A fondamental problem in the discussion on the foundations of mathematics is to clarify what an axiom is. This is especially important in the light of the most recent advances in set theory where new axioms have been proposed whose legitimacy is highly controversial (for example, large cardinal axioms); this paper is a contribution to this discussion. By analysing the view of Poincaré and Hilbert on axioms, we observe that, despite the deep differences in their philosophical thinking, the two logicians came to the same conception of the axioms of geometry as definitions in disguise. We revisit and generalise this view by arguing that any axiomatic system (set theory in particular) is the definition of some concepts.