Hacker News Reader

Comment by thisiscorrect | Hacker News Reader

“What exactly is spacetime? Is it a real thing like an atom, or just a mathematical construct that is used to describe how mass 'generates' gravity?”

This is a false dichotomy. Atoms too are "just" a mathematical construct used to describe how certain observable, physical systems behave. It's even called the atomic _theory_, and like all other good theories it knows its limits. At low enough energies, it's too fine and at high enough energies it's too coarse a model. Spacetime is exactly as real as any other current model of observable physical phenomena that physics can offer.

View on HN

I disagree with you. Your argument could be made about anything, there is no such thing as a "human" either, it's just a big bunch of wave-particles that does not have some clear limits. Yet when we talk about a human, we all know what that is, and when a human is seating on a chair we don't think the chair is part of it.

The same applies to atoms, we all agree on what they are, and we can show pictures of them etc.

The question here is whether spacetime is something like that, or if it is a proxy concept with no grounding whatsoever in the universe, but with good predictive powers?

For example, for a long time we believed that Aether filled space, and although we now know this is not the case, the mathematical model provided by the Aether theory was pretty useful. You could very well imagine a world where we know Aether is not real, but continue using it for lack of a better theory as it is a useful framework.

> It's even called the atomic _theory_

Theory, in scientific theory, means the opposite of theory in our everyday language though. Theory does not mean a guess, it means a body of work accepted as valid and robustly tested.

> The question here is whether spacetime is something like that, or if it is a proxy concept with no grounding whatsoever in the universe, but with good predictive powers?

Consider that "human", "chair", and "atom" are also just concepts are minds apply to certain patters of energy. We all "know what that is" because it is advantageous for our minds to place things in these categories for the purpose of modeling. You may know what a "chair" is, but there are an infinity of gradations between "chair" and any two other objects, like the beanbag chair sitting between "chair" and "pillow".

The concept of a chair has no grounding in the universe, it's all in the mind. There are no physics that only apply to chairs.

> For example, for a long time we believed that Aether filled space, and although we now know this is not the case, the mathematical model provided by the Aether theory was pretty useful. You could very well imagine a world where we know Aether is not real, but continue using it for lack of a better theory as it is a useful framework.

Yes, that's how it works. When we have a better model we use that unless a simpler one will do. Newtonian gravitation isn't "real" by your definition, but we still use it in a lot of circumstances even though we have more accurate models.

Right, and that was the point I was trying to make with the human analogy.

So the question is whether spacetime is one of those patterns we can point to, like a "human" a "chair" or an "atom", which is what we call "real" in our everyday language, or whether it is only a useful modelization tool, like the Aether. There are no pattern anywhere in the universe matching the Aether.

It is a valid question.

I'm saying the differentiation is meaningless. We have a model of space time that makes accurate and measurable predictions and that is as real as real gets.

So then the Aether was real for a while and stopped being real 200 years ago?

If we discover a new model that makes better prediction than our current spacetime based model, just like relativity was to Newton's theory, will spacetime stop being real?

The article was clearly using "real" with the definition we attach to it in our everyday language, not in deep philosphical sense.

> So then the Aether was real for a while and stopped being real 200 years ago?

Yeah, pretty much.

> If we discover a new model that makes better prediction than our current spacetime based model, just like relativity was to Newton's theory, will spacetime stop being real?

In so far as Newtonian gravitation is also not-real. We still use that model for a lot of things of course. I don't know much about Aether theory or its predictions, so instead I'll use another long gone theory for my example: People used to model the motion of "heavenly bodies" using epicycles, complicated circles-within-circles patterns that assumed the Earth was the center of the universe and everything was taking a very complicated path around it. These epicycles were actually quite good at predicting the motion of planets and stars, so as far as anyone was concerned they were real. Some clever people eventually came around to the idea that all of that could be much more easily and more accurately modeled if we put the sun at the center. Of course we know today that that isn't quite right either.

We don't, and possibly cannot, ever know if the models we use to describe the world around us are the most accurate or not. They are real when they work, because we have no better metric to determine their reality.

