[jleyank]

(1) seems incorrect. Large scale cosmology wrestles with the observational or physical bounds of the universe and what that means. Small scale goes granular due to the uncertainty principle. (Planck length).

The others seem to run together. In general, classical modeling works with problems appearing at the very large (relativity) and very small (quantum mechanics). At the very small, the act of observing something influences it’s future behaviour (or, that behaviour becomes unpredictable). But the overall behaviour is statistical and thus predictable (quantum effects like diffraction or tunneling).

Check out Brian cox, Timothy ferris, hawking or other science popularizers.