If one is discovered before 2050, resolves to that. Otherwise resolves to "None before 2050"

What does "discovery" of a graviton entail?

In most cases, discovering a particle is the same as discovering its field. But the graviton is the hypothesized quantum of the GR spacetime metric (gravitational field), which there's already abundant evidence for in the classical regime.

cf. the axion, which is fairly likely to be discovered in the classical-field regime if at all, and at any rate probably not due to individual particle interactions.

So does discovering a graviton mean evidence for the quantum nature of gravity? What does that mean exactly and how are you going to operationalize it?

@AMS I’d argue that the discovery of a graviton means that there’s some prediction in a QFT-like theory of gravity which differs from the classical prediction and is verified experimentally.

I’d apply the same threshold for axions, even if the experiment is mostly classical, the axion is required for a theoretical understanding of the measurement.

@Eel13 by "the classical prediction" you mean GR?

I think this is reasonable but not actually analogous to the axion case, because a discovery of the axion *field* could be adequately modeled by a classical field theory analogous to GR (e.g. Maxwell EM + classical axion field) and should still count as discovering the axion.

In other words, "discovering a particle" = any evidence of beyond Standard Model degrees of freedom, except for GR metric because we already know about it. So we need a new standard for gravity, and I guess "anything gravitational beyond GR" seems fine to me.

"QFT-like theory of gravity" seems both too easy and too hard. IIRC there's a nonrenormalizable effective field theory of Standard Model + gravity which includes graviton interactions as a way of modeling gravitational effects we already observe (light bending, etc). On the other hand, an actual theory of quantum gravity probably won't have QFT-like degrees of freedom.

@AMS I think the two cases are different because one is a force mediating boson and one isn’t. Massless gauge bosons like gluons are confirmed only because of the consequences of their existence in QFT. QFT says that forces are mediated by particles, so discovering a force that can’t be explained by another framework (like GR) can mean discovering a particle.

Axions are different because they’re not force mediating, so the particle has to be discovered by directly observing it interact with photons in a specific mass state.

@Eel13 By “QFT-like” I mean some theoretical framework that describes the gravitational force in terms of the exchange of gravitons. If there is some quantum gravity theory that makes no reference to gravitons, then I’d say it’s far enough away from being “QFT-like” that this wouldn’t count as the discovery of gravitons even if it was experimentally confirmed.

@Eel13 I don't see the principled difference between e.g. the design of the DM Radio experiment and e.g. an interferometric measurement of gravitational redshift.

In both cases you have one field (the EM field) interacting via classical field theory with another field (axion or spacetime metric), causing coherent observable changes in the EM field.

Also, the axion is a boson and does mediate a force. It's not a gauge boson but the graviton isn't the usual kind of gauge boson either.

I think your description of "QFT-like" standard would either include the interferometer redshift measurement or *exclude* DM Radio, since those are both classical field effects that can be described as scattering processes involving coherent states of gravitons/axions in QFT. It would also clearly exclude string theoretic modifications of gravity, since gravitons aren't a fundamental degree of freedom in string theory (they show up in the EFT, but we can already write down a similar EFT for GR effects). This doesn't seem viable. Discovering string theory experimentally has gotta count as discovering gravitons.

@AMS I think the difference between those cases would be that the gravitational experiment could have explanations that don’t require the existence of a graviton, while the axion experiment necessarily requires an axion with a specific mass to explain the results.

Seeing something strange in a gravitational experiment doesn’t immediately tell you that gravitons exist. Seeing a signal in an axion experiment immediately tells you that axions exist.

I’m basically saying that not everything outside the standard model necessarily counts as a new particle.

@Eel13 I think the real problem is that, by the standards we would apply to any other new field, we already *have* discovered gravitons. Axion particles aren't strictly required by axion field detection -- "axions but they're a classical field and just sort of awkwardly couple to EM without acknowledging QM" is pretty analogous to the way people think about the established evidence for GR and the SM. But we would immediately infer that the axion field is quantized, as a matter of low-energy effective theory, and we can in fact do the same with GR -- people are just more hesitant to make that inference about gravity.

