of the talks, but I did get to see Leonard Susskind talk about his new slogan, GR=QM.

For some time now, researchers have been uncovering deep connections between gravity and quantum mechanics. Juan Maldacena jump-started the field with the discovery of AdS/CFT, showing that theories that describe gravity in a particular curved space (Anti-de Sitter, or AdS) are equivalent to non-gravity quantum theories describing the boundary of that space (specifically, Conformal Field Theories, or CFTs). The two theories contain the same information and, with the right “dictionary”, describe the same physics: in our field’s vernacular, they’re dual. Since then, physicists have found broader similarities, situations where properties of quantum mechanics, like entanglement, are closely linked to properties of gravity theories. Maldacena’s ER=EPR may be the most publicized of these, a conjectured equivalence between Einstein-Rosen bridges (colloquially known as wormholes) and entangled pairs of particles (famously characterized by Einstein, Podolsky, and Rosen).

GR=QM is clearly a riff on ER=EPR, but Susskind is making a more radical claim. Based on these developments, including his own work on quantum complexity, Susskind is arguing that the right kind of quantum mechanical system automatically gives rise to quantum gravity. What’s more, he claims that these systems will be available, using quantum computers, within roughly a decade. Within ten years or so, we’ll be able to do quantum gravity experiments.

That sounds ridiculous, until you realize he’s talking about dual theories. What he’s imagining is not an experiment at the absurdly high energies necessary to test quantum gravity, but rather a low-energy quantum mechanics experiment that is equivalent, by something like AdS/CFT, to a quantum gravity experiment.

Most people would think of that as a simulation, not an actual test of quantum gravity. Susskind, though, spends quite a bit of time defending the claim that it really is gravity, that literally GR=QM. His description of clever experiments and overarching physical principles is aimed at piling on evidence for that particular claim.

What do I think? I don’t think it matters much.

The claim Susskind is making is one of metaphysics: the philosophy of which things do and do not “really” exist. Unlike many physicists, I think metaphysics is worth discussing, that there are philosophers who make real progress with it.

But ultimately, Susskind is proposing a set of experiments. And what justifies experiments isn’t metaphysics, it’s epistemology: not what’s “really there”, but what we can learn.

What can we learn from the sorts of experiments Susskind is proposing?

Let’s get this out of the way first: we can’t learn which theory describes quantum gravity in our own world.

That’s because every one of these experiments relies on setting up a quantum system with particular properties. Every time, you’re choosing the “boundary theory”, the quantum mechanical side of GR=QM. Either you choose a theory with a known gravity partner, and you know how the inside should behave, or you choose a theory with an unknown partner. Either way, you have no reason to expect the gravity side to resemble the world we live in.

Plenty of people would get suspicious of Susskind here, and accuse him of trying to mislead people. They’re imagining headlines, “Experiment Proves String Theory”, based on a system intentionally set up to have a string theory dual, a system that can’t actually tell us whether string theory describes the real world.

That’s not where I’m going with this.

The experiments that Susskind is describing can’t prove string theory. But we could still learn something from them.

For one, we could learn whether these pairs of theories really are equivalent. AdS/CFT, ER=EPR, these are conjectures. In some cases, they’re conjectures with very good evidence. But they haven’t been proven, so it’s still possible there’s a problem people overlooked. One of the nice things about experiments and simulations is that they’re very good at exposing problems that were overlooked.

For another, we could get a better idea of how gravity behaves in general. By simulating a wide range of theories, we could look for overarching traits, properties that are common to most gravitational theories. We wouldn’t be sure that those properties hold in our world…but with enough examples, we could get pretty confident. Hopefully, we’d stumble on things that gravity has to do, in order to be gravity.

Susskind is quite capable of making these kinds of arguments, vastly more so than I. So it frustrates me that every time I’ve seen him talk or write about this, he hasn’t. Instead, he keeps framing things in terms of metaphysics, whether quantum mechanics “really is” gravity, whether the experiment “really” explores a wormhole. If he wants to usher in a new age of quantum gravity experiments, not just as a buzzword but as real, useful research, then eventually he’s going to have to stop harping on metaphysics and start talking epistemology. I look forward to when that happens.

6 thoughts on “Epistemology, Not Metaphysics, Justifies Experiments

  1. Alejandro Baroni

    Good article. I agree that science is not metaphysics, or should not be. As I understand, epistemology covers philosophy of science, its history and sorroundings. Many, too many scientists are epistemological methaphysicalists. I think your point, strong one, is that scientists work expanding knowledge, and not grasping ‘reality’ as ‘it is’.
    Thank you for your posts.
    Alejandro Baroni,
    Montevideo, Uruguay


  2. Pingback: GR=QM, el nuevo eslogan de Susskind | Ciencia | La Ciencia de la Mula Francis

  3. ohwilleke

    You often see something similar in the popular press when a compressed matter physics result analogous to a quantum system produces something that is widely accepted as not existing in the quantum system, like a magnetic monopole.

    Liked by 1 person

  4. Keith Douglas

    On the other hand, think of (say) David Bohm’s little book on causation. We now know, by taking arguments about very fundamental questions about reality into consideration, where to create new theories to further the enterprise of physics (etc.) Bohm did not realize, what the nature of the sub-quantum would have to be, if any, but we now do: if there is anything, it is non-local (this is the lesson of the Bell experiments). But a completely antimetaphysical attitude would have just left us “hanging” with no place to look for a successor theory. In this context it is important to realize, at least for scientific metaphysics (like mentioned) that the metaphysics itself is of course just another system of theories, albeit a hypergeneral one. These shade immediately into traditional “scientific” theories – I for one cannot find a dividing line. Some contemporary physicists like Lee Smolin seem to recognize this.


    1. 4gravitonsandagradstudent Post author

      Sure, that’s a fair point. I don’t think that it’s inappropriate to try to answer metaphysical questions scientifically, precisely because there isn’t a clear dividing line: people often label things as “metaphysics” merely because no-one has figured out how to address them experimentally, and the metaphysical question of whether x “really is” y becomes scientific as soon as you fix what you mean by “really is”.

      There’s a difference between trying to answer a metaphysical question and using a metaphysical argument to avoid answering a question, though. You do still need to answer “what can this experiment teach us that we didn’t already know?” and what frustrated me with Susskind’s talk was the lack of emphasis on that.



Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s