I'm certainly liking the approach here. It also seems to suggest that rationality can be broader than, say, concepts like logical validity, and include other kinds of discursive games or dialogue goals that shape what constitutes "rational" (i.e., justified/justifiable or prescribed) speech. At the same time, it seems obvious that certain kinds of dialogue, or discursive games, are paradigmatic to us when discussing what reasoning looks like externally. The paradigmatic case still benefits from seeing how it is embedded into other speech activities. I feel like this is what would allow us to truly have the discussion, for example, about whether LLMs can reason or not.