hypothetico-deductive method a method of testing hypotheses. Thought to be preferable to the method of enumerative induction, whose limitations had been decisively demonstrated by Hume, the hypothetico-deductive (H-D) method has been viewed by many as the ideal scientific method. It is applied by introducing an explanatory hypothesis resulting from earlier inductions, a guess, or an act of creative imagination. The hypothesis is logically conjoined with a statement of initial conditions. The purely deductive consequences of this conjunction are derived as predictions, and the statements asserting them are subjected to experimental or observational test. More formally, given (H • A) P O, H is the hypothesis, A a statement of initial conditions, and O one of the testable consequences of (H • A). If the hypothesis is ‘all lead is malleable’, and ‘this piece of lead is now being hammered’ states the initial conditions, it follows deductively that ‘this piece of lead will change shape’. In deductive logic the schema is formally invalid, committing the logical fallacy of affirming the consequent. But repeated occurrences of O can be said to confirm the conjunction of H and A, or to render it more probable. On the other hand, the schema is deductively valid (the argument form modus tollens). For this reason, Karl Popper and his followers think that the H-D method is best employed in seeking falsifications of theoretical hypotheses. Criticisms of the method point out that infinitely many hypotheses can explain, in the H-D mode, a given body of data, so that successful predictions are not probative, and that (following Duhem) it is impossible to test isolated singular hypotheses because they are always contained in complex theories any one of whose parts is eliminable in the face of negative evidence. See also CURVE-FITTING PROBLEM , DUHEM , TESTABILIT. R.E.B.