I do agree with your point, but I think we are seeing the limit of language: when we don't attach the exact same definition to words, we can misunderstand each others while trying to convey the same thing.

The way you use "real" is the way I would use "valid": A theory is valid until proven otherwise.

And the way I use "real" (and I think it is the same way the article used it), is whether something has any grounding in our universe (some pattern as you said before).

For example, wormholes are possible following the theory of general relativity, but we don't know if they are "real", e.g. can there actually be a wormhole in our universe. If we do end up finding one, then they are real. If we however find conclusive evidence that they cannot exist in our universe, then they are not real, but only an artifact of our (incomplete) theory of general relativity.

> And the way I use "real" (and I think it is the same way the article used it), is whether something has any grounding in our universe (some pattern as you said before).

Many of the things you described as "real" don't have grounding in our universe, i.e. things like "chair". It's possible I misunderstand your usage of the term "grounding".

> For example, wormholes are possible following the theory of general relativity, but we don't know if they are "real", e.g. can there actually be a wormhole in our universe. If we do end up finding one, then they are real. If we however find conclusive evidence that they cannot exist in our universe, then they are not real, but only an artifact of our (incomplete) theory of general relativity.

Sure, but how does that apply to spacetime? We know, so far as anything can be known, that space and time are two different perspectives of the same thing. Relativity described it, and we have measured and the effects it predicts with startling accuracy.

Here's a question: It matters if wormholes can or cannot exist regardless of if the current model permits them or not. Does it matter if spacetime fits your definition of "real" or not?

Sabine's version of "real" - that the universe cannot exist without this is a better one imo

Worm holes aren't real until we can't describe phenomenon without them.

Same with the aether. Black holes and chairs are real because we can't make a theory that doesn't include them while still making accurate prediy

but the universe can exist without chairs

Aether is real again as Higgs field. Speed of light in vacuum is variable again (gravitational waves). Vacuum is no longer an empty space.

> Speed of light in vacuum is variable again (gravitational waves).

Hm? Can you elaborate on this?

We have a photo of a building that makes accurate and measurable predictions of things like the roof being above the door, and how far, and which parts are transparent (windows) and which parts opaque (walls), and how many people could fit inside it. Is that picture as real as buildings get?

Is your claim that the picture is the building? Because that's the equivalent of claiming that the a Penrose diagram is spacetime, instead of a representation of the model.

The photo of the building is an accurate diagram of the model of that one individual building. It describes the building, it is not actually the building. However, your model (roof height, walls, location, etc) is the building. If that building is destroyed in a fire and remade to those exact specs, do we not say it is still a building?

> "Is your claim that the picture is the building?"

You said "a model which makes accurate predictions is as real as real gets". That the picture is the building is your claim here, that there's nothing more real than an accurate model, which I'm countering by absurdum. Wind can turn a turbine, a weather predicting model of wind cannot. You say "the differentiation [between a model and reality] is meaningless", I say a model is a leaky abstraction - when you send a single atom through a Young's double slit experiment and it interferes with itself and generates a wave interference pattern, that's leaking a reality outside the billiard-ball model of atoms, that's one time when the distinction between model and non-model is meaningful.

> "If that building is destroyed in a fire and remade to those exact specs, do we not say it is still a building?"

If the building is destroyed in the fire, and your model of it isn't, how can you say there is no difference between the building and the model?

> You said "a model which makes accurate predictions is as real as real gets". That the picture is the building is your claim here, that there's nothing more real than an accurate model, which I'm countering by absurdum.

My claim is that the model describes a real thing, not that it itself is a real thing, and the extent to which that thing being described is real is a matter of the accuracy of the model. You can model the motion of planets with geocentric epicycles, but that is less real than a model that uses heliocentric ellipses.

Let me turn this around for a second:

> when you send a single atom through a Young's double slit experiment and it interferes with itself and generates a wave interference pattern, that's leaking a reality outside the billiard-ball model of atoms, that's one time when the distinction between model and non-model is meaningful.

