The Sapir-Whorf hypothesis, also known as (the) linguistic relativity (hypothesis), more or less states (in one of the most specific descriptions by Whorf): users of markedly different grammars are directed by their grammars towards a different evaluation of reality ('Language, Thought and Reality' (selected writings by Benjamin Lee Whorf), 1940: 221).
To my knowledge, research into this hypothesis has to date not touched on the paradox it introduces (and please let me know when I'm mistaken). How is this hypothesis paradoxical?
When someone denies the hypothesis, it means that he or she has arrived at a different evaluation of reality than Whorf, as witnessed by his articulation of the hypothesis. Understanding that Whorf not necessarily meant 'actual wording' by 'grammar' (implying that it's not because those people don't understand the sentence that they reject it), upon asking for clarification of their rejection, they will rationalize their opinion as to why they consider the hypothesis to be wrong. In other words: although in all likeliness expressed in English, their rationalization will reflect an underlying logic different to the one they will say to have read in the hypothesis they rejected. That underlying logic (expressed in a configuration of arguments) will therefore indicate a different 'grammar' which orients the intention of the words they use.
The paradox is that, when people reject linguistic relativity, they can only attempt to make their rejection understood by creating a different grammar intended to persuade their opponents to share their evaluation of reality, not Whorf's. Two grammars, two evaluations of reality (hypothesis is wrong, hypothesis is right). They will, at that point, have paradoxically proven the hypothesis.
Or am I wrong?