Can laws of physics suddenly change? At the present time, we have a collection of variables and laws that describe, maybe in a non deterministic fashion, the evolution of the universe.
Is it possible that these laws suddenly change? Say, for example, that the sign of an interaction change, and all the other laws are modified coherently.
What would happen in this case? Of course there are apocalyptic scenarios, in which all things disgregate instantaneously. Is there a scenario that would not result in a complete mess?
Edit. Let me clarify that I am not looking for an answer concerning "what we think is a law" or "the knowledge we Have of the actual laws"; this would pertain the subject of pure logic.
Say for example you have a collection of polynomial equations (like are most of the classical laws). For example, the gravitational interaction force for two planets can be written as
$$ F_{12} R^2 = G M_1 M_2 $$
This can be restated as
$$ \frac{d}{d M_1 } \frac{d}{dM_2} ( F_{12} R^2 ) = G = constant $$
In fact with this procedure all the polynonial equations can be restated in terms of something = constant.
A (not small) change in these constants would be a change in the actual physics of the universe, and not only that we knew things in a wrong way. Things would start to interact differently. I don't know if all the laws of physics can be stated as something being constant - as Nielsen noticed, Noether's theorem give us a lot of other candidates to be constant. In any case, maybe I explained myself better!
 A: One way to think of laws of physics is as update rules for some set of states, a bit like how a computer has an operating system allowing various applications to interact with each other in particular ways. From this perspective, where the laws of physics are not inherent in the physical states themselves, anything could happen. The operating system could after all have a hidden rule that after next Tuesday the allowed interactions would be something completely different, and there would be no way to determine this since we can only observe the past regularities of interactions (as far as we have data... and that data could in principle have been faked at the last reset). There is no way of limiting how often or how radically physics could change in this case.
Another way of thinking about laws of physics is to think of them as the patterns we see in repeated experiments or observations that describe or predict natural phenomena. The complete set of physical laws would be what we would know if we did every possible observation and constructed the best model possible from them. This is the "inside" perspective, and it does not and can not give evidence for any radical changes. We cannot exist in a reality with different coupling constants or number of spatial dimensions, so observed laws (as observed by us) cannot change fundamentally.
Note that small changes may be survivable and memorable - if several dimensionless physical constants changed together in a minor way (see Victor Stenger's critique of "fine tuning") life would go on but measurements may reliably tell that something really weird happened. That would force the models in the second approach to incorporate the possibility of such weird changes somehow.
In the end the main answer is anything could happen, but what we really care about is prediction ability: what can we predict given what we know (or could know)? Random physics changes is mostly decision-irrelevant since we cannot predict it. One could argue it should set our long-term discount rate of the future to the expected rate of physics change, but constructing a sensible prior without any evidence likely ends up with a scale-free prior for the rate of physics change, which implies some rather weird hyperbolic discounting.
A: At the time I am adding this, there are some good answers, but none that make the following statement completely clearly and upfront:
the 'laws' that science deals in are descriptive, not imperative
That is to say, a 'law of nature' is simply some observed regularity in natural phenomena. The observed regularity gets called a 'law' when it is of sufficiently wide occurrence or relevance. Asking whether the laws can change is no more and no less than asking whether regularities which have been observed up till now might not be quite as universal as was thought. The answer to that is yes. For example, it was thought for a long time that there was a law of conservation of mass. It turns out that was wrong, but not entirely wrong. Once we understand the relationship between mass and energy then the mass conservation 'law' is seen to be the form that conservation of energy takes in some circumstances.
In chemistry it was thought that the chemical elements were immutable, and then radioactivity was discovered which shows that there are processes which change one element into another.
It was also thought for a long time that if one connects three points with lines of minimal distance, then the triangle thus formed will have internal angles summing to 180 degrees (i.e. half a full rotation). Up until about 1917 this would have been said to be a very concrete and universal regularity of nature. It turns out it is not in fact true in general; it is just approximately true in ordinary circumstances.
Now it will be said that the question asked is not about human misunderstanding about laws of nature, but the laws themselves. Let's translate that. It is asking "can the regularities of nature themselves change?" But if they change then they were not regularities after all! Well not universal ones anyway. So if they were universal, describing phenomena equally well at one time as another, then by definition they cannot change.
Philosophers of science have of course given a good deal of thought to all this, and a very common conclusion is that one cannot give a good definition of "laws of nature" except the working definition "our best summary so far of the regularities which have been perceived". The whole of science is based on induction, which amounts to saying it is based on the reasonable conjecture that the universe will carry on being like it has been in the past, just changing over very long timescales. For example, the values of some of the parameters that physics normally treats as constant might possibly change over many billions of years. Experimental searches for such effects have been carried out, taking advantage of the precision of atomic clocks and the long time-scales offered by astronomy.
