"If the packet is to be reduced, the interaction must have produced knowledge in the brain of the observer. If the observer forgets the result of his observation, or loses his notebook, the packet is not reduced. We are again led to emphasize the fact that the wave function of a pure-state assemblage is merely a mathematical tool for computing from all previous observations what the relative probabilities are for different results when we make our next observation."
January 1, 1970