This conference paper examines Quentin Meillassoux's critique of correlationism, with special attention to facticity and to the role of scientific methods in securing rational access to reality. It reconstructs the argumentative path from the distinction between primary and secondary qualities to the thesis that only contingency is necessary, and it assesses the philosophical scope of this position in light of Kantian critical limits. The paper also advances an alternative reading: rather than abandoning metaphysics altogether, it proposes an experimental metaphysics of the sciences that preserves methodological rigor while formulating absolute yet revisable hypotheses.