testability in the sciences, capacity of a theory to undergo experimental testing. Theories in the natural sciences are regularly subjected to experimental tests involving detailed and rigorous control of variable factors. Not naive observation of the workings of nature, but disciplined, designed intervention in such workings, is the hallmark of testability. Logically regarded, testing takes the form of seeking confirmation of theories by obtaining positive test results. We can represent a theory as a conjunction of a hypothesis and a statement of initial conditions, (H • A). This conjunction deductively entails testable or observational consequences O. Hence, (H • A) P O. If O obtains, (H • A) is said to be confirmed, or rendered probable. But such confirmation is not decisive; O may be entailed by, and hence explained by, many other theories. For this reason, Popper insisted that the testability of theories should seek disconfirmations or falsifications. The logical schema (H • A) P O not-O not-(H • A) is deductively valid, hence apparently decisive. On this view, science progresses, not by finding the truth, but by discarding the false. Testability becomes falsifiability. This deductive schema (modus tollens) is also employed in the analysis of crucial tests. Consider two hypotheses H1 and H2, both introduced to explain some phenomenon. H1 predicts that for some test condition C, we have the test result ‘if C then e1’, and H2, the result ‘if C then e2’, where e1 and e2 are logically incompatible. If experiment falsifies ‘if C then e1’ (e1 does not actually occur as a test result), the hypothesis H1 is false, which implies that H2 is true. It was originally supposed that the experiments of J. B. L. Foucault constituted a decisive falsification of the corpuscular theory of the nature of light, and thus provided a decisive establishment of the truth of its rival, the wave theory of light. This account of crucial experiments neglects certain points in logic and also the role of auxiliary hypotheses in science. As Duhem pointed out, rarely, if ever, does a hypothesis face the facts in isolation from other supporting assumptions. Furthermore, it is a fact of logic that the falsification of a conjunction of a hypothesis and its auxiliary assumptions and initial conditions (not-(H • A)) is logically equivalent to (not-H or not-A), and the test result itself provides no warrant for choosing which alternative to reject. Duhem further suggested that rejection of any component part of a complex theory is based on extra-evidential considerations (factors like simplicity and fruitfulness) and cannot be forced by negative test results. Acceptance of Duhem’s view led Quine to suggest that a theory must face the tribunal of experience en bloc; no single hypothesis can be tested in isolation. Original conceptions of testability and falsifiability construed scientific method as hypothetico-deductive. Difficulties with these reconstructions of the logic of experiment have led philosophers of science to favor an explication of empirical support based on the logic of probability.
See also CRUCIAL EXPERIMENT , DUHEM, HYPOTHETICO -DEDUCTIVE METHOD , PROBA — BILIT. R.E.B.