Physicists nowadays adopt an approach called “model-dependent realism” when using theories or models to explain how the world around us, from the sub-atomic level to the entire universe, works. The idea is that when a theory or model used is successful in explaining a particular physical situation/observation then we assume that it is true and valid. By “successful” we mean that the results predicted by the model agree with experimental results. It may be that an entirely different model for the same physical situation explains it just as well, but uses completely different laws and theories, that may have nothing in common with the ones used in the previous model.
In that case we really have no way of attributing more or less validity to any one of the two models and they are both considered to be equally valid and both offer an adequate description of reality. Adopting “Model-Dependent Realism” became necessary when theories such as Quantum Mechanics used to describe the sub-atomic world, predicted models and behaviours that differed wildly from the physical models based on classical theories, such as Newton’s Gravitation Law and Einstein’s Special and General Theories of Relativity, which brake down on the atomic level.
The important distinction to make here is that the specific model we are using to describe a particular physical situation does not have to be in accordance with our intuition or with what we see around us. After all, what we see is just a representation of the physical world put together by a combination of our minds and our eyes, and/or our other sensory organs. No matter how crazy the model and how completely impossible or absurd its implications, if the predictions it makes are always in accordance with experimental results then the model is assumed to be valid.
Perhaps the most spectacular example of this is Richard Feynman’s sum-over-histories theory which is used to explain the otherwise unexplainable results obtained in the (in)famous double slit experiment where discrete sub-atomic particles are fired towards a barrier containing two slits. In this experiment -which in Feynman’s own words “contains all the mystery of quantum mechanics”- interference patterns are formed on the other side of the slits as if there was a wave that passed through the slits and not discrete particles.
Even more, when the same experiment is carried out by firing single photons and ensuring that the they reach the screen one at a time the same interference pattern is obtained in the end. Amazingly enough, if one of the slits is closed or opened the particles seem to be able to acquire this information somewhere during their journey from the source to the screen and the interference pattern brakes up or is formed accordingly.
In order to explain this behaviour physicist Richard Feynman who “worked at the California Institute of Technology but also played the bongos at the strip joint down the road” proposed a radical theory which says -wait for it…- that the particles we shoot towards the barrier do not just take a single path from the source through one or the other slit until they reach the screen in the back but instead they take ALL POSSIBLE PATHS to travel from the source to the screen, and that all these paths are taken SIMULTANEOUSLY!! ALL POSSIBLE PATHS means that the particles can go to the other end of the universe and then through your coffee before they pass or not pass through one of the slits and hit the screen.
One can see how by having a particle take all possible paths simultaneously the interference pattern obtained on the screen can be explained. But the idea is still completely insane. The reason why apples or humans or pickup trucks will not form interference patterns and won’t travel to the end of the universe before they hit the screen is that on a larger scale this kind of behaviour brakes down. Whole universes -on the other hand- are not so lucky. If the theory of the Big Bang is true then at some point in the very distant past our universe was so small in size that it would be governed by quantum theory and therefore it would take all possible paths (that would satisfy the no-boundary condition) and only some of those paths would lead to universes with physical laws similar to what we know today. Most probably a great number of other universe also exist with entirely different physical laws, depending on which path the universe took.
Feynman developed a complete mathematical formulation for calculating the sum over histories and a specific graphical representation and even drove around with a truck painted with Feynman diagrams. An important implication of the sum over histories model is that there is no single past history but all possible past histories. An even more astounding implication of that, although maybe not immediately apparent, is that in the quantum world the (unobserved) past is not fixed but like the future it is indefinite and exists only as a spectrum of possibilities. And so it is with our universe and its past.