A civil engineer designs a bridge. The bridge is built. Traffic drives across it. The bridge holds or it does not. If it holds, the design was adequate. If it collapses, the design was wrong. No committee vote changes the outcome. No consensus of bridge experts can prevent the collapse if the steel is undersized. No peer review process can make a failed bridge hold weight. The bridge answers to gravity, and gravity does not read journals. A pharmaceutical chemist designs a molecule to bind a receptor on a bacterial cell wall. The molecule is synthesized. It is introduced to a bacterial culture. The bacteria die or they do not. If they die, the molecule works. If they do not, the molecule does not work. The chemist's institutional prestige, publication record, and funding history have no influence on whether the molecule binds. A semiconductor engineer designs a transistor at a three-nanometer process node. The transistor switches or it does not. It switches at the specified speed or it does not. It holds its state for the specified duration or it does not. There is no debate. There is no interpretation. There is measurement, and there is function. Every person reading this paper trusts these outcomes implicitly. Nobody debates whether bridges hold. Nobody argues about whether antibiotics kill bacteria. Nobody questions whether their phone's processor computes. These things work, and the fact that they work is self-evident to everyone who uses them. Ask a person on the street what produced these outcomes, and they will say: science. This is where the confusion begins.
Geoffrey Howland (Fri,) studied this question.