He says that in real science when a prediction is found to be false, scientists assert the experimental conditions and assumptions and drive the refutation through unless they can clearly say what the problem is with either the experimental conditions or the assumptions. Adjustments can be made to the experimental conditions or the assumptions if they meet two criteria. The adjustments must account for the failed prediction and they must allow new successful predictions to be made. If the criterion is not met then the scientists are forced to drive the refutation through, therefore making the theory false. When a prediction is found to be true the scientists can not say that the theory is true or probably true, but they must say they have failed to falsify. This line of argument conflicts with the thinking of confirmationalist or inductivist philosophers including Hempel, Nagel, and Duhem. They would argue that if the prediction is true, the theory is true or probably true. Popper rejects this argument because of its use of induction which he also rejects. .
Popper is sensitive to the problem of induction formulated by Hume in the 1700's which can be shown through the use of an example. An inductive argument could be every morning for the past 3000 years the sun has come up, therefore the sun will probably come up tomorrow. This argument sounds good, but what actually warrants the claim the sun will come up tomorrow? Hume proposed that to inductive arguments we need to add a principle, such as, nature is uniform or the future resembles the past. This would then turn our argument into for the past 3000 years the sun has come up and the future resembles the past therefore the sun will probably come up tomorrow. This also sounds good, but still makes use of induction to get the rule the future resembles the past. Russell proposed to add to inductive arguments the laws of nature or in other words laws of physics.
