There is no absolute "knowing" in science. Physicists constructed a model and that model is then used to make predictions which are checked againt experiments. And so far quantum mechanics turns out to be an exceptional accurate model.

It doesn't mean that we know it is true. But so far sticking to this weird model with all it's quirks allowed us to build amazing gadgets

Other than some issues with wording (i.e., the misuse of "state"), those are some good questions.

It turns out under reasonable assumptions that any theory that attempts to always assign deterministically evolving "hidden" definite values to measurable quantities while reproducing the predictions of quantum mechanics must be nonlocal. This does not mean that such nonlocal hidden variables theories are necessarily wrong, but introduces issues such as the incompatibility of the dynamics of the hidden variables with the theory of relativity. However, the "standard" Copenhagen interpretation has the same issue of nonlocality in the case of wavefunction collapse.

A second issue with such hidden variables theories that are faithful to the predictions of standard quantum mechanics is that they are often essentially standard quantum mechanics with added complexity in the form of the hidden variable dynamics, which would be undesirable from the perspective of Occam's razor, which disfavors unnecessary complexity.

A third issue is the question of how measurement of a quantity would reveal the true, definite hidden variable value. The Copenhagen interpretation has a similar issue with the question of how measurement causes wavefunction collapse.

One may ponder the hidden variable theories that disagree with the predictions of standard quantum mechanics, but experiments investigating these differences in predictions have repeatedly favored the predictions of standard quantum mechanics to an overwhelming degree.

The particle does have a state before it's observed—it just might not be an eigenstate with respect to the variable that shall be measured, but rather there is a well-defined distribution in said variable which comes from the wavefunction.