Published on November 2, 2025 6:03 AM GMT

It’s November! I’m not doing Inkhaven, or NaNoWriMo (RIP), or writing a short story every day, or quitting shaving or anything else. But I (along with some housemates) am going to try to write a blog post of at least 500 words every day of the month. (Inkhaven is just down the street a bit and I’m hoping to benefit from some kind of proximity effect.)

Today: Llamamoe on Discord complains about

people who respect science but say "and in the past we thought the earth was flat and everything we currently think is impossible might end up possible" and refuse to acknowledge that some things are in fact fundamentally impossible, like true FTL travel

And elaborates:

Like in principle FTL could be possible. But it would require everything we think we know about physics to turn out to have been wrong, and not slightly but completely, with no real exceptions.

I'm actually gonna side with the FTL believers on this one, with some caveats.

(Content warning: physicist discussing philosophy)

Map and territory

The case for FTL being "fundamentally impossible" is pretty straightforward: relativity is generally accepted as a "correct" physical theory; relativity says FTL is fundamentally impossible; therefore FTL is fundamentally impossible.

For the purposes of this post, I think it's basically true that FTL is fundamentally impossible according to relativity. (Tachyons, in theories which contain them, are probably better-modeled as something similar to the classic "shadows can move faster than light" brainteaser.)

The flaw in this argument is that "correctness" of theories in physics doesn't go as far as we might like it to. Sure, relativity has passed many difficult experimental tests with flying colors. This is enough for us to accept it as a highly accurate model of reality. When it makes quantitative predictions, we will happily adopt those predictions with pretty high confidence. But I want to distinguish the following claims:

• FTL travel is probably impossible
• FTL travel is impossible
• FTL travel is "fundamentally" impossible

The first claim requires relativity to be right about most things (and for FTL travel to not be a likely exception); the second requires relativity to be right about FTL travel in particular; and the third claim requires relativity to be right about everything, such that we can adopt not just its predictions but its internal ontology. As I'll explain below, I think this last claim is a lot stronger than the other two, and requires some nonobvious philosophical commitments.

Predictions and ontologies

There's a long track record of physical theories being extremely good models, but ultimately wrong in a way that is fatal for their basic ontology of the world. Newtonian physics (and Galilean relativity in particular) is a good example. At this point, some skepticism towards the ontology is warranted.

I think working physicists vary a lot in how strongly they believe in scientific realism. The actual work of physics only requires (some degree of) consensus on the trustworthiness of theories in terms of their predictions; individual physicists are free to treat the internal language of the theories (in terms of electrons, fiber bundles, wavefunctions, etc) as a literal description of reality, as a formal symbol-game with no truth value, or anything in between.

If you take an anti-realist stance, then physical theories are really just tools for making predictions about the world, with some colorful mnemonics attached to the prediction-making machinery. If you take a realist stance, then physical theories are not just making predictions, but also telling us all kinds of wonderful things about an unseen world of electrons and fiber bundles and so on. On the other hand, according to the realist stance, most physical theories to date have been wrong about the interesting part, and successful at predictions kind of by accident.

I'm going to try to take an awkward middle-of-the-road position here: one that's realist enough to let us ascribe some truth to the colorful stories our theories tell, but anti-realist enough to survive the ontological apocalypses that happen whenever a theory is superseded by a more correct one.

Kuhn

I'm going to throw in a shout-out to a blog post on Kuhn, "Science Cannot Count to Red. That’s Probably Fine.", by Lou Keep. In particular:

Newtonian physics makes several ontological claims (the universe is corporeal particles), Ptolemaic astronomy the same (circles are fitting for the heavens due to their divinity), etc. Both of these are wrong. Newtonian physics, however, can solve many more puzzles. "Amount of puzzles solved" is commensurable - it carries from one scientific set to another, there's a quantifiable, comparable idea of progress. The ontologies of the paradigms display no such progress.

I think this is a bit too strong. I'd say we make some ontological progress too: just as the quantitative predictions of wrong-but-useful models are approximately correct, I think the ontological claims are often approximately correct in an appropriate sense.

Approximate ontological correctness

As an example, Newtonian physics claims that spacetime is invariant under the Galilean group Gal(3). Relativity claims it's invariant under the group SO(3, 1). The former is a group contraction of the latter, so we can view Newtonian physics as making a kind of "qualitatively approximately correct" claim: spacetime is invariant under something that is approximately Gal(3), in the appropriate limit.

Similarly, atomic nuclei are not indivisible point particles, but they are approximately so, on the scale at which chemistry happens.

There's something kind of absurd about this, to be sure. The ontologies of physical theories have a reassuring crisp, absolute flavor to them; trying to believe them only in an "approximate" sense means throwing the crispness while trying to keep everything else. But I think it's in line with how we use informal ontologies in everyday life. When we claim something is rectangular, we're not insisting on geometrical perfection; we're saying that it is "approximately" a four-sided shape with four right angles. (Note that we're not even claiming it has "approximately four" sides; Colorado's border is officially defined by 697 straight boundary lines.)

All models are wrong?

Some say that "all models are wrong, but some models are useful." This is a fairly anti-realist stance; I would probably modify that to "most models are wrong", and add that we don't know which of our models, if any, are right.

Personally I'd bet on the perfect correctness of quantum mechanics, against that of quantum field theory, and very tentatively in favor of some version of relativity, but probably not in 3+1 dimensions.

What I mean by this is that I think quantum field theory is merely approximately ontologically correct, but that QM is exactly ontologically correct -- the true substance of reality is something "ontologically approximately like" a bunch of quantum fields, but it's precisely a wavefunction in an appropriate Hilbert space. And likewise, there is probably something worth calling spacetime that is in some sense a Lorentzian manifold, but probably not a 3+1-dimensional one. For example, it might be a 10+1-dimensional manifold compactified onto a 3+1-dimensional base.

Back to FTL travel

It's actually pretty unclear where that leaves FTL travel (through the ordinary 3+1-dimensional spacetime of general relativity).

My best guess is actually that it's "approximately fundamentally impossible": FTL travel is arguably possible, but only in situations where (3+1-dimensional) spacetime itself is close to breaking down.

As an example, the "ER=EPR correpondence" speculates that strong enough quantum entanglement between distant objects can be usefully modeled as a wormhole physically connecting the objects. (One thought experiment involves Alice and Bob creating a pair of black holes far apart from each other, entangling them by throwing in a bunch of Bell pairs, and then diving through the event horizons to meet each other in the wormhole's interior.)

To the extent that spacetime is a real thing, you can't move faster than light through it, just as relativity says. But the claim that "spacetime exists and has 3+1 dimensions" is itself only approximately true.

Discuss