Falsification , also falsification (from Latin falsificare "to recognize as wrong") or refutation , is the proof of the invalidity of a statement , method , thesis , hypothesis or theory . Statements or experimental results that can prove invalidity are called “falsifiers”.
A falsification consists of the proof of immanent inconsistencies or contradictions ( contradiction ) or the incompatibility with instances accepted as true (contradiction to axioms ) or of the detection of an error . Methodically, one confronts the contradicting statements that follow from the initial assertion as a counter-hypothesis or antithesis .
In science , falsification is a result of validation alongside verification . In the philosophy of science according to Karl Popper , the falsifiability of a theory or hypothesis plays a central role. Falsified statements, theses, theories are worthless for science as a method of gaining knowledge and are rejected. They only make sense in the historical perspective, in order to learn lessons from wrong approaches. Details and problems of this approach are covered under Falsificationism . It should be noted that falsifications take on the function of an "activity trigger" to investigate the problem more closely. Methodologically, it does not necessarily follow that the falsification of a statement immediately results in the rejection of the underlying theory (see also paradigm and Duhem-Quine thesis ).
In classical propositional logic , falsification is the assignment of the truth value incorrectly to a statement within the framework of the principle of two- valued , likewise in the calculus of Boolean algebra .
Evaluation of experimentally obtained data
The evaluation of data obtained through experiments is also used for falsification, in empirical stochastics for example in the context of a sample plan , in order to describe the results of observations that do not correspond to an assumed probability, or to assess causalities (null hypothesis) . See in detail hypothesis (statistics) .
In computer science, falsifying processes are used to determine the presence of errors in software, for example through static code analysis .