Now ask yourself: if we have a 100% accurate model of physics that employs a concept of spacetime such that there are no leaks, no unaccounted for observations whatsoever, then would you say that it matters if spacetime is just some mathematical concept and the real mechanism works differently but produces exactly the same results?

I submit that it doesn't. A model that produces perfectly accurate descriptions describes reality in so far as any description of reality can be said to. We don't have perfect models of reality, that is known, however lacking a more accurate model it is meaningless to say that our current ones (which are astonishingly accurate despite being imperfect) don't describe reality.

> You can model the motion of planets with geocentric epicycles, but that is less real than a model that uses heliocentric ellipses.

This example actually weakens your argument. Heliocentric ellipses are chosen because they explain our model well and importantly are simpler than epicycles; but we can make epicycles arbitrarily accurate by piling on as many of them as you like.

So which is real? That Earth and the other planets revolve around the sun or that the sun and other planets revolve around the Earth?

> if we have a 100% accurate model of physics that employs a concept of spacetime such that there are no leaks, no unaccounted for observations whatsoever,

Let's say I have a box with a green LED, a red LED, and a button. One of the LEDs is always lit. Whenever I press the button, that LED becomes unlit, and the other becomes lit. If I try to open the box, it explodes. My model of how this box works is very simple: A demon is sitting inside the box and touches the LED that becomes lit. When I press the button he gets poked with a tack and switches to touching the other LED. If I try to open the box, he gets angry and blows up.

This model is 100% accurate. Every one of these boxes I have seen always behaves this way. Is there a demon sitting in the box?

> This example actually weakens your argument. Heliocentric ellipses are chosen because they explain our model well and importantly are simpler than epicycles; but we can make epicycles arbitrarily accurate by piling on as many of them as you like.

Epicycles allowed us to map the motions very well but they were incredibly complicated and didn't provide for explanations of how they came into being. Simpler models that provide the same results are considered to be better. Given two models that make the same predictions, it makes more sense to use the simpler one than the more complicated one (Occam's razor). The heliocentric model no doubt had its detractors until it was refined enough to make the same or better predictions, but its simplicity made it much more attractive as an explanation.

> This model is 100% accurate. Every one of these boxes I have seen always behaves this way. Is there a demon sitting in the box?

Typically we apply Occam's razor in this. If we know how we could construct such a device without a demon, a being for which we have no other evidence of its existence, then our simpler and more readily explainable solution is what we generally go with. However, if this box is indeed unopenable, and the light always changes, then it doesn't really matter does it? Neither the demon nor a more conventional explanation makes any difference to how we interact with the box, nor could it ever.

The "reality" of what happens inside the box is not relevant and some would argue not even something that exists in the realm of scientific thought. It's like, say you live in Conway's Game of Life, and ask "what are cells made of?", it is completely unknowable and even if you somehow had the answer it would be meaningless since cells always behave in exactly the same way that is 100% explained by a simple model.

You said:

> ... the extent to which that thing being described is real is a matter of the accuracy of the model. You can model the motion of planets with geocentric epicycles, but that is less real than a model that uses heliocentric ellipses.

This statement doesn't suggest simplicity as a measure of realness and it is unclear why reality would necessarily prefer the simpler thing. We prefer it, of course. However, the two models on offer are mathematically equivalent and differ only in: (1) simplicity, and (2) explanation. So I'm not sure if you're modifying your stated position or not when you apply Occam's razor here.

As for whether it matters, I think that's rather the point: We don't know if it matters until we find out it does. I don't think we can know if matters until the moment it definitely does. A 100% accurate full-knowledge model is, by our own models, an impossibility. And even if we had such a model, how would we ever know it?

> This statement doesn't suggest simplicity as a measure of realness and it is unclear why reality would necessarily prefer the simpler thing.

Because it doesn't matter if the results are the same, and simpler things are inherently preferred. Assume that both of these models are 100% correct in their predictions: there is no reason whatsoever to believe that the more complicated one is a "truer" model of reality, nor is there a reason to believe the simpler one is, so we proceed on the notion that the simpler one is because that is more to our liking and the end results are the same.

