Maybe instead of just epistemic and aleatoric uncertainty, we need a third option which is how certain you are that there aren't bugs in your analysis code.
Maybe instead of just epistemic and aleatoric uncertainty, we need a third option which measures how uncertain you are about which of these words to use