(A measurement of the axion mass doesn't imply the field is quantized and has particles, because the field has a parameter corresponding to the mass.)

I'm probably being too pedantic. If "graviton" includes confirmations of any intrinsically-quantum theory involving gravity, excluding the low energy EFT of GR, I'm happy to take that as clear enough resolution criteria despite philosophical quibbles.

@AMS I understand what you’re saying to some extent. I think the difference is that an axion experiment signal would very clearly extend the existing standard model with another particle, while it’s still possible that gravity could never fit in in that way.

I guess maybe I’m thinking of the definition of “particle” as “something that could extend the standard model without upending the whole idea of QFT too much”. And then people can still debate how much is “too much”.

I’m not an expert, but I’d say that string theory isn’t “too much” different from QFT in the sense that both transmit forces by mediating particles. But our current theory of GR is “too much” different from QFT in the sense that there are no mediating particles at all. It’s to be determined which of those two the true theory of gravity looks more like.

@Eel13 I guess I'm saying that an axion experimental signal *doesn't* clearly add a particle to the model, without additional assumptions. And the analogous additional assumptions, in the case of GR, would imply gravitons are "real".

In general, particles aren't fundamental in QFT, they're just an approximate consequence of combining field theories (like Maxwell EM, or GR) with QM. ("Mediating particles" is not really dependent on particles; even in QFT it's really the field that mediates, just like in GR. And forces in this sense are not a fundamental thing either.) So it's hard to imagine getting a theory of quantum gravity that doesn't involve gravitons. But people do try sometimes.

Similarly you can imagine that an axion field exists but isn't quantum and has no particles. It would still mediate very weak forces between electrically charged particles.

So the distinction rests on the fact that:

Nobody has hypothesized that the axion field is real but un-quantum (no particles), because it would be awkward and there's no reason to.

Some people do hypothesize that the gravitational field is real but un-quantum (no gravitons), even though it's awkward. String theorists and their biggest critics (LQG) both expect gravitons, but gravity is weird and special so there's a variety of weirder opinions beyond that.

In other words it's all about what people have *already* hypothesized, ie what's at least marginally popular with theorists, not about what's theoretically possible.

Actually I think this does matter, not for gravitons but for modified gravity and non-minimal couplings of QFT fields to gravity. (These are hypotheses that live in the gray area between "GR is a little different" and "QFT has a fun extra particle".)

An attempt at criteria:

Gravitons are discovered if an observation is made which is predicted by any theory that has a quantized gravitational field and

*not*predicted by "semiclassical quantum gravity" (QFT in*fixed*curved spacetime + non-quantum gravity coupled to*averaged*QFT matter). In particular, Hawking radiation wouldn't count. Also, modified gravity doesn't count as a graviton if its interactions with QFT fit the semiclassical structure I gave.Any new field that couples in the usual way to gravity counts as a new particle, even if it could conceivably be a classical field in the same sense GR could be. Axions count even if the evidence can be modeled via classical field theory.

(Option A) Modified gravity with a new field (TeVeS etc) would count as one new particle for each of the new fields, but no new particle for the tensor (metric) field. Downside: two particles for three fields is very weird if the evidence doesn't distinguish then like that.

(Option B) TeVeS confirmed = zero new particles, treating the other two fields like the tensor. And also, fields that are expected by theorists to have non-minimal coupling to gravity only count as particles if they are confirmed to be quantized. Downside: seems quite unfair to the nonminimally coupled fields in e.g. inflation models. Why should adding a hypothesized coupling to your model disqualify a particle discovery?

(Option C) TeVeS doesn't count as particles but nonminimally coupled fields do. Downside: there is not a principled difference here so we'd have to draw a fuzzy line between very similar models that "aren't really

*about*modifying gravity" and those that are. This will cause fights.

(Yes, I'm causing trouble, but I do think these gray area models are pretty plausible discoveries! Especially if we learn a lot more about inflation.)