I guess one of the fundamental impasses here is that I am not convinced, on a philosophical level, that there is any deeper "realness" than can be modeled. If we can't model it, or rather if modelling it serves no purpose, then it may as well not exist.

> As for whether it matters, I think that's rather the point: We don't know if it matters until we find out it does. I don't think we can know if matters until the moment it definitely does.

Right. We must assume reality is what it appears to be until we make observations and tests that prove otherwise, then we construct a new model. We must assume that new model represents reality because we simply have no alternative until we build a better model.

> A 100% accurate full-knowledge model is, by our own models, an impossibility. And even if we had such a model, how would we ever know it?

To a certain extent we can't. No matter how many correct predictions we make there is always the possibility that we will eventually make an observation that defies our model. That's sorta my point. What we call reality is, by necessity, a product of observation, deduction, testing, and modelling. We have to construct our reality from that, and talking about whether or not an accurate model represents some deeper "reality" or not is pointless without an alternative.

If right now we question the "reality" of spacetime, what changes? Can we even build an accurate model of reality as we currently know it without spacetime? Can that model be made simpler than the ones with spacetime? If so, no one has done it yet. Therefore we have no choice but to believe that the universe isn't deceiving us and spacetime actually does exist, because believing the opposite is completely pointless.

"The same applies to atoms, we all agree on what they are, and we can show pictures of them etc."

This is where you bury the lede. It's clear if you read many of the comments here that there is no such agreement. Some people _think_ of an atom as something you can reach out and touch, or grok in some deep sense distinct from performing specific experiments. Others take an approach that's more in line with theory in the sense I use it (https://en.wikipedia.org/wiki/Scientific_theory). Your definition of theory is too loose: it's not just a body of work. That sounds like a pile of experimental evidence to me. A theory is a robust model or family of related models that match the experimental evidence. Atoms exist in that sense: they're a well tested model of experimental phenomena.

I'm certainly not claiming that atomic theory is a wild speculative guess which you seem to imply I've done.

> It's clear if you read many of the comments here that there is no such agreement

I guess I should have said "all experts agree" instead? Lack of knowledge about something by some people doesn't change the definition of that thing. I would add however, that you can reach out and touch an atom with a certain definition of touch (as touch is mostly a macroscopic concept).

> I'm certainly not claiming that atomic theory is a wild speculative guess which you seem to imply I've done.

You will have to explain this one to me then. You said "it's *even* called a theory" to reinforce your point that it was just a mathematical tool (and hence not real following your definition). This point is valid only if theory means guess, which it does not.

Touch isn't any different than measuring radiation interacting with something. These are metaphysical questions, not scientific ones.

Scientists, like everyone, have a realist perspective. They consider reality to correspond with appearences. But this is a convention or a metaphysical belief or even a religous belief, it's not scientific.

> I would add however, that you can reach out and touch an atom with a certain definition of touch (as touch is mostly a macroscopic concept).

Don't be so sure... atoms and even relatively large molecules definitely experimentally exhibit [1] wave particle duality in double slit experiments demonstrating they move as waves. Given that absolutely nothing is stationary at any temperature above absolute zero, what are you touching and with what?

[1] see end of article https://arstechnica.com/science/2021/01/the-curious-observer...

> I would add however, that you can reach out and touch an atom with a certain definition of touch

And unfortunately for your argument, the bit you elided here is the critical bit. If we understand "touch" meaning to interact with something "physically" rather than "chemically" or indirectly (light, sound), then "touch" refers essentially to electromagnetic force and the Pauli exclusion principle. Since the latter can only be defined in terms of yet more "matter", in fact we have no way to observe matter in any way but electromagnetism in its various forms, viz. "light" in a wide sense, chemical interaction, and via magnetic repulsion on macroscopic or microscopic scales. Given this, I would say it's largely correct to understand atomic theory as useful abstraction; indeed, an abstraction which dates back hundreds of years which has been refined and found still useful. That said, something being an abstraction doesn't make it not real! Indeed I'm not entirely sure I know what else "real" is supposed to be.

