Science Puzzle
The Theory That Explains Everything
A theory is presented to you. It explains why it rained, and it equally explains why it did not. It explains her recovery and it explains her decline. Whatever happens, its supporters can show how the theory accounts for it. They regard this universal explanatory reach as the strongest argument in its favour.
Why is this a fatal weakness rather than a strength?
The Answer
Because explanatory power and predictive power are not the same thing, and only one of them is worth anything. A theory that can explain any outcome after the fact has, by that very fact, forbidden no outcome in advance. And a theory that forbids nothing predicts nothing.
Work through what that means operationally. To test a theory you must be able to say, before you look: if the theory is right I should see this, and if I see that instead, the theory is in trouble. A theory compatible with every observation makes that sentence impossible to write. There is no observation that would embarrass it, so gathering evidence cannot tell you anything about whether it is true, so no amount of evidence can ever count in its favour either. It floats entirely free of the world.
This is what Popper was pointing at when he compared Einstein with the theories he regarded as pseudoscientific. Einstein's general relativity predicted a specific deflection of starlight, a number, measurable during an eclipse. If the number had come out differently the theory would have been dead. That risk is what made the 1919 confirmation mean something. A theory that had been able to accommodate any deflection would have gained nothing from the eclipse at all.
The confusion arises because universal explanation feels like strength. It is enormously satisfying to have a framework that makes sense of everything you encounter, and its defenders experience each new event as further confirmation. That feeling of endless confirmation is precisely the symptom. Genuine confirmation is rare and expensive, because it requires a prediction that could have failed and did not.
So the diagnostic question to put to any theory, including your own, is not "what does this explain?" It is: "what would this theory rule out? What could I observe tomorrow that would show it to be false?" If the answer is nothing at all, then the theory is not competing with rival explanations of the world. It is not making a claim about the world in the first place.
The principle: Explanatory power versus predictive power. A theory that accommodates every possible outcome forbids none, and so predicts nothing and can never be tested. Universal explanation is a symptom of emptiness, not strength.