The self-reinforcing successes of mathematics and experimentation, which science depends on much more than language, would seem to escape the linguistic relativism and false sense of confidence you're trying to present (if I understand your comment correctly).

In general, though, personal confidence in our personal understanding is indeed a very fuzzy thing that probably does have a lot to do with popular acceptance. Eventually neuroscience will help us understand this better.

But, the scientific method and mathematics together provide a way to escape the limitations of perceived correctness now. Instead of our brains judging what's correct, we have logically independent methods (science proper and math) that can make the decisions for us.