Theory is indeed separated into several disciplines, speaking about mathematization: Phenomenology, Model building, Mathematical Physics. FWIW I think Phenomenology is where a lot is going on, it's about looking what is there and how it compares to the models.

The fully formalized ("axiomatic") mathematical descriptions oftentimes lag behind the theoretical models ("postulated") in use that are used to compare with real data for instance.

That said, I would say Atoms are as real as it can get. There are so many experiments about them and applications as well, for instant nuclear fission. How can they not be real? :-) Speaking about Spacetime, of course it is real. There is such a crazy amount of predictions based on Special Relativity.

> Atoms are as real as it can get

Well put, and also not any specific degree of real, if you accept that the simulation hypothesis can't be disproven. The specific 'real' discovered properties of atoms or spacetime are actually artifacts of simulation.

Is simulation time continuous or discrete? If there is only 1x forward speed for simulation time, perhaps relativity could continue to work, as that fixed single speed would be discrete (speed of light).

I'm not sure how to postulate this idea further, but I've had it for years.

I like to imagine the world as being like a distributed game engine running on a computer outside of time. That would mean that life could have existed before physical space or the earth did. History being built piecemeal and in reverse order, a past being created by collective effort and attention. And somewhere along the way Santa planted dinosaur bones.

Maybe when you close your eyes you fade to black, and existing is a constant choice. Maybe situation and circumstance are handshake deals made in the collective unconscious, an onion growing layer by layer from the kernel.

Just something to think about :)

I think if we're in a simulation, it's hard to do more than guess at what its properties are. We could try to use mathematics and computer science and physics to see what it looks like from inside, what seems possible to compute (making the probably good assumption that computation and math are universally true), etc. But if you get right down to it, a sufficently advanced/evil simulation could paper over whatever it wants just by modifying our brains, removing troublemakers from existence and memory, etc. Gives me the creeps thinking about it much.

> Some people _think_ of an atom as something you can reach out and touch, or grok in some deep sense distinct from performing specific experiments.

Right, with an atomic force microscope.

...but it almost sounds like you're suggesting that atoms and molecules are NOT something you could individually pick up with an AFM?

The parent was referring to the perceived physicality of atoms which you cannot touch or understand except in terms of the forces between your instrument and the bundle of quanta we call an atom.

> The question here is whether spacetime is something like that, or if it is a proxy concept with no grounding whatsoever in the universe, but with good predictive powers?

I'd add: is it something one can develop an intuition for?

I don't think there is anyone who says gluons or general relativity are intuitive. You have to do the calculations (or memorize heuristics) to get something useful. Atoms, on the other hand, or electromagnetic radiation as another example, can be visualized (and mentally simulated) more or less correctly. (Note: atoms. Not nuclear mechanics.)

I don't believe anybody can intuitively "visualize" an atom which leads to accurate predictions, especially the electron cloud and it's collapse. Atoms which manage to stay isolated for long enough are not even localized in space.

Well you can visualize them well enough for basic chemistry if you just view them as spheres with a certain fixed number of lines (bonds) sticking in or out of them. I don't think anyone would call this model real, but it's arguably more relevant to daily life than more fundamental descriptions.

I think you and the author are talking about different concepts. To your credit, for the author's 'real thing like an atom' here, the term real is ambiguous.

The question the author is talking about isn't whether a particular theory is correct, or even whether a theory contains explanatory power. I believe the author is assuming general relativity for the sake of argument, and discussing the concept of existence within the general relativity theory's framework.

For example, take the three-body problem [1]: does the _theory_ of general relativity require us to add initial conditions to the three body problem? In other words, are there more fundamental initial conditions not derivable that must be inserted as inputs into the equations when we use general relativity to solve the problem instead of say Newtonian mechanics? The answer is No. While the equations will change, the initial conditions (the degrees of freedom) do not.

