4 ms·
I'm one of those students who questioned it. Never got a satisfactory answer. For example the gravity between two objects depends on the relative speed betwee
by ars 7y ago
I'm one of those students who questioned it.
Never got a satisfactory answer.
For example the gravity between two objects depends on the relative speed between them (since gravity is proportional to energy, not rest mass). Are we calculating that for each and every sun orbiting in a galaxy?
It's a lot of interactions, each sun interacting with all the rest.
It gets worse - gravity is also proportional to potential energy. So you need to include how much energy you would have if every sun fell into every other sun, not just their rotation.
I just can't shake the feeling that dark matter is just this unaccounted for energy in a galaxy.
- philwelch 7y agoThat sounds somewhat plausible to me because it fulfills the criterion of something that would mess up our observations of gravitational force but nothing else. Is the n-body problem really so intractable we can't even handwave our way to an approximate answer that accounts for this stuff?
- dodobirdlord 7y agoEveryone has thought "maybe it's just a bad gravity equation?" at some point. It's a fun idea, but it's contrary to observation. https://en.wikipedia.org/wiki/Bullet_Cluster https://en.wikipedia.org/wiki/Bullet_Cluster
- andbberger 7y agoOof, a gravitational lensing based result at 8 sigma... so much for this fun little theory
- heavenlyblue 7y ago>> gravitational lensing based result What's wrong with basing your results on gravitational lensing? Is it related somehow to the complexity of inferring on how much the images are wrapped by gravitational lensing?
- andbberger 7y agonothing, I'm talking about OP's theory. A gravitational lensing based result is particularly strong due to it being model independent. OP's theory cannot account for lensing inferred CoM differing from the baryonic CoM in the bullet cluster.
- jiggawatts 7y agoThis always gets trotted out, but it isn't ironclad proof of anything. Everything about the Bullet Cluster is conjecture based on secondary observations. They're not measuring "mass", they're measuring light, from which they're inferring mass. The galaxy collision is an unusual case, being at relatively high velocity, so the models may not even cover it well. They're also assuming that the majority of the mass is in the gas, not the stars or invisible brown dwarvf stars or even black holes. It's layers of assumptions a mile deep being used as irrefutable proof.
- andbberger 7y agoI'm not sure what you mean. The n-body problem has no analytical solution in general. You have to simulate. The author is suggesting that assumptions you make (such as representing the earth moon system as a point mass at their CoM) in order to make simulation feasible might lead to a catastrophic divergence from reality over time. This is all a consequence of n-body dynamics being chaotic [1] https://en.wikipedia.org/wiki/Chaos_theory https://en.wikipedia.org/wiki/Chaos_theory
- philwelch 7y agoI don't know a lot about physics, but I do know a little bit about computer science, and it seems to me that--barring some crazy breakthrough in computation--any simulation accurate enough to overcome the chaotic nature of this problem may very well be fundamentally impossible. That in itself would be a pretty fascinating result.
- andbberger 7y agoThe problem isn't the fidelity of the simulation persay, it's that chaotic systems are quite literally infinitely sensitive to initial conditions.
- philwelch 7y agoSure; the possible result that the physical universe is fundamentally uncomputable is the result I find interesting.
- andbberger 7y agoThat is _not_ the result though. This has nothing to do with computability. It's about the behavior of a certain pervasive class of dynamical systems.
- philwelch 7y agoOkay, I’m missing something then.
- sanxiyn 7y agoN-body we can do, for example Illustris Project. To address the concern of this post, one would need non-linear general relativity simulation. That's more challenging.
- chadcmulligan 7y agoWill it be feasible in the near future to run these equations, or is it a millions of years on a million supercomputers type of question?
- Cogito 7y agoBased on my very limited knowledge, there is no way to calculate these equations exactly for the entire universe, the way they are currently formulated. Perhaps more importantly, we don't even have complete information, but only measurements that are necessarily 'averages'. We don't even have a formulation of this model that allows us to understand the 'error' in these measurements, where here error is about making future predictions based off those measurements. To use a perhaps tenuous analogy, say we have a model that describes exactly how air molecules interact with each other. We take measurements of parcels of the atmosphere and are able to describe what the molecules within are doing - pressure, temperature, humidity, etc. If we want to describe what happens to the entire atmosphere, however, we need a different model that describes how the parcels of air interact - cold fronts, wind patterns, rainfall etc. We have continuity between these two models, because as far as we can tell the way these parcels of air interact is consistent with the way individual molecules interact, but the actual models are very different to each other. With general relativity we have a model that works incredibly well for understanding how a small number of things interact with each other, and we assume that it describes the interactions of every object in the universe. We currently use the exact same model to talk about parcels of objects (eg galaxies) in just the same way as we talk about individual objects (atoms? stars?), and we don't know if that is reasonable or not. It's obviously reasonable at a small scale, as when we model things like our solar system we are accurate to an extremely high level of precision but it's not obvious that this continues to hold as we scale up. When trying to work out a weather forecast we don't simulate every single molecule, but instead large parcels of air (at a surprisingly fine level of detail, it must be said, but still relatively huge). It's not even reasonable for us to measure every single molecule, even if we had the computing power to simulate it. Even so, it is far more tractable to simulate every molecule of air in our atmosphere than to simulate GR for every body in the universe, or even every galaxy! We need a model for the interaction of large collections of objects in the universe, and at the moment the best we have is GR. What we don't have is a way of measuring how accurate that is, or even a way of formulating the model that would be consistent when applied to the kinds of collections of objects we're capable of measuring. It's possible that dark energy and dark matter are real things we can't (currently) directly observe, or they could be artefacts that arise when you try and apply GR to 'averages' of large collections of objects, but right now we don't have a way to quantify how accurate those models are in order to rule out the possibility one way or the other.
- XorNot 7y agoExcept this is trivially testable because it should produce defined gravitational lensing effects which follow baryonic matter, rather then interactions like the bullet cluster where gravitational lensing has detached from the observable matter.
- dodobirdlord 7y ago> For example the gravity between two objects depends on the relative speed between them (since gravity is proportional to energy, not rest mass). Are we calculating that for each and every sun orbiting in a galaxy? Disk stars in the Milky Way only have velocities on the order of a few hundred km/s. The relativistic contribution to gravity from such velocity is negligible.
- teh_infallible 7y agoI believe yours is the standard response that the author of the article was referring to, and taking issue with.