Abstract
Kuhn defended the so-called relativised apriori. Even if the apriori is not taken apodictically, it still preserves a constitutive function.
For there is, according to Kuhn, something more to experience than a simple perception. To put the idea in Fleckian terms, to see is to look and to know.
We have to know the meanings of "paradigmatic propositions" in order to have an experience. If all observations are "paradigm-induced", di) erent paradigms are constitutive of different experiences.
Nonetheless, paradigms are not wholly independent on empirical input. Paradigmatic propositions are of empirical origin: they are, as Poincaré used to say, "hypotheses erected into principles".
The role of empirical input in knowledge is hence somewhat paradoxical: it motivates the part of knowledge that is independent on it.