What is Science?

This work addresses the demarcation problem by proposing an axiomatic approach to distinguish science from pseudoscience. It is shown that both Popper's falsification criterion and the verification criterion of logical positivism fail to meet the requirements of a reasonable demarcation line. Building on three simple and intuitive axioms of demarcation, this work introduces a criterion based on evidence, according to which a thesis is scientifically valid if and only if its antithesis can be rejected with sufficient statistical significance. The evidence criterion satisfies all demarcation axioms and applies to both empirical and analytical sciences, while Popper's falsification criterion violates the third axiom of demarcation when it is translated into statistical terms. Moreover, the criterion of evidence aligns closely with modern empirical practice grounded in statistical inference. Through illustrative examples from medicine, biology, astronomy, and mathematics, this work argues that contemporary science already operates implicitly according to the evidence criterion. Hence, the given analysis suggests that the demarcation problem has effectively been resolved in empirical research for nearly a century, even though this resolution has not been made explicit in philosophy.