But you can argue against someone misunderstanding what a theorem says and applying it in the wrong context.

The question is about tokenization, not parsing.