What does that mean? It demonstrates indivisible properties of the universe given a set of theories. With general relativity now 'energy' and 'mass' are interchangeable, but the existence of mass/energy isn't a fictitious force, it can't be explained away by coordinate transformations. Gravity can't either, but a gravity well points to an underlying reality of mass/energy, not the other way around. If it were the other way around, we would say gravity is the 'real' thing, and mass/energy are resulting effects.

[1]: https://en.wikipedia.org/wiki/Three-body_problem

While not a perfect definition of "real", don't we already have images of atoms?

We do. Thinking a bit about what that means, we have photographs showing how the physical entity we model as atoms interact with light, and can confirm that the theory lines up with experiment.

If we adopt that same definition for spacetime, we also have photographs of it, e.g. https://en.wikipedia.org/wiki/Solar_eclipse_of_May_29,_1919#... and https://en.wikipedia.org/wiki/Gravitational_lens#Gallery.

Atoms aren’t just something we’ve modeled — they’re a physical thing, i.e. matter. We can sense them directly, not just infer their existence by measuring related phenomena. I realize when you get into theoretical physics the “thingness” of concepts becomes more and more abstract, but atoms don’t fall into that bucket.

I don't see the distinction you're making between "sensing" and "measuring related phenomena." What's different from using a specialized microscope to take a picture of an atom and using a telescope to take a picture of black hole?

The amount of grant funding! ;P

>We can sense them directly

What does this even mean? Unless you're talking about some supernatural "atomic sense", there's obviously no way to "sense them directly" - we sense them through their interaction with other things combined with our model of those things and the interaction.

Okay, that's fair. But taken to its extreme, we sense everything through its interaction with other things. But obviously, things exist; not everything can be just a mathematical model.

Why not?

Mathematics is too powerful. It can express anything, both real and unreal things. Physical world has limitations.

Check out Max Tegmark's writings on the mathematical universe, some very interesting stuff in that direction.

Will do, thanks for the rec!

You sense spacetime directly too. It's what holds you do this planet we inhabit.

I think atomic phenomena may feel somehow more familiar because it takes far less energy to make them dynamic than it does for gravitational phenomena. If we were somehow huge beings made of binary stars or something, gravity would probably be the very familiar force and atomic phenomena almost inconceivably small to you.

Atomic phenomena are almost inconcievably small to me. There's around a hundred trillion atoms in a human body cell, and around a hundred trillion cells in a human body, plus or minus an order of magnitude or two. It's astonishing to me that people in the early 1900s could measure and calculate the charge on an electron, or that Marie Curie had to process tons of pitchblende to purify micrograms of Radium - and could, and did. Or that Rutherford could say where most of the mass inside an atom is, and split the nucleus of Nitrogen atoms - 1/1000th of the size of the whole atom - and tell what was happening down at that scale - using the technology of 1917.

As a kid I read the whole Michelson–Morley experiment in a public library. It was a quite a peek into how the science of that era was done.

Basically you start with some base information(axioms and postulates), then you gather some experimental evidence for those axioms and postulates. You establish some rules on which your experimental and observational universe works. Then what follows is volumes and volumes of Math.

If Im not wrong relativity itself was developed over a decade. And I'd assume it took tons and tons iterations, retries and tweaking to make it work.

Where common people like us get lost is looking at Symbol manipulation directly. At that stage Math scares the day light out of us.

The right approach is to start with your own axioms, postulates, truths and symbols and start working from there.

This is really like some newbie seeing a super massive code base and thinking it could be outside their reach to ever work with such a thing. But if you look at it carefully. All there is to it is loops, condition statements, operations, variables and basic abstractions.

Spacetime is defined as [x,y,z;t] - 4d array of data.

> You sense spacetime directly too. It's what holds you do this planet we inhabit.

Nope.

Those images, as I understand, constructed by heavy calculations based on a theory of atoms. Isn't it? So it is arguable that images are not "real", they are visualizations of a data.

At the same time not every real image is real. We could sometimes see two Suns in the sky, or we can see a rainbow, but it doesn't mean that the second Sun and the rainbow are real.

We have images of black holes merging too.