Just as we have no guarantee, when venturing out for a walk, that some calamity might not happen, such as a rock fall or an earthquake, equally we have no absolute guarantee that the fabric of the universe will not undergo some profound change which it has been building up to, all unknown to us. So what shall we do? Better accept the uncertainty and go out for that walk, and also carry on with all our other efforts to do something creative.
A: There's a difference between the "laws of physics" which are our current best model for how the universe works, and the true laws of physics, the underlying behaviour and nature of the universe.
We could quite easily encounter a situation that defies the former - we already know that our two best models, for astronomical and microscopic scales, don't mesh with each other, so we must be missing something. If we had such an encounter, that doesn't really mean anything about the universe has changed - it just means that our ideas of how it was presumed to work are wrong, and furthermore, always were wrong, although possibly only by a little bit. (It turns out that Newton always had been wrong about the nature of gravity and mechanics, but he was so close to being right that we didn't notice for centuries, and even now we have noticed, we still assume his model because it's close enough. Einstein will probably be discovered to have always been slightly wrong in the same way, but with an even smaller "slightly.")
Considering the "true" laws, it's not obvious what it would mean for them to change. No amount of experimental evidence can rule out a future encounter that breaks your theories, so in practice it's not really possible to distinguish between the true laws changing and you just being mistaken about what they were to begin with.
A: If any of the fundamental laws of physics suddenly became invalid, mayhem would ensue and carnage would result on all fronts- and the universe as we know it would end, and so would we.
But be not dismayed. One of the most basic laws governing the universe is the conservation of energy, which requires that all the rest of the laws of nature cannot simply stop working at any time. In other words, if the laws DID quit on us, energy would not be conserved anymore- and magic would become real, because magic is based on energy nonconservation!
A: We cannot prove that the laws of physics can't change
This is a branch of the philosophy of science explored by Nelson Goodman and David Hume, referred to as the Problem of Induction. Essentially we can't experimentally differentiate between statements like "This emerald is green" and "This emerald is grue", with grue being defined as a colour that changes from green to blue at time X. This generalises to laws of physics. We think that we're in a universe where F=MA, but we could be in a universe where F=MA before time X. Then after time X; F=MA^2
https://existentialcomics.com/comic/2
https://plato.stanford.edu/entries/induction-problem/#GruParNewRidInd
A: This is an interesting question, and as has been stated, our known laws are often approximations of more fundamental laws of physics. It is obviously impossible to know if any of our known laws is a "fundamental" according to mother nature, or even if the concept is sensible.
As Anders touched in his answer, constants with dimensions (like gravitational constant) are not very good probes for changes in laws of physics. All such "constants" as measurements of something against a known measuring stick. How would you know it was not the stick that changed?
There are dimensionless constants, the fine structure constant alpha as the best example. These would probably be better probes.
Any kind of conclusive answer would be the discovery of a fundamental no-free-parameters-theory. Such a theory would require no tuning, or experimental measurements, and there would be no arbitrary natural constants! Everything would just somehow pull itself out of pure mathematics. The universe would be how it is, because to exist is to be identical to it. It's hard to imagine how one would even begin such a theory.
A: The actual laws of physics can’t change — by assumption, they are valid at all times and in all places. But our knowledge of the laws may be (almost certainly is) sufficiently defective that something happens to change our understanding of them. Just for example, parity conservation was one of the presumed laws of physics until 1957, when an experimental result changed that law.
A: This is an interesting question. I do not know much about the law of physics, but I believe what you are suggesting, in some sense, happens continuously in the social realms. In the latter realm, there is no "complete mess" scenario, up to a relatively conservative view regarding what "a complete mess" is (are world wars a complete mess?). What I mean specifically is the following.
I guess there could be two views about the "causal foundations" of social phenomena. They could be described by stochastic laws, which probably are more volatile than physics laws, or alternatively, there are no laws which describe it. In the second case, I guess there is sign reversion all the time. In the first scenario, there is also sign reversion. People may react qualitatively differently to different stimuli, meaning that laws are stochastic at the individual level, but not in the aggregate (I guess this is different in physics). In none of these cases the outcome is a complete mess.
Hence, my point is twofold:

*

*A brave analogy between the social and the physical realm could lead us to conclude that sign reversion may not imply necessarily complete destruction.


*A vision according to which the social realm is just a more complicated physical realm could point in the direction that what we believe are interactions with a constant sign are only a realisation of aggregate stochasticity of which we do not view the other side. In this case, I guess everything is fine.
Sorry for the probably meaningless philosophical diversion.
