Title: A quantitative approach to semantic informativity

Abstract: The present article defines a way to measure semantic informativity of deductions. This is done by means of dynamic concepts of complexity and relevancy, which are, in turn, defined according to explicit operations of insertion and deletion on databases. Hence, with respect to finite databases, it proposes a unified solution to Bar-Hillel-Carnap paradox, showing how some tautologies may be informative but contradictions cannot, and to the Hintikka' scandal of deduction, explaining how some deductions can be informative and others not.