We have devices that register electromagnetic radiation spectrum emitted (by ”atoms”) which can be visualized as an image. But this is at the heart of the parent comment: does the atom exist, or is it just a concept that describes the end result when strong interaction, electromagnetic force and gravity make particles stick together in a (usually) consistent way? If we call the atom real, we can call other aggregations like the economy or game theory real as well. But they are just concepts

It isn't much of an argument. Everything we think of as "real" in the sense that you describe it is actually just a conceptual model we have for how certain patterns of energy in the universe behave. The universe has no special rules for apples, floors, wifi access points, cars, or basketballs.

Yeah but what are we actually seeing? An atom is made of protons, neutrons, and electrons, and 99.9% empty space. But you can't really look at an image of an atom and say "there's a proton there" so what is the image of exactly? Rinse and repeat with quarks.

I dont agree its a false dichotomy.

Yes every theory is just a model to fit observations etc etc. But Atoms are conceivable - they work as a model, but we can also imagine how to fit into the reality that we experience. We can 'see' them to an extent with special instruments. And their abilities/behaviours are evident in all the physical world of compounds and elements around us. You can disolve salt in water, you can light up a neon sign.

Quantum Mechanics is a model that fits experiment results very well, but is very hard to conceive of. Wave function collapse has no obvious analogue in the world as we experience it. Thats the point of schrodingers cat. You either have to go for the many worlds interpretation, or get into spooky stuff about observers.

Spacetime/relativity is similar, we can do stuff like fly a clock around the world in an airplane, or observe mercury transitioning past the sun, but its much less conceivable based on our everyday world. So spacetime might just be a constuct, like wave function collapse, that doesnt map easily onto our experience of the world. Certainly its a question worth asking.

Our ability to conceive of something should not be in the definition of reality. Quantum mechanics and relativity have just as much, if not more experimental evidence to support them as atomic theory, and make even better predictions. This has implications in the "real world." If we base our definition on what is "intuitive" instead of what is provable, we end up with things like humour theory in medicine.

I'm not saying we should reject anything if it isn't intuitive. I'm just saying the statement Is Spacetime a real thing like an atom, or just a mathematical construct that is used to describe how mass 'generates' gravity? is not a false dichotomy

"But Atoms are conceivable - they work as a model, but we can also imagine how to fit into the reality that we experience."

"Quantum Mechanics is a model that fits experiment results very well, but is very hard to conceive of."

These two statements stand in contradiction. Atoms only work as a model if quantum mechanics holds. In a non-quantum world, electrons collapse into a nucleus, shedding electromagnetic radiation and lowering their energy as they do so. That's also necessary to understand why neon signs have their colors.

More broadly, if you want to compare one scientific theory to another and contrast their "realness" you have to take quantum mechanics along with your atomic theory.

I guess I'm referring to the theory of atoms as it was before QM came along. In the same way that newtonian mechanics is still useful even though its has been superceded by relativity.

If you're saying I 'have' to take QM along with my atoms, then you also have to take Relativity along with your QM, and at the moment they are irreconcilable. Until we have a 'theory of everything' its ok to use simpler models within boundaries

Agree with this - the jury is still out on "reality" or otherwise of a wavefunction, and people expend their "real" research efforts differently according to their opinion! Saying "well everything is just mathematics" is a bit like advocating solipsism: people will roll their eyes and stop inviting you to cocktail parties, its a boring position.

Another historical example would be electric and magnetic field lines. My impression is that even Faraday originally had doubts about their reality; others certainly did. They could have just been forever considered a useful mathematical construct. At some point it became clear that thinking of them as really existing, permeating space, and having physical properties akin to those of accepted real stuff (momentum etc) was more useful. Many years after GR was formulated there were arguments about similar mathematical objects (dynamical components of a tensor, or "gravitational waves") should be considered real or not. The story is that Feynman convinced many with a simple thought experiment about how they transmit energy.

A scientific theory is a body of knowledge. The special use of the term seems like a false monocotomy.

Reply on news.ycombinator.com