OK, so “valid” reasoning is not fallible? A noisy or imperfect conclusion is not “valid” reasoning? Because I think we see such imperfect, fallible reasoning not only in many animals, but in ourselves as well. It is not a stretch to think that natural selection could produce neural circuits that have that kind of reasoning.