The talk consisted in critical reconstruction of Quentin Meillassoux's argument in favor of absolute contingency, where I show that his argumentation is deeply flawed and circular. Moreover I tried to show that it is generally impossible to prove something like "ontological contingency" from an empirical as well as rationalist standpoint.