Lecture
The theoretical and experimental hypotheses stand in an implicative relation to each other. It is represented by the scheme: if the proposition P is true, then at the empirical level a regularity q follows from it (corresponding to the EH).
Example: Festinger's theory: dissonance → change of attitude.
There can be many such particular consequences derivable from the statement P.
The statement «if P, then q) is not purely logical, since q belongs to the domain of empirical reality, while P remains part of the «world of theories».
A scientific hypothesis implicitly assumes that the consequence q will occur if the theory P is true. However, there is no constraint implying that q cannot be derived from some other theory. (the analysis of competing explanations is the third condition of causal inference).
Distinguishing the rules of logical inference from the problem of relating empirical data to statements requires understanding that experimental hypotheses are always open to further testing.
The asymmetry of inference consists in the different consequences accepted when empirical data are obtained «for» and «against» the experimental hypothesis.
That is, an equality of probabilities of obtaining data both for and against is required.
And if the results correspond to the EH, then we may refrain from rejecting it as a relatively correct description of the empirical dependency (i.e. until other data are obtained).
The norm of reasoning here is described by the modus tollens syllogism. If the statement q (reflecting the empirical content of the EH) is a consequence of P (a generalized statement), then obtaining not-q as a result of the experiment under such a premise requires refuting the truth of P.
[( p→q) & not-q] → not-p
This principle of asymmetry of inference from experimental data was examined in detail by K. Popper as the fundamental norm of hypothetico-deductive inferences.
Comments