“Whenever a fact contradicts one of our beliefs, we are prompted to restore consistency by revising some of the beliefs in our web. But in choosing what to revise, we are no longer guided by facts alone. Starting from the anomalous evidence, we look at the contested belief and its supporting justification(s), and assess how consistency can be most parsimoniously restored in light of the full web of our beliefs. We can end up revising anything, from doubting that we actually observed the anomalous evidence in the first place, to the principles of logic and mathematics that lie at the